OPINION AND ORDER
GOLDBERG, Senior Judge:This case concerns the third administrative review of a countervailing duty order on citric acid and certain citrate salts- from the People’s Republic of China (“China” or “the PRC”). See Citric Acid and Certain Citrate Salts from the People’s Republic of China, 79 Fed.Reg. 108 (Dep’t Commerce *1292Jan. 2, 2014) (final admin, review) (“Final Results ”) (covering imports from January 1 to December 31, 2011). Plaintiffs, the RZBC Group Shareholding Co. and related companies (“RZBC” or “Plaintiffs”), sue to reduce the final countervailing duty rate imposed on them by the U.S. Department of Commerce (“Commerce” or “the agency”). The Government of the People’s Republic of China, Ministry of Commerce (the “GOC” or “Plaintiff-Intervenor”), also sues, making arguments above and beyond those lodged by RZBC. Constituents of the U.S. domestic industry — including Archer Daniels Midland Company, Cargill, Incorporated, and Tate & Lyle Ingredients Americas (the “Defendant-Intervenors”)' — ■ side with the agency in defending the countervailing duty rate against these attacks.
After carefully considering the parties’ briefs and the record, the court remands one issue to Commerce for reconsideration: the calculation of world benchmarks for the steam coal, sulfuric acid, and calcium carbonate subsidies. The agency must properly address whether to render the benchmarks using weighted averages or simple averages. Otherwise, the court sustains the Final Results in all respects.
GENERAL BACKGROUND
Countervailing duties serve the same purpose as their better-known cousins, an-tidumping duties: They level the playing field between U.S. manufacturers and their overseas competition. But each regime addresses a different problem. Anti-dumping duties (“ADs”) were made to fight price discrimination, so if a foreign producer sells goods in the United States for less than in the home market, ADs bring the U.S. price back to fair value. See 19 U.S.C. § 1673 (2012). Countervailing duties (“CVDs”), by contrast, were created to correct the cost-distorting effect of subsidies. When a foreign government lends support to a producer, CVDs boost the producer’s U.S. prices to offset the net benefit from the subsidy. See id. § 1671(a).
This appeal challenges the Commerce Department’s CVD procedure and calculations from soup to nuts. The background that follows sets the table.
A CVD investigation usually starts with a petition. The purpose of the petition is, quite simply, to alert the agency to the possibility of a subsidy. In this sense, the petition is like a civil complaint. It must allege the rough contours of the subsidy, and it has to contain “information reasonably available to the petitioner supporting those allegations.” Id. § 1671a(b)(l). But Commerce cannot refuse to investigate unless it “is convinced that the petition and supporting information fail to state a claim upon which relief can be granted.” S.Rep. No. 96-249, at 47 (1979), reprinted in 1979 U.S.C.C.A.N. 381, 433. The bar for launching a CVD inquiry is low.
After Commerce accepts a petition and begins investigating, it must decide if the alleged subsidy really exists. By statute, a subsidy may occur when a foreign government “provides a financial contribution ... to a person and a benefit is thereby conferred.” 19 U.S.C. § 1677(5)(B). Financial contributions come in all shapes and sizes, and can include “the direct transfer of funds, ... tax credits or deductions,” and the provision of “goods and services.” Id. § 1677(5)(D). And while all subsidies must spring from a foreign government, it does not matter if “the subsidy is provided directly or indirectly” to the producer. Id. § 1677(5)(C). Commerce can find a subsidy exists even if the foreign authority funneled its donation to the recipient through private parties. See Delverde, SrL v. United States, 202 F.3d 1360, 1366 (Fed.Cir.2000) (“[I]n order to find that a person received a subsidy, Commerce [must] determine that that person received ... a *1293financial contribution and benefit, either directly or indirectly.... ”).
Once the agency pinpoints a subsidy, it must decide if the subsidy is eountervaila-ble, or eligible for CVDs. Because the statute defines “subsidy” so broadly, it is simply impossible to countervail all the benefits that foreign producers take from their governments. (Imagine what chaos would ensue if Commerce tried to slap CVDs on every government loan, tax loophole, and public project in a foreign jurisdiction). For this reason, the law lets Commerce countervail subsidies only if they are “specific.” 19 U.S.C. § 1677(5)(A). A domestic subsidy is specific if the foreign authority limits the pool of recipients by law — a de jure subsidy— or if the subsidy is given to a select number of industries or enterprises — a de fac-to subsidy. See id. § 1677(5A)(D).
Next, after it’s found a specific subsidy, Commerce measures the benefit to the foreign producer or “adequacy of remuneration.” See id. § 1677(5)(E). If a foreign power furnished subsidized goods and services, the agency calculates benefit using the following formula. First, Commerce finds the price that the foreign producer actually paid for the subsidized goods. Second, it determines the price that the producer would have paid had it bought the goods on the open market. See id. § 1677(5)(E)(iv). The agency can do this in at least two ways. It can draw unsubsidized market prices from the producer’s home jurisdiction, 19 C.F.R. § 351.511(a)(2)(i) (2015), or it can use “a world market price where it is reasonable to conclude that such price would be available to purchasers in the country in question,” id. § 351.511(a)(2)(h). If more than one world price is available, the agency can build a market price or “benchmark” by averaging available prices, “making due allowance for factors affecting comparability” to the home country. Id.
Commerce then subtracts the price actually paid for the subsidized goods from the market price. If the former is less than the latter, then the producer garnered a benefit from the subsidy. The agency can levy a CVD on the producer’s exports in an amount equal to the net subsidy, expressed as a percentage of the foreign product’s U.S. sale price. See 19 U.S.C. § 1671(a).
In the review below, the agency imposed a 35.87% total CVD on RZBC’s citric acid exports. Final Results at 109. The duty aimed to offset the benefit RZBC received from steam coal, sulfuric acid, calcium carbonate, land-use, and other subsidies from the Chinese government. See Issues & Decision Mem. (“I & D Mem.”) at 17-37, PD 233 (Dec. 27, 2013) (listing subsidies examined). Now, in their appeal, RZBC and the GOC contest elements of the Final Results, including (1) the agency’s decision to investigate calcium carbonate subsidies; (2) the finding that the calcium carbonate subsidy was specific; (3) the decision to countervail subsidies passed through private trading companies; (4) the choice to offset the benefit from real estate not used to make citric acid; (5) the agency’s estimation of benchmarks for steam coal, sulfuric acid, and calcium carbonate; and (6) certain adjustments made to the benchmarks for international freight costs. See generally Mem. of Law in Supp. of Pis. RZBC’s Mot. for J. on Agency R., ECF No. 44 (“RZBC Br.”); Mem. of Law in Supp. of Pl.-Intervenor’s Rule 56.2 Mot. for J. on Agency R., ECF No. 43 (“GOC Br.”).
The court has jurisdiction to hear these claims under 28 U.S.C. § 1581(c). The court must also uphold the agency’s results unless they are “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(l)(B)(i).
*1294 DISCUSSION
In view of these standards, the court must return one matter to Commerce: the calculation of world benchmarks for the steam coal, sulfuric acid, and calcium carbonate subsidies. The Plaintiffs’ and Plaintiff-Intervenor’s other claims lack merit, so the court sustains the agency’s reasoning in all other respects.
I. Commerce Properly Investigated the Calcium Carbonate Subsidy
RZBC’s first claim concerns a new subsidy petition filed during the third review period. The petition, lodged by U.S. industry, alleged that RZBC received subsidized calcium carbonate as an input into its citric acid. Commerce accepted the petition and later found a countervailable subsidy, but Plaintiffs argue this was a mistake, because RZBC never used a type of calcium carbonate mentioned in the petition. See RZBC Br. 21-24.
The court finds no flaw in the subsidy allegation or in the decision to investigate.
A. Background
The chemical at issue here, calcium carbonate (CaC03), is a versatile commercial compound. It comes in two main forms: ground calcium carbonate and precipitated calcium carbonate. Ground calcium carbonate (“GCC”) is made of crushed limestone, and it’s often used in industrial and pharmaceutical applications. When employed as a purifying agent, GCC may be called limestone flux. Precipitated calcium carbonate (“PCC”) bears the same chemical signature as GCC (CaC03), but it is created by a chemical process that yields a fine-textured powder. PCC is especially useful in the production of paper. See Pet’rs’ New Subsidy Allegations (“NSA”) at Ex. 30, PD 44-18 (Sept. 26, 2012); see also The Condensed Chemical Dictionary 616 (Gessner G. Hawley, ed., 10th ed. 1981) (giving makeup and uses of limestone).
