Appellants1 appeal an August 28,1992 decision of the Court of International Trade *1056(CIT), Creswell Trading Co. v. United States, 797 F.Supp. 1038 (Ct.Int’l Trade 1992), sustaining a Department of Commerce (Commerce) determination countervailing the entirety of certain rebates provided by the Indian government pursuant to India’s International Price Reimbursement Scheme (IPRS)2 to exporters of iron-metal castings3 for the period from January 1, 1985 to December 31, 1985. For the reasons set forth below, we reverse and remand.
The CIT and Commerce have failed to recognize the appropriate burdens of proof and production to be applied in the type of countervailable duty inquiry at issue herein. Although the ultimate burden of proving countervailability in this case rested with Commerce, the existence of India’s IPRS program by itself constituted presumptive evidence that the IPRS rebates were coun-tervailable. The evidence relied upon by Appellants, however, sufficiently rebutted this presumption. Thus, the burden of production shifted back to Commerce to prove that this evidence was either inaccurate or insufficient, which it failed to do. On this basis alone, reversal is required. Nevertheless, the CIT also erred in failing to recognize the inherent procedural unfairness in Commerce’s ruling that the evidence of record was too ambiguous and deficient for it to render a decision, given that Commerce had indicated during the initial stage of its investigation that it need not consider the type of evidence that it later found missing and given that Commerce indicated affirmatively during a later stage of its investigation that it had all of the information it needed.
I.
The CIT had jurisdiction to consider this case pursuant to 19 U.S.C. § 1516a(a)(2)(B)(iii) (1988) and 28 U.S.C. § 1581(a) (1988). We therefore have jurisdiction over this appeal pursuant to 28 U.S.C. § 1295(a)(5) (1988).
The issue before us is whether the CIT erred in sustaining Commerce’s ruling that the IPRS rebates India made to its exporters in 1985 were countervailable in their entirety. In reviewing Commerce’s ruling, the CIT was guided by a statutory standard of review which provided that Commerce’s ruling should be held unlawful if it was “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a(b)(l)(B) (1988). To determine whether the CIT correctly applied this standard in reaching its decision, we must apply anew this statutory standard of review to Commerce’s ruling and affirm the CIT unless we conclude that Commerce’s ruling is not supported by substantial evidence on the record or is otherwise not in accordance with the law. Nitta Indus. Corp. v. United States, 997 F.2 1459, 1460 (Fed.Cir.1993); see also PPG Indus., Inc. v. United States, 978 F.2d 1232, 1236 (Fed.Cir.1992); Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1189 (Fed.Cir.1990); Atlantic Sugar, Ltd. v. United States, 744 F.2d 1556, 1559 n. 10 (Fed.Cir.1984). “Substantial evidence is more than a mere scintilla. It means such relevant evidence [considering the record as a whole] as a reasonable mind might accept as adequate to support a conclusion.” Matsushita Elec. Indus. Co., Ltd. v. United States, 750 F.2d 927, 933 (Fed.Cir.1984).
II.
On December 10,1990, Commerce issued a first determination that the 1985 IPRS re*1057bates were countervailable in their entirety. First Commerce Determination, 55 Fed.Reg. 50747. Commerce stated that it was “irrelevant” whether the 1985 