In September 2012, domestic industry alleged that Chinese state-owned enterprises (“SOEs”) sold calcium carbonate to RZBC for less-than-adequate remuneration. The new subsidy petition was agnostic regarding the type of calcium carbonate Plaintiffs purchased, though. In the main body of the allegations, the petitioners referred to calcium carbonate generally, never saying which form of the chemical RZBC used to make citric acid. NSA at 27-35. The exhibits to the petition were equally unenlightening. One document explained the uses and varieties of calcium carbonate, but it did not specify which variety the Plaintiffs consumed. Id. at Ex. 30. Another set of exhibits described three calcium-carbonate-producing SOEs, yet one of those companies produced both PCC and GCC. Id. at Exs. 38-41. If anything, this implied that Plaintiffs could use either type of calcium carbonate to make its goods.
The petitioners came close to choosing between PCC and GCC in an exhibit estimating the benefit of the alleged subsidy. To calculate benefit, petitioners created a world benchmark for calcium carbonate using data from Harmonized Tariff Schedule (“HTS”) 2836.50. Id. at Ex. 42. HTS 2836.50 covers calcium carbonate including PCC, but it excludes ground natural limestone, or GCC. RZBC Comments on Pet’rs’ New Subsidy Allegations Benchmarks at Attach. 1, PD 114 (Mar. 25, 2013). By cobbling the benchmark from this information, the petitioners hinted, indirectly, that RZBC used PCC to make citric acid.
Commerce began investigating the alleged subsidy in January 2013. See Decision Mem. on New Subsidy Allegations (“NSA Decision Mem.”) at 14-16, PD 77 (Jan. 25, 2013). Soon thereafter, RZBC urged Commerce to drop the inquiry because it “used limestone flux (powder) *1295rather than pure or precipitated calcium carbonate” in production — in short, they thought the agency was investigating the wrong subsidy. RZBC’s New Subsidy Questionnaire Resp. at 11, CD 44-50 (Mar. 1, 2013). Commerce continued anyway, and in later submissions, both respondents and petitioners furnished data from HTS 2521.10 to make calcium carbonate benchmarks. See RZBC’s New Subsidy Allegations Benchmarks (“Pis.’ NSA Benchmarks”) at Attach. 1, PD 113 (Mar. 18, 2013); Pet’rs’ Benchmark Pricing for New Subsidy Allegations (“Pet’rs’ NSA Benchmarks”) at Ex. 4, PD 111-12 (Mar. 18, 2013). HTS 2521.10 covers GCC or limestone in powdered form.
At the close of the investigation, the agency held that RZBC had received coun-tervailable calcium carbonate subsidies. It measured the benefit using information from HTS Chapter 25, which covers GCC. See I & D Mem. at 25-26.
B. Discussion
RZBC argues that it was wrong to investigate the calcium carbonate subsidy, because the petition misidentified the type of calcium carbonate it used to make citric acid. RZBC Br. 21-24. But Plaintiffs read the law and the allegations too narrowly.
To trigger an investigation, petitioners must allege a subsidy as defined by statute. They also have to pad their allegations with any “information reasonably available” to them at the time of filing. 19 U.S.C. § 1671a(b)(l). If a petition meets these requirements, Commerce must proceed. The agency cannot refuse to investigate based on conjecture that the subsidy does not exist. This means most subsidy petitions are granted unless the allegations “are clearly frivolous, not reasonably supported by the facts alleged or ... omit important facts which are reasonably available to the petitioner.” H.R.Rep. No. 96-317, at 51 (1979).
These easygoing standards gave Commerce little choice. In the petition, domestic industry alleged all the elements necessary to prove a calcium carbonate subsidy. They started by claiming that government authorities furnished the input. Then they alleged that the respondent received a financial contribution from those authorities, rendering a benefit. The petitioners rounded out their allegations by discussing specificity. See NSA at 27-35. And to support each of these elements, petitioners supplied evidence that was available to them. This included, among other things, the profiles of state-owned companies producing PCC and GCC, see id. at Exs. 38-41, and documents listing chemicals manufacturing as a state-protected industry, see id. at Exs. 31, 33, 35. So in all material respects, the petition complied with the law. See 19 U.S.C. § 1671a(b)(l) (listing petition requirements). Commerce had to begin investigating, even if the precise contours of the subsidy were still unknown.
With a touch of irony, Plaintiffs counter that the petition was too specific to trigger an investigation. Though the text of the petition alleged a general calcium carbonate subsidy, one of the exhibits estimated benefit using export data for PCC. See NSA at Ex. 42. RZBC argues that this exhibit narrowed the scope of the petition from calcium carbonate in both its forms (including GCC and PCC) to PCC only. And because it used GCC to produce citric acid, Plaintiffs say the petition was ill-suited to propose an investigation. RZBC Br. 21-24.
The court cannot subscribe to this cherry-picked argument. While it’s true that the benefit estimate drew data from HTS 2836.50 — the subheading for PCC — other exhibits discussed calcium carbonate in both of its commercial guises. See, e.g., NSA at Exs. 30 (addressing industrial uses *1296of GCC and PCC), 38 (naming SOE Jianxi Taihua as a manufacturer of both GCC and PCC). None of the exhibits said which form of the chemical was used to make citric acid. In light of these ambiguities, it would be unfair to read the petition as an attack on PCC alone. The petition directed its allegations at calcium carbonate generally, and the provisional benefit calculations did not change that fact.
It may help, in closing, to recall the purpose of a petition. Just as a complaint proposes a lawsuit, a CVD petition proposes an investigation. See Duferco Steel, Inc. v. United States, 296 F.3d 1087, 1096 (Fed.Cir.2002) (discussing petition in context of antidumping scope decision). At the outset, when a petition is drafted, domestic industry may lack the data it needs to make firm factual allegations. Hence a degree of imprecision is acceptable, so long as the parties back their claims with information reasonably available to them. See 19 U.S.C. § 1671a(b)(l); S.Rep. No. 96-249, at 47 (stating Commerce must investigate unless it is “convinced that the petition and supporting information fail to state a claim upon which relief can be granted”); H.R.Rep. No. 96-317, at 51 (“The Committee views the rigor of the requirements of this threshold test as roughly analogous to the rigor of the requirements necessary to make out a cause of action for purposes of civil litigation.”). The goal is to direct Commerce to investigate the right subsidy.
The petition here accomplished this purpose without question. Though it skirted whether respondent used GCC or PCC, the petition ultimately led Commerce to examine the correct input: limestone flux. RZBC also concedes that the agency used the right tariff subheading to finalize the benefit calculation. See RZBC Br. 5-6, 23. So whatever the petition’s flaws, Commerce made no mistake when it inspected the calcium carbonate (limestone flux) subsidies. The decision to investigate was in accord with law and planted in substantial evidence.
II. Commerce Correctly Held that the Calcium Carbonate Subsidy Was Specific
The next claim — this one brought by the GOC — challenges the finding that the calcium carbonate subsidy was specific. Commerce applied adverse facts available (“AFA”) to reach this result, but the GOC rejoins that it complied fully with the agency’s requests for information. See GOC Br. 12-18.
The court rules for Commerce. Though it claims to have answered the agency’s questions to the best of its ability, the GOC could have done more to discover the amount of calcium carbonate used and purchased by Chinese industry during the review period. AFA was a just and reasonable response to the GOC’s intransigence.
A. Background
After confirming the existence of the calcium carbonate subsidy, Commerce had to decide if the subsidy was specific. By statute, a domestic subsidy is specific if “[t]he actual recipients of the subsidy ... are limited in number,” if “[a]n enterprise or industry is a predominant user of the subsidy,” or if “[a]n enterprise or industry receives a disproportionately large amount of the subsidy.” 19 U.S.C. § 1677(5A)(D)(iii)(I)-(III). These criteria seem to demand a mathematical approach. In general terms, they make specificity a function of the amount of subsidy taken by the subject industry or enterprise, divided by the total subsidy given to recipients as a whole. See, e.g., Samsung Elecs. Co. v. United States, 38 CIT -, -, 37 F.Supp.3d 1320, 1324-26 (2014) (finding Commerce reasonably analyzed specificity *1297by measuring plaintiffs benefit relative to other companies).
In that spirit, Commerce asked the GOC for hard statistical data to decide if the calcium carbonate subsidy was specific. In the new subsidy questionnaire, the agency requested a list of all the industries in the PRC that purchased calcium carbonate. Then Commerce asked for more granular details, including:
the amounts (volume and value) purchased by the industry in which the respondent companies operate, as well as the totals purchased by every other industry. In identifying the industries, please use whatever resource or classification scheme your government normally relies upon to define industries and to classify companies within an industry.
GOC New Subsidy Allegation Questionnaire Resp. (“GOC NSA Resp.”) at 28, CD 54 (Mar. 1, 2013). The purchase data, broken out by volume and value, would reveal whether the chemicals industry took a disproportionate share of the subsidy.
The GOC filed a response, but the information provided was meager. Rather than furnishing purchase data by volume and value as requested, Plaintiff-Intervenor remarked that calcium carbonate was “among the most important and versatile materials used by industry.” See id. at 24. The documents affixed to the GOC’s response were also general; they noted a growing demand for calcium carbonate in China, and they spotlighted the chemical’s diverse forms and uses, but they offered none of the mathematical rigor that Commerce wanted. Id. at Ex. 3. The GOC closed by adding that its State Statistical Bureau did “not collect any calcium carbonate data based on sales, let alone based on sales volumes by industrial sectors.” Id. at 24.
Unimpressed by the GOC’s answers, Commerce drafted another questionnaire on the specificity issue. This time, instead of asking for a list of industries that purchased calcium carbonate, Commerce requested a list of industries that used or consumed calcium carbonate. Like before, the GOC was to collate the data by volume and value. GOC Second Supplemental Questionnaire Resp. (“GOC 2d Supp. Resp.”) at 6, CD 72 (May 3, 2013).
Plaintiff-Intervenor rehashed its earlier answer in the second response. After claiming that it did not “maintain statistics on either calcium carbonate production or consumption,” the GOC urged Commerce to use data from the exhibits of the original questionnaire response to decide specificity. Id. It also invoked 19 C.F.R. § 351.301(c)(2)(iv) (2012), a regulation that lets respondents request the agency’s help when they struggle to submit information in the form and manner required. Apparently, the GOC hoped Commerce would waive its demand for statistics and rely on the general commercial tracts it previously rejected.
The agency dismissed the GOC’s pleas in the post-preliminary results. Instead of using Plaintiff-Intervenor’s exhibits to decide specificity, Commerce held that the GOC’s answers were unresponsive. Commerce also found that the GOC had not worked to the best of its ability to provide data. As a consequence, Commerce invoked AFA and held that the calcium carbonate subsidy was specific, notwithstanding the lack of evidence on the matter. Post-Prelim. Results Decision Mem. (“Post-Prelim. Mem.”) at 6-7, CD 110 (Nov. 8, 2013). The agency ratified its reasoning in the Final Results. I & D Mem. 47-50.
B. Discussion
The GOC now claims that it was wrong to settle the specificity question using AFA. Yet the court finds nothing amiss in the agency’s approach. Plaintiff-Interve-nor squandered multiple opportunities to *1298give statistical data (so Commerce could use facts available), and the record shows that the GOC spent less-than-best efforts to collect information (so Commerce could draw an adverse inference). The court explains its conclusions in more detail below.
1. Commerce Reasonably Used Facts Available to Decide Specificity
At the outset, the GOC argues that it lacked a fair chance to put passable specificity data on the record. Had it received all the opportunities to submit evidence that were its due under the statute, then Plaintiff-Intervenor might have produced compliant information, and Commerce might have held that the calcium carbonate subsidy was not specific. See GOC Br. 12-18.
But these arguments misconstrue the law and the help it affords struggling parties. In 19 U.S.C. § 1077e(a), Congress gave Commerce the power to complete trade remedy cases even when necessary information is not available, or where parties withhold evidence from the authorities. See Uruguay Round Agreements Act, Statement of Administrative Action, H.R. Doc. No. 103-316, voj. 1, at 869-70 (1994), reprinted in 1994 U.S.C.C.A.N. 4040, 4198-99 (“SAA”). When key data are missing from the record, the law allows the agency to use facts otherwise available to reach a decision. In other words, Commerce can take proof from the far reaches of the record to close eviden-tiary gaps that the parties never filled.
This power comes with caveats, however. Before Commerce can use facts available, it must give the parties a chance to comply with the agency’s requests for evidence. The law provides that if a party’s submission is deficient, the agency “shall promptly inform the person submitting the response of the nature of the deficiency and shall, to the extent practicable, provide that person with an opportunity to remedy or explain the deficiency in light of [relevant] time limits.” 19 U.S.C. § 1677m(d). Then, if the party’s supplementary response falls short, Commerce may “disregard all or part of the original and subsequent responses,” id., unless the data are otherwise fit for consideration, see id. § 1677m(e).
The statute also makes Commerce assist those who are unable to submit data in the agency’s preferred form and manner. If a party explains why it cannot give the information in the form requested, if it suggests alternative ways to package the data, and if it notifies the agency of its plight within fourteen days of receiving the questionnaire, then Commerce must “consider the ability of the interested party to submit the information in the requested form and manner and may modify, such requirements to the extent necessary to avoid imposing an unreasonable burden on that party.” Id. § 1677m(c)(l); see also 19 C.F.R. § 351.301(e)(2)(iv) (2012). The idea is to help respondents who face technical barriers to filing their answers. The provision does not excuse parties from submitting data altogether. See SAA at 865 (“Section 782(c)(1) is intended to alleviate some of the difficulties encountered by small firms and firms in developing countries, particularly with regard to the submission of data in computerized form. It is not intended to exempt small firms from the requirements of the ... countervailing duty laws.”).
Here, Commerce justifiably used facts available to decide specificity. On two occasions, the agency asked the GOC for specific statistical data regarding the use and purchases of calcium carbonate by Chinese industry. And on two occasions, Plaintiff-Intervenor responded that it kept no relevant statistics. Instead, the GOC supplied information devoid of meaningful analytical content, something more akin to *1299an infomercial than to focused market research. See GOC NSA Resp. at 24 & Ex. 3; GOC 2d Supp. Resp. at 6. Becausé the specificity inquiry demanded some mathematical rigor — and because the GOC gave no data to populate the analysis — Commerce reasonably invoked facts available to decide the matter. See 19 U.S.C. § 1677e(a).
The GOC counters that the agency never alerted it to deficiencies in its questionnaire responses. GOC Br. 18. It is right' in one sense. Commerce never expressly stated, “This information is useless,” or “Please try again,” in its letters to the GOC. That said, the agency’s communications left the clear impression that Plaintiff-Intervenor’s data fell short. In its first new subsidy questionnaire, Commerce asked the GOC for the volume and value of calcium carbonate purchases by industry. The GOC replied with some articles of general interest, but it added that it kept none of the data Commerce wanted. GOC NSA Resp. at 23-24. The agency then circulated another questionnaire ' craving statistics on calcium carbonate use by volume and value. GOC 2d Supp. Resp. at 6. The second questionnaire made plain that the answers to the first questionnaire were deficient: Had the first response passed muster, Commerce would not have demanded similar data in its second query. So despite its protestations, ¡the GOC had notice under § 1677m(d) that its first answers were flawed. See Ta Chen Stainless Steel Pipe, Inc. v. United, States, 298 F.3d 1330, 1337-38 (Fed.Cir.2002) (holding agency need not give formal notice of deficiency where “party informs Commerce that it will not provide the information requested”). When the GOC recycled the same evidence in its second response, Commerce reasonably disregarded it and relied on facts available.
The GOC also complains that it never got the help it demanded under § 1677m(c)(l). - This argument fails for two reasons. First, the pleas for assistance were not made on time. To secure the agency’s help, parties have to tell Commerce of their hardship “within 14 days after the date of receipt of the initial questionnaire.” 19 C.F'.R. § 351.301(c)(2)(iv) (2012). But the GOC requested assistance in the second supplemental response, which was filed months after the first questionnaire. See New Subsidy Allegation Questionnaire, PD 78 (Jan. 25, 2013); GOC 2d Supp. Resp. (May 3, 2013). The agency had no duty to accept this late-coming petition.'
Second, the GOC never offered compliant data in an alternative form. See 19 U.S.C. § 1677m(c)(l). None of the materials that Plaintiff-Intervenor filed— whether in the initial or supplemental response — came close to equipping the statistical data that the agency needed. Instead, the GOC wrote bluntly that its State Statistical Bureau did “not collect any calcium carbonate data.” GOC NSA Resp. at 24. It made no additional effort to pull statistics from alternative sources or to compile data in another form. The agency did not have to help a party that was so unwilling to help itself. See China Steel Corp. v. United States, 27 CIT 715, 732-33, 264 F.Supp.2d 1339, 1357-58 (2003) (finding criteria to 'trigger assistance not met where plaintiff failed to “suggest alternatives for providing the [requested] information”).
2. Commerce Reasonably Drew an Adverse Inference
Next, after invoking facts available to decide specificity, Commerce inferred that the subsidies were specific to’the chemicals industry. The inference selected was adverse to the GOC’s interests; Commerce explained that it took this tack because the GOC had not spent best efforts to supply information. I & D Mem. 47-5Q. The *1300GOC now argues that it did its utmost to gather data about calcium carbonate purchases and use. Plaintiff-Intervenor also argues that the decision to draw an adverse inference was unsupported by evidence. See GOC Br. 12-17.
These arguments do not persuade. The law provides that if a party fails to answer questionnaires to the best of its ability, Commerce may draw an adverse inference when choosing among facts available. 19 U.S.C. § 1677e(b). The statute does not define what it means to act “to the best of [one’s] ability,” but precedent sheds some light. Before the agency makes an adverse inference, the Federal Circuit requires Commerce to show (1) that a “reasonable [party] would have known that the requested information was required to be kept,” and (2) “that the failure to fully respond is the result of the •respondent’s lack of cooperation” in keeping proper records or in obtaining data. Nippon Steel Corp. v. United States, 337 F.3d 1373, 1382-83 (Fed.Cir.2003).
The evidence supports Commerce’s decision here on both fronts. For starters, the agency made plain in prior reviews of the citric acid- order that it needed statistics to decide specificity. In the first review, for instance, Commerce asked for data regarding steam coal use. See Archer Daniels Midland Co. v. United States (“Archer II”), 38 CIT -, -, 968 F.Supp.2d 1269, 1272-73 (2014); Citric Acid and Certain Citrate Salts from the People’s Republic of China, 76 Fed.Reg. 77,206 (Dep’t Commerce Dec. 12, 2011) (final admin, review) and accompanying I & D Mem. at cmt. 6. When the GOC answered that it kept no relevant statistics, Commerce postponed ruling on the issue. But the court later reversed the decision not to rule. See Archer Daniels Midland Co. v. United States, 37 CIT -, -, 917 F.Supp.2d 1331, 1339-40 (2013). Then on remand, Commerce revisited the issue and found that the steam coal subsidies were not specific — yet it only did so because it lacked the evidence it needed to go the other way. Archer II, 38 CIT at-, 968 F.Supp.2d at 1272. It would be unreasonable to infer from this back-and-forth that the GOC was exempt from keeping data on subsidy use. The agency clearly wanted it.
This point was hammered home in the second review. There, Commerce again requested data regarding steam coal use. When the GOC failed to give the information requested, Commerce determined that the steam coal subsidy was specific. Citric Acid and Certain Citrate Salts from the People’s Republic of China, II Fed.Reg. 72,323 (Dep’t Commerce Dec. 5, 2012) (final admin, review) and accompanying I & D Mem. at cmt. 4. These back-to-back reviews put the GOC on notice that it should collect subsidy consumption data going-forward.1
The GOC also shirked its duty to keep statistics that it knew it had to collect. See Nippon Steel, 337 F.3d at 1382-83. There were a number of groups besides the State Statistical Bureau that gathered data on the chemicals business, but the *1301GOC never consulted them. For instance, the record shows that the state-affiliated China Inorganic Salts Industry Association (“CISA”) compiled statistics on inorganic salts manufacturing. GOC 2d Supp. Resp. at Ex. 1. Had the GOC tried its best, it might have asked CISA for data on the industry’s calcium carbonate use. The GOC might also have taken data from calcium-carbonate-producing SOEs. See NSA at Exs. 38-41. These companies likely recorded their calcium carbonate sales (if they were properly managed, of course), and Commerce might have used this information to decide specificity. The GOC had no excuse for failing to explore these untapped resources.
The decision to use AFA was grounded in substantial evidence and accorded with law.
III.. Commerce Properly Countervailed Inputs Purchased from Trading Companies
The next claim highlights an important facet of the definition of a subsidy. During the review, Commerce countervailed certain inputs that RZBC purchased from private trading companies. But the statute requires that a subsidy’s benefit result from a foreign government’s financial contribution. See 19 U.S.C. § 1677(5)(B) (“A subsidy [occurs when] an authority provides a financial contribution ... to a person and a benefit is thereby conferred.” (emphasis added)). In the GOC’s view, the decision to countervail the inputs was contrary to law, because Commerce never established a “causation connection” between the government’s contribution to the traders and the benefit received by RZBC. See GOC Br. 20. The court disagrees.
A. Background
During a prior review, RZBC reported that it purchased steam coal and some of its sulfuric acid from private trading companies. The trading companies bought those inputs from state authorities. See Confidential App. to Pl.-Intervenor’s Rule 56.2 Mot. for J. upon Agency R. at Ex. 7, DE 48 (“GOC Conf. App.”).2 The situation remained the same during the third review period. See GOC Br. 19.
In its preliminary results, Commerce decided to countervail the sulfuric acid and steam coal purchases. It did so even though RZBC bought some of the inputs from private traders. The GOC argued that this was improper, because Commerce had not established “a definitive causal connection between the government action and the benefit bestowed.” I & D Mem. 50.
The agency rejected the GOC’s argument in the Final Results. It reasoned that when the government sold subsidized goods to trading companies, the traders passed some of the benefit to producers who bought the inputs for below-market prices. Hence there was no need to prove that the government directed trading companies to dispose of their wares for less-than-adequate remuneration. The traders could sell for low prices, and the producers could reap a benefit, even without the threat of coercion. See id. at 51-52.
B. Discussion
The logic in the Final Results is spot on. As mentioned before, the law defines a subsidy as a financial contribution from a government authority to a person, imparting a benefit thereby. See 19 U.S.C. *1302§ 1677(5)(B). But the statute does not say-how the contribution must be conferred to count as a subsidy. In fact, the law leaves that matter wide open. The statute reads, in relevant part, “The determination of whether a subsidy exists shall be made without regard ... to whether the subsidy is provided directly or indirectly on the manufacture, production, or export of merchandise.” Id. § 1677(5)(C).
This provision gives Commerce a broad range of options. Naturally, it means Commerce can countervail contributions passed straight from a foreign government to a producer — a direct subsidy. See id. § 1677(5)(B)(i). It also means that the agency can countervail contributions from private parties acting at the direction of a public body — an indirect subsidy springing from a private source. Id. § 1677(5)(B)(iii). And more important for our purposes, the provision allows Commerce to counteract contributions which originate with the government, but which pass through an intermediary before going to the ultimate user. In the last variation, the middleman may skim some of the benefit by reselling the subsidized inputs at a markup. But if the marked-up cost that the final user pays is less than the market cost, the user still gets a benefit. Commerce can countervail these sales because the authorities made a contribution to a person (the intermediary), thereby conferring a benefit to the producer (in a sum reduced by the middleman’s cut). See id. § 1677(5)(B); Beijing Tiankai Indus. Co. v. United States, 39 CIT -, -, 52 F.Supp.3d 1351, 1360-67 (2015) (holding subsidy exists even if contribution and benefit received by different people); Guangdong Wireking Housewares & Hardware Co. v. United States, 37 CIT-, -, 900 F.Supp.2d 1362, 1379-80 (2013) (holding subsidy passed through private trading companies was countervailable); SAA at 926 (“The Administration plans to continue its policy of not permitting the indirect provision of a subsidy to become a loophole when unfairly traded imports enter the United States and injure a U.S. industry.”).
The disputed steam coal and sulfuric acid subsidies fall in this final bucket. The record reflects that Chinese authorities sold inputs to private trading companies, which then resold the goods to RZBC. See GOC Conf. App. at Ex. 7. The record also indicates that RZBC bought steam coal and sulfuric acid from the traders for less-than-adequate remuneration, though the amount of benefit remains in play. I & D Mem. 20-24; see also infra Part V. Under the circumstances, it was reasonable for Commerce to infer that RZBC’s discount stemmed from the government’s contribution to the traders. The trading companies likely resold the inputs for more than they paid, of course, but the difference between the resale price and the market price was a benefit to RZBC. The agency could countervail this amount, because the government’s financial contribution allowed the traders to sell sulfuric acid and steam coal for less-than-market price. See 19 U.S.C. § 1677(5)(B).
In response, the GOC hints that RZBC’s benefit was caused by something other than government largesse. “It is certainly plausible,” they posit, “that the trading companies purchased the inputs from the government authorities at higher prices and subsequently resold them to RZBC at lower prices due to prevailing market conditions or for some other reason.” GOC Br. 22. If this were the case, then perhaps the benefit would not stem from an official contribution as the law requires. But the GOC marshals no facts to support its claim. It cites no data to prove that the purchase price from the government exceeded the sale price to RZBC. See id. And it conjures no evidence to show that the trading companies kept all the benefit *1303they got from the state. See Delverde, 202 F.3d at 1368 (reversing CVD decision where record indicated that subsidy recipient passed no benefit to successor in interest of pasta factory). Without proof to the contrary, the court must sustain the finding that the Chinese government’s contribution benefitted RZBC. See Yangzhou Bestpak Gifts & Crafts Co. v. United States, 716 F.3d 1370, 1377 (Fed.Cir.2013) (“Commerce’s determination will be sustained unless it is ‘unsupported by substantial evidence on the record, or otherwise not in accordance with law.’ ” (quoting 19 U.S.C. § 1516a(b)(l)(B)(l))). We will not upend the agency’s reasonable inference based on speculation that the traders sold their goods for a loss.
The GOC also suggests that Commerce could not countervail the subsidies unless the authorities ordered the traders to sell for below-market prices. GOC Br. 21-23. Yet the cases cited for this proposition address an irrelevant issue. In AK Steel Corp. v. United States, 192 F.3d 1367, 1370 (Fed.Cir.1999), and Hynix Semiconductor Inc. v. United States, 29 CIT 995, 997-99, 391 F.Supp.2d 1337, 1340-41 (2005), Commerce countervailed preferential loans that originated with private banks. To decide if the loans sprang from an authority, the courts had to resolve whether the government ordered the banks to lend money. See AK Steel, 192 F.3d at 1376; Hynix, 29 CIT at 1000-01, 391 F.Supp.2d at 1343-44. Here, by contrast, no one disputes that the steam coal and sulfuric acid inputs came from an authority. Though the government passed its contribution to the traders first, RZBC reaped a benefit to the extent that it bought the inputs for less-than-adequate remuneration. So again, the purchases met all the criteria of a subsidy as defined in statute. See 19 U.S.C. § 1677(5)(B); Beijing Tianhai, 39 CIT at -, 52 F.Supp.3d at 1360-67. There was no need to show that the GOC directed the sales.
The decision to countervail the steam coal and sulfuric acid subsidies was based in substantial evidence and accorded with law.
IV. Commerce Reasonably Countervailed Land-Use Subsidies
RZBC brings the next claim, which concerns land-use subsidies. During the review period, Plaintiffs purchased the rights to three plots of land for below market value. RZBC did not use the land to manufacture citric acid, but Commerce countervailed the subsidies anyway. The Plaintiffs now claim this was error. See RZBC Br. 42-43. The court finds none.
A. Background
At the dawn of the third review, the petitioners alleged that RZBC bought subsidized land-use rights from provincial and local government authorities. NSA at 3-10. The agency decided to investigate shortly thereafter. See NSA Decision Mem. at 2-5.
In their questionnaire responses, Plaintiffs reported that one of their companies, RZBC Juxian, bought fifty-year rights to a plot of land in November 2010. Another company, RZBC Co., bought rights to two more parcels in September 2011. Plaintiffs said the Juxian plot was not used to make citric acid. And the other two plots owned by RZBC Co. were vacant during the review period and slated to be rezoned for commercial use. Resp. to New Subsidy Questionnaire Section A (“RZBC NSA Section A Resp.”) at 2-5, CD 55-64 (Mar. 8, 2013).
The agency chose to countervail all three subsidies over Plaintiffs’ protest. Though none of the plots were ever used to make subject merchandise, Commerce argued that the statute permitted it to counter subsidies that did not affect production. See I & D Mem. 65.
*1304B. Discussion
The agency is right again. To qualify as a subsidy, a government contribution must confer a benefit. 19 U.S.C. § 1677(5)(B). That said, the agency is not required to “to consider the effect of the subsidy in determining whether a subsidy exists.” Id. § 1677(6)(C); see also 19 C.F.R. § 351.503(c) (“In determining whether a benefit is conferred, the Secretary is not required to consider the effect of the government action on the firm’s performance, including its prices or output, or how the firm’s behavior otherwise is altered.”). A subsidy is a subsidy even if the government action has no influence “on the price or output of the class or kind of merchandise under investigation or review.” SAA at 926.
In that spirit, Commerce acted lawfully to countervail the land-use subsidies. Even if RZBC never used its three plots to make citric acid, the law presumes that the subsidies rendered a benefit. The agency was not required to decide exactly how much the benefit reduced the price of the subject merchandise or altered RZBC’s production process. See 19 C.F.R. § 351.503(c). Furthermore, Plaintiffs offered no proof that the land-use subsidies were tied to a product other than citric acid. See id. § 351.525(b)(5)(i) (stating agency must attribute subsidy to particular product if subsidy tied to production or sale of that product). While RZBC mentioned a desire to rezone one of its plots for commercial use, that parcel was not earmarked for such a purpose during the review period. See RZBC NSA Section A Resp. at 4-5. Accordingly, Commerce could countervail the land-use subsidies against citric acid, without deciding the price effect on the subject goods, and without limiting the benefit to other products.3
V. Commerce Did Not Properly Calculate World Benchmark Prices
In the next round of claims, the Plaintiffs focus their fire on benefit — specifically, the methods and data Commerce used to make price, benchmarks for the steam coal, sulfuric acid, and calcium carbonate subsidies. They argue, among other things, that it was unlawful to average prices to create benchmarks; that it was error to exclude information from countries that shipped to China; and that the combination of simple averages and aberrant price data yielded unrealistic benchmarks. See GOC Br. 28-33; RZBC Br. 24-35.
Of these arguments, only the last merits a remand.
A. Background
The mathematical theory driving the benefit calculation sounds simple, but the devil is in the details. In general, when goods are provided, Commerce derives benefit by subtracting the price paid for the subsidized item from the item’s price on the open market. See 19 U.S.C. § 1677(5)(E)(iv). The agency must determine “the adequacy of remuneration ... in relation to prevailing market conditions for the good or service being provided ... in the country which is subject to the investigation or review.” Id.
The real complications arise before the subtraction, when Commerce estimates the *1305subsidized item’s price on the open market. In ordinary cases, Commerce lifts this number from “actual transactions in the county in question.” 19 C.F.R. § 351.511(a)(2)(i) (2015). If there are no useable market prices for the input, however, then Commerce moves to its second, or “tier-two” methodology. Under the tier-two method, Commerce adopts the world market price as a benchmark or proxy for the price in the producer’s home country. The agency may use the world price only if it is “reasonable to conclude that such price would be available to purchasers in the country in question.” Id. § 351.511(a)(2)(ii). Furthermore, if the record boasts more than one viable world price, Commerce must average “such prices to the extent practicable, making due allowance for factors affecting comparability.” Id. These factors ensure that the composite benchmark reflects prevailing market conditions in the home country. They include “price, quality, availability, marketability, transportation, and other conditions of purchase or sale.” 19 U.S.C. § 1677(5)(E).
In this case, Commerce invoked its tier-two method to construct benchmarks for the steam coal, sulfuric acid, and calcium carbonate subsidies. I & D Mem. at 21-26. It began by requesting data for each input examined, and the parties responded with export prices from a smattering of countries. For sulfuric acid and calcium carbonate, the data provided came almost exclusively from the Global Trade Atlas (“GTA”), a reporter that lists each country’s exports by transaction with price and quantity terms. The transaction-level data were averaged to make country-level export prices by month (or by year for RZBC’s calcium carbonate data). See Pis.’ NSA Benchmarks at Attach. 1 (calcium carbonate); Pet’rs’ NSA Benchmarks at Ex. 4 (calcium carbonate); Pet’rs’ Factual Information Submission (“Pet’rs’ Factual Sub.”) at Tab 20, PD 64-70 (Nov. 20, 2012) (sulfuric acid); RZBC Factual Information Comments (“RZBC Factual Cmts.”) at Ex. 2-C, PD 56-62 (Nov. 19, 2012) (sulfuric acid). The parties also furnished GTA data for steam coal, but to this, the petitioners added information from the International Monetary Fund (“IMF”) and the Platts International Coal Report (“Platts”). See RZBC Factual Sub. at Ex. 1 — B; Pet’rs’ Factual Sub. at Tab 17. The IMF and Platts listed steam coal export prices by country by month; they did not give transaction-specific , quantity information like GTA.
With this evidence in hand, Commerce set about assembling its benchmarks. For steam coal, Commerce first purged the dataset of country-level prices reflecting shipments to China. The agency reasoned that government intervention could distort the price of China-bound exports, making the data unreliable for the purpose of calculating benefit. Prelim. Results Calculation Mem. (“Prelim. Calcs.”) at 5-7 & n. 18, CD 95 (June 7, 2013). Commerce then simple averaged the remaining prices to form world average prices by month. The agency used simple averages instead of weighted averages because the IMF and Platts data lacked quantity terms, and it wanted to keep the IMF and Platts data in the mix to support a robust calculation. I & D Mem. at 87-88.
The agency followed a similar approach with sulfuric acid. After excluding prices from China-bound exports, Commerce simple averaged the prices that remained to form world averages by month. Id.; Prelim. Calcs, at 3-4. But here, the agency gave no reason why it simple averaged the data instead of weight averaging them. The sulfuric acid prices that remained all had quantity terms. See, e.g., Pet’rs’ Factual Sub. at Tab 20. Presumably the agency could have weight averaged the prices had it so chosen.
*1306Finally, for calcium carbonate, Commerce rejected RZBC’s information outright. Unlike the other submissions— which provided country-level prices by month — these data gave country-specific prices averaged over the entire review period. In Commerce’s view, the year-long averages would yield less precise benefit estimates than month-long averages. See I & D Mem. at 83-84; Post-Prelim. Mem. at 11 n. 60. The data also included exports to China. See Pis.’ NSA Benchmarks at Attach. 1. Hence the agency used only the petitioners’ prices and merged the data into worldwide simple averages (again without explaining why it eschewed the weight-averaging approach). I & D Mem. at 87-88.
Prior to the Final Results, RZBC and the GOC challenged aspects of the benchmark calculations. They alleged that the law barred Commerce from averaging country prices to render benchmarks; that Commerce made benchmarks using country-level prices that were not reasonably available in China; that small export quantities made some of the data unreliable; and that Commerce should not have simple averaged country prices by month to create world benchmarks. Id. at 86-87; RZBC Case Br. at 12-30, CD 114 (Nov. 18, 2013). The agency rejected these arguments and others in the Final Results. See I & D Mem. at 74-89.
Later, RZBC lodged ministerial error comments. It noted that the datasets for calcium carbonate and sulfuric acid listed price and quantity values uniformly, meaning Commerce could have calculated benchmarks for those inputs using weighted instead of simple averages. RZBC Ministerial Error Comments (“Min. Error Cmts.”) at 1-3, PD 235 (Jan. 6, 2014). The agency again rejected the arguments, declaring that it had used simple averages deliberately. Ministerial Error Allegation Mem. at 2, PD 240 (Jan. 27, 2014).
B. Commerce’s Tier-Two Methodology Complies With Law
Now on appeal, the Plaintiffs and' Plaintiff-Intervenor muster many of the same claims made below. The GOC leads the charge with a legal challenge. By statute, the agency must decide the extent or “adequacy” of benefit “in relation to prevailing market conditions for the good or service being provided.” 19 U.S.C. § 1677(5)(E)(iv). In light of this law, Plaintiff-Intervenor would abolish the tier-two method in 19 C.F.R. § 351.511(a)(2)(ii) because it forms benchmarks from average prices. An “average” price is more than an “adequate” price, the GOC claims, because savvy producers always choose the lowest prices available to source their inputs. See GOC Br. 28-33.
The court does not question the Plaintiffs’ commitment to frugality. But it does take issue with the notion that the tier-two methodology violates the statute. As mentioned before, the law requires Commerce to measure benefit, or the adequacy of remuneration, “in relation to prevailing market conditions.” 19 U.S.C. § 1677(5)(E)(iv). In other words, remuneration for a subsidized input is “adequate” if it conforms to prevailing market norms. The statute lists the conditions the agency should consider to determine adequacy (price, quality, availability, etc.), but it does not explain what it means for conditions to “prevail” in the home market. Id.
Without guidance on the matter, it was reasonable for Commerce to equate “prevailing market conditions” with average market conditions. See Chevron, U.S.A., Inc. v. Nat’l Res. Def. Council, Inc., 467 U.S. 837, 843, 104 S.Ct. 2778, 81 L.Ed.2d 694 (1984) (“[I]f the statute is silent or ambiguous with respect to [a] specific issue, the question for the court is whether the agency’s answer is based on a permis*1307sible construction of the statute.”). A legal dictionary defines the verb “to prevail” as “to be commonly accepted or predominant.” Black’s Law Dictionary 1380 (10th ed. 2014). And when you take an average, all you are doing is finding the predominant or typical case within a sample. See id. at 162 (defining “average”). So when Commerce averages country-level prices under the tier-two method, it is doing just what the statute commands: It is finding the prevailing world price for the subsidized input and using that to measure the adequacy of remuneration. See 19 U.S.C. § 1677(5)(E)(iv). The court sees nothing unreasonable in this approach.
The GOC retorts that “adequate” prices are found “at the lower end of available market prices, not the middle.” GOC Br. 81. Yet this argument turns a blind eye to the statute’s text. The law defines adequate remuneration as an input price that aligns with “prevailing market conditions.” See 19 U.S.C. § 1677(5)(E)(iv). By definition, the lowest input values are always less than prevailing or average prices. So even if penny-pinching producers buy their inputs from the cheapest sellers, that does not mean Commerce must use the lowest input values to make its benchmarks. And contrary to the GOC’s claims, this court has never held otherwise. See Dorbest Ltd. v. United States, 30 CIT 1671, 1687-88, 462 F.Supp.2d 1262, 1278-79 (2006) (critiquing use of import data to create AD surrogate values when domestic data available, but omitting comment on whether agency must choose lowest values on record to price inputs); Hebei Metals & Minerals Imp. & Exp. Corp. v. United States, 28 CIT 1185, 1193-95, 2004 WL 1615597 (2004) (same); Yantai Oriental Juice Co. v. United States, 26 CIT 605, 616-18, 2002 WL 1347018 (2002) (same).
The agency’s tier-two methodology accords with law.
C. Commerce Used Prices that Were Reasonably Available in China
RZBC builds its next claim on the regulation that the GOC tried to tear down. Under 19 C.F.R. § 351.511(a)(2)(ii), Commerce measures the adequacy of remuneration using world market prices “where it is reasonable to conclude that such price would be available to purchasers in the country in question.” RZBC argues that the values used to make benchmarks here did not represent prices available in China, for two main reasons. See RZBC Br. 29-35.
First, the data underlying the benchmarks excluded shipments to China. After taking information from the parties, Commerce purged the datasets of country-level prices reflecting exports to the PRC. The agency did so out of concern that intervention by the PRC government artificially deflated export prices. See, e.g., Prelim. Calcs, at 3 (excluding China-bound sulfuric acid export prices). RZBC alleges that by omitting these values, Commerce deleted the only prices on record that were reasonably available to purchasers in China. See RZBC Br. 30-35.
Second, the remaining data for steam coal, sulfuric acid, and calcium carbonate included regional export transactions. The calcium carbonate prices from Greece, for example, were based solely on shipments to Albania, a country adjacent to the exporter. Pet’rs’ NSA Benchmarks at Ex. 4. The same can be said for sulfuric acid (which included prices from Egypt, Colombia, and others to nearby countries) and steam coal (which also included prices from exporters to their neighbors). See Pet’rs’ Factual Sub. at Tabs 17, 20. Because these export prices stemmed from regional trades, RZBC argues that the data do not represent 'world market prices reasonably available in China. See RZBC Br. 30-35.
*1308Perhaps these twin claims have merit in the abstract. Sometimes it’s not sensible to use prices from one region to represent prices in another region; for example, you would not adopt European electricity prices to estimate Latin American electricity prices, because power cannot be shipped across the globe. See Countervailing Duties, 63 Fed.Reg. 65,348, 65,377 (Dep’t Commerce Nov. 25, 1998). Yet here, the inputs in dispute are commodities, or items that are routinely exported around the world. Because these inputs are traded so widely, the agency reasonably presumed that they have a similar price everywhere — regardless of whether they were shipped to China or to another country, and regardless of whether they were exchanged among regional partners. See id.; High Pressure Steel Cylinders from the People’s Republic of China, 77 Fed.Reg. 26,738 (Dep’t Commerce May 7, 2012) (final admin, determination) and accompanying I & D Mem. at cmt. 8 (using data from Iran, Italy, and Ukraine to evaluate steel tube prices in China). The Plaintiffs furnished no concrete evidence in rebuttal, at least with respect to sulfuric acid and calcium carbonate.
An exhibit about steam coal caused the court some concern, but not enough for a remand. In a supplemental submission, RZBC included an article explaining that the global coal trade is divided into Pacific and Atlantic spheres. RZBC Additional Factual Information Comments at Ex. 6, PD 71-74 (Nov. 21, 2012). The article seems to suggest that prices in each region differ, and that coal prices along the Atlantic Rim are not available in China. Even so, the exhibit did not mention whether prices are higher or lower in either region. Id. Because the evidence failed to prove how coal prices vary by region, Commerce reasonably relied on the prices available, exclusive of PRC-bound shipments, to make its benchmarks.
D. Simple Averages Distorted Commerce’s Benchmarks
All this does not mean the agency escapes unscathed. In its next claim, RZBC says the agency erred when it simple averaged country prices to create benchmarks. This approach gave undue weight to small-quantity, high-cost shipments, and in Plaintiffs’ estimation, the method returned benchmarks that were higher than prices in China. Remember, the law requires benchmarks to align with market conditions prevailing in the exporter’s home country. 19 U.S.C. § 1677(5)(E)(iv). So couched in terms of our standard of review, Plaintiffs argue that the benchmarks were unfounded in substantial evidence and contrary to law. See RZBC Br. 24-29.
RZBC’s arrow hits the target, and some background helps to explain why. Before this review, Commerce generally set world benchmarks by simple averaging country-level prices. See Utility Scale Wind Towers from the People’s Republic of China, 77 Fed.Reg. 75,978 (Dep’t Commerce Dec. 26, 2012) (final affirm, determination) {“Wind Towers ”) and accompanying I & D Mem. at cmt. 15 (citing Certain Steel Wheels from the People’s Republic of China, 77 Fed.Reg. 17,017 (Dep’t Commerce Mar. 23, 2012) (final affirm, determination) and accompanying I & D Mem. at cmt. 15). The agency began by gathering price data for exports of the good in question. Then it melded the data into an average, assigning each price an equal weight regardless of the quantity shipped. The method was particularly useful (and indeed the only averaging technique available) when parties provided export price data without listing the volume shipped at each price. See id.
Yet even in 2012, Commerce seemed to recognize that benchmarks made with simple averages were inferior those made with *1309weighted averages. See id. at cmt. 15 (explaining that agency lacked “information on the record that would allow [it] to weight-average the prices properly”). Like simple averaging, the weight-averaging method blends country-level prices into world benchmarks. But unlike simple averages, weighted averages assign each price a weight proportional to the quantity shipped at that price. What results are benchmarks that favor prices from large-quantity shipments. The method ensures that high prices from countries with low-volume exports do not skew the benchmarks upward. Commerce now prefers to use weighted averages when the parties report price and quantity in a uniform manner. See Steel Concrete Reinforcing Bar from the Republic of Turkey, 79 Fed. Reg. 54,963 (Dep’t Commerce Sept. 15, 2014) (final affirm, determination) (“Steel Rebar ”) and accompanying I & D Mem. at cmt. 1 & n. 128 (weight averaging prices after excluding data lacking quantity terms); Certain Oil Country Tubular Goods from the Republic of Turkey, 79 Fed.Reg. 41, 964 (Dep’t Commerce July 18, 2014) (final affirm, determination) and accompanying I & D Mem. at cmt. 4 (“Using weighted-average prices where possible reduces the potential distortionary effect of any specific transactions (e.g., extremely small transactions) in the data.”).
In this case, the agency used simple averages to build benchmarks for sulfuric acid and calcium carbonate. Yet Commerce never explained why it opted against weighted averages. See I & D Mem. at 87-88. To repeat, the datasets for sulfuric acid and calcium carbonate contained both price and quantity terms. Commerce could have used this information to draw weighted averages, and presumably, if the agency were incapable of taking weighted averages, it could have made that known during the administrative process. This lapse in explanation strikes the court as arbitrary. See Burlington Truck Lines, Inc. v. United States, 371 U.S. 156, 167, 83 S.Ct. 239, 9 L.Ed.2d 207 (1962) (invalidating decision where agency failed to state “the basis on which [it] exercised its expert discretion”); Nat’l Org. of Veterans’ Advocates, Inc. v. Sec’y of Veterans Affairs, 314 F.3d 1373, 1380-81 (Fed.Cir.2003) (remanding where agency failed to articulate rationale for legal interpretation). Commerce could not favor one method over another without saying why.4
The choice to take simple averages was not just arbitrary in the abstract, however. The method also caused real distortions in the benchmarks Commerce created. A simple average, unlike a weighted average, gives equal weight to all prices regardless of the quantities sold. High prices from small transactions can balloon the average to absurd proportions, and that seems to be what happened here. Take sulfuric acid, for example. The agency’s November world market price was $376.29 per metric ton (“MT”), about $124 more than the next highest world value for the year. Final Results Calculation Mem. (“Final Calcs.”) at Attach. 1, CD 120-21 (Dec. 27, 2013). And what might have inflated this *1310figure? The Indian export price of $5,397.99/MT stands out as a likely culprit. Id. This country-level average was based in part on one low-quantity load (34.125 MT) for a stratospheric sum ($2,984,697.00, or $87,463.12/MT). Pet’rs’ Factual Sub. at Tab 20. Yet despite the small volume of this underlying shipment, India’s country-level price was weighted equally with other prices to form a world benchmark. See I & D Mem. at 87-88. These data, together with others like them, combined to form simple-average benchmarks that outstripped what Commerce might have calculated had it weighted its averages. The table below demonstrates how much the simple-average sulfuric acid world prices exceeded their weighted-average counterparts.
Sulfuric Add World Average Price Comparison Chart for 2011 (USD/MT)5
Jan. ' Feb. Mai’. Apr. May June July Aug. Sept. Oct. Nov. Dec.
Simple 106.62 99.59 125.10 119.37 252.43 112.47 125.74 112.06 126.91 113.80 376.29 112.33
Weighted 67.42 80.51 84.04 84.14 85.14 84.55 97.50 90.18 100.62 91.06 95.90 87.70
We see the same pattern among the calcium carbonate benchmarks. In October 2011, for instance, the agency’s world price was $314.08/MT — lower than the May zenith of $622.82/MT, but higher than all the other world prices for the year. Final Calcs, at Attach. 1. The number was bloated by .the Australian country-level price, which was based on a small, 18 MT shipment for $3,207.67/MT. Pet’rs’ NSA Benchmarks at Ex. 4. By packing its simple averages with these and similar data. Commerce amplified the effect of high-price transactions and produced outsize benchmarks. The following table illustrates the difference between the simple-average and weighted-average world price estimates.
[[Image here]]
On remand, the agency must reconsider whether to make the sulfuric acid and calcium carbonate benchmarks with weighted averages. If it chooses not to, Commerce must give some logical reasons why simple averages yield more reliable benchmarks than weighted averages. The agency should keep in mind the effect of small-quantity, high-price transactions as it writes up its reasoning. The goal is to' make world benchmarks that reflect prevailing market conditions in the home country, as the statute commands. See 19 U.S.C. § 1677(5)(E)(iv).
Steam coal presents a different challenge, but Commerce should revisit these benchmarks on remand, too. As noted *1311earlier, the agency used country-level information from GTA, IMF, and Platts to estimate world prices. The GTA dataset boasted price and quantity terms, but IMF and Platts did not. Because the latter two sources were missing quantity information, Commerce blended the GTA, IMF, and Platts data using simple averages. Commerce thought the world prices would be more robust if they took statistics from all three reporters. I & D Mem. at 87-88.
Ordinarily this approach would cut the mustard. Commerce has been known to simple average prices wheré quantity terms were absent, for obvious mathematical reasons. See Wind Towers at 75,978 and accompanying I & D Mem. at cmt. 15. But here, the agency’s explanation fell short. When Commerce averaged the steam coal data, high prices from low-quantity shipments swelled the benchmarks considerably. A single 2 MT shipment of steam coal from Turkey cost $1,457/MT, for example; Plaintiffs estimate that this transaction raised the September world price from $198.84/MT to $241.69/MT. See RZBC Br. 37-38; Final Calcs, at Attach. 1; Pet’rs’ Factual Sub. at Tab 17. In view of this distortion, Commerce should have addressed the choice before it: Would the agency make better benchmarks using robust simple averages (including all three datasets) that gave undue weight to small shipments? Or was it better to use less robust weighted averages (excluding IMF and Platts) that gave proper weight to small shipments? Commerce dealt with similar questions in Steel Rebar, but here, the agency simply assumed that a more robust simple average was the right approach. See Steel Rebar at 54,963 and accompanying I & D Mem. at cmt. 1. It never compared the two averaging methods to decide which would better capture prevailing market conditions in China. The agency must reconsider its choice on remand and give an explanation backed with substantial evidence.6
RZBC pitches arguments other than those made above, but the court will not decide them now. For example, Plaintiffs claim that some of the benchmark data came from shipments for less than what RZBC consumed; that Commerce averaged prices from transactions with zero quantity; and that benchmarks averaged by month gave undue weight to high-price, low-volume shipments. See RZBC Br. 30-34, 38-41. Yet each of these arguments seeks to temper the effect of small-quantity exports on the agency’s simple averages. If Commerce reverses course and uses weighted averages on remand, many of these concerns could evaporate. And if the agency sticks to its original decision, Plaintiffs can resurrect their arguments in the remand comments.
YI. Commerce Properly Selected Freight Rates for Its Benchmarks
RZBC’s final arguments challenge adjustments the agency made to its benchmark values. Before calculating benefit, Commerce increased the calcium carbonate benchmarks by the sum an importer would pay to ship its inputs in flat-rack collapsible containers. Plaintiffs rejoin that this rate included irrelevant charges, and that the agency could not use rates from countries with no benchmark data. RZBC Br. 43-45. The court rejects both of these last-ditch appeals.
A. Background
When forging benchmarks, Commerce tries to estimate the amount producers *1312would pay if they imported the goods they received for a subsidy. As a first step, the agency calculates the price of the goods under market conditions. The court already discussed this process in detail. Then Commerce adjusts the benchmarks to “include delivery charges and import duties.” 19 C.F.R. § 351.511(a)(2)(iv). These charges include the cost of international shipping.
In this case, Commerce asked the parties to suggest international freight rates for calcium carbonate and other items. Only the petitioners responded for calcium carbonate. In their factual submission, they provided price quotes from the Maersk Line for 40-foot flat rack collapsible containers. Pet’rs’ Factual Sub. at Tab 23. The quotes were for shipments to China from a number of countries, including Canada, Argentina, Russia, the Netherlands, and the United States. The prices also included a special equipment service charge comprising a substantial chunk of the quote. (For example, to move a container from Los Angeles to Shanghai cost $6,604.30, and the special charge made up $5,175.00 of the total). Id. Commerce ultimately adopted the Maersk information, over objections that (1) there was no proof that importers paid special service charges to ship calcium carbonate, and (2) shipping prices from countries with no benchmark data did not represent the actual cost to import to China. I & D Mem. at 91-93.
B. Discussion
RZBC unfurls the same two grievances now, but neither one leaves the harbor. First, the record supports including the special service charge in the freight price. The evidence here was sparse, of course. Petitioners were the only ones to offer information, and the proof they provided did little more than list prices and note that flat rack collapsible containers carried “heavy cargo that requires special attention.” Pet’rs’ Factual Sub. at Tab 23. The evidence did not reveal how people usually ship limestone across the ocean. It also failed to describe when Maersk levies the special equipment service charge. See id. For lack of relevant data, Commerce reasonably relied on the flat rack rates cum service charges to estimate freight expenses. See Ames True Temper v. United States, 31 CIT 1303, 1312-13, 2007 WL 2460615 (2007) (sustaining agency’s choice to use only brokerage and handling surrogate on record). If the Plaintiffs wanted Commerce to use a different price — or if it wished to exclude the special service charge as an extraneous fee- — then it should have furnished some proof to that effect. See QVD Food Co. v. United States, 658 F.3d 1318, 1324 (Fed.Cir.2011) (quoting Tianjin Mach. Imp. & Exp. Corp. v. United States, 16 CIT 931, 936, 806 F.Supp. 1008, 1015 (1992)) (“[T]he burden of creating an adequate record lies with [interested parties] and not with Commerce.”).7
*1313Second, there is no evidence that freight costs from countries with benchmark data are any different than prices from countries without them. Cf. supra Part V.C. (holding exports did not need to go to China to represent Chinese prices). RZBC could have submitted such evidence had it been diligent, but it neglected to do so. See QVD, 658 F.3d at 1324. Absent proof to the contrary, Commerce reasonably averaged freight costs using prices from countries without calcium carbonate, sulfuric acid, and steam coal benchmarks. This made for a robust approximation of prevailing shipping prices to China, and fulfilled the statute’s aim. See 19 U.S.C. § 1677(5)(E)(iv).
CONCLUSION AND ORDER
The court has reached the end of the road. Many of the arguments considered along the way lacked merit, but the issue of greatest probable import to the Plaintiffs — the calculation of world benchmarks for steam coal, sulfuric acid, and calcium carbonate — will return to the agency for review and revisions. The court expects Commerce to return a countervailing duty rate that complies with the statute, the regulations, and this opinion.
Upon consideration of all papers and proceedings herein, it is hereby:
ORDERED that the final determination of the International Trade Administration, United States Department of Commerce, published as Certain Citrate Salts from the People’s Republic of China, 79 Fed. Reg. 108 (Dep’t Commerce Jan. 2, 2014) (final admin, review), be, and hereby is, REMANDED to Commerce for redetermi-nation; it is further
ORDERED that Plaintiffs’ Rule 56.2 Motion for Judgment on the Agency Record be, and hereby is, GRANTED as provided in this Opinion and Order; it is further
ORDERED that Plaintiff-Intervenor’s Rule 56.2 Motion for Judgment on the Agency Record be, and hereby is, DENIED as provided in this Opinion and Order; it is further
ORDERED that Commerce shall issue a redetermination (“Remand Redetermination”) in accordance with this Opinion and Order that is in all respects supported by substantial evidence and in accordance with law; it is further
ORDERED that Commerce shall reevaluate world benchmarks for the steam coal, sulfuric acid, and calcium carbonate subsidies, considering whether to calculate world average prices using weighted or simple averages in light of small-quantity, high-price transactions in the underlying data, and complying with the mandate to measure the adequacy of remuneration in light of prevailing market conditions in the country subject to review; it is further
ORDERED that Commerce shall recalculate Plaintiffs’ countervailing duty rate consistent with the reevaluated benchmark prices for steam coal, sulfuric acid, and calcium carbonate; it is further
ORDERED that Commerce shall have ninety (90) days from the date of this Opinion and Order in which to file its Remand Redetermination, which shall comply with all directives in this Opinion and Order; that the Plaintiffs, Plaintiff-Intervenor, and Defendant-Intervenor shall have thirty (30) days from the filing of the Remand Redetermination in which to file comments thereon; and that the Defendant shall have thirty (30) days from the filing of Plaintiffs’, Plaintiff-Interve-nor’s, and Defendant-Intervenor’s comments to file comments.
. The GOC adds in its reply that Commerce did not collect subsidy purchase or use information in other proceedings. Pl.-Intervenor’s Reply Br. 8-11, ECF No. 60; see also Galvanized Steel Wire from the People’s Republic of China, 77 Fed.Reg. 17,418 (Dep't Commerce Mar. 26, 2012) and accompanying I & D Mem. at IV.Á.1; Circular Welded Austenitic Stainless Pressure Pipe from the People’s Republic of China, 74 Fed.Reg. 4936 (Dep’t Commerce Jan. 28, 2009) and accompanying I & D Mem. at cmt. 7. But these cases have no bearing here. The agency has a proven track record of soliciting statistical information in reviews of the citric acid order, and that was enough to alert the GOC to Commerce’s demands.
. The exhibit documenting RZBC’s purchases from trading companies is dated September 27, 2011. This is a date long before the third administrative review began. See Administrative Review Request, PD 1 (May 14, 2012). It seems that the exhibits were furnished in error, but the court assumes for now that RZBC bought its inputs through trading companies as alleged. See GOC Br. 19.
. RZBC also argues that Commerce used the wrong information to estimate the benefit from the land subsidies. RZBC Br. 42-43. The agency calculated benefit using Asian Marketview data for Thai industrial parks, but RZBC notes that its parcels were not developed during the review period. See Prelim. Results Calculation Mem. at 8, CD 95 (June 7, 2013). The court will not consider these arguments, however, because RZBC did not raise them below. See I & D Mem. at 64; 28 U.S.C. § 2637(d) (”[T]he Court of International Trade shall, where appropriate, require the exhaustion of administrative remedies.”).
. In its brief, the defendant blusters that Commerce simple averaged the sulfuric acid and calcium carbonate prices for good reasons. Def.'s Resp. to Pis.’ & Pl.-Intervenor’s Mots. For J. on Agency R. 24-25, ECF No. 55. Apparently, weight averaging the sulfuric acid prices would have caused rounding error, and the calcium carbonate data were not amenable to weighted averages because they were given in hard copy. These excuses sound flimsy. The court will withhold judgment, however, because the agency never raised these arguments below. The court cannot sustain an agency decision backed by post-hoc justifications. SEC v. Chenery Corp., 318 U.S. 80, 95, 63 S.Ct. 454, 87 L.Ed. 626 (1943).
. The simple-average world prices in this table (sulfuric acid) and the table following (calcium carbonate) come from the agency’s final calculations. Final Calcs, at Attach. 1. The weighted-average world prices come from the Plaintiffs’ ministerial error comments. Min. Error Cmts. at Exs. 1 (sulfuric acid), 2a (calcium carbonate). The court lists these prices for demonstrative purposes only; the agency may calculate new world prices as necessary to comply with the statute, the regulations, and this opinion on remand.
. RZBC suggests that Commerce could simple average the IMF and Platts data, but weight average the GTA data. The agency is welcome to consider this or another reasonable approach on remand. See Oral Argument at 41:09.
. RZBC counters that the agency excluded the special service charge from freight costs in past cases. RZBC Br. 43-44. But this observation goes nowhere. In prior proceedings, Commerce deleted the charges after seeing evidence that no special services were needed to ship steel billet and rounds. Certain Oil Countiy Tubular Goods from the People’s Republic of China, 74 Fed.Reg. 64,045 (Dep’t Commerce Dec. 7, 2009) (final admin, determination) (“PRC OCTG Determination ”) and accompanying I & D Mem. at cmt. 13.D; see also Certain Oil Country Tubular Goods from the People’s Republic of China, 78 Fed. Reg. 9368 (Dep't Commerce Feb. 8, 2013) and accompanying I & D Mem. at l.A (relying on PRC OCTG Determination to hold that shipping billet did not incur special fees); Certain Seamless Carbon and Alloy Steel Standard, Line, and Pressure Pipe from the People’s Republic of China, 75 Fed.Reg. 57,444 (Dep't Commerce Sept. 21, 2010) and accompanying I & D Mem. át cmt. 9.C (same). We have no evidence here of comparable utility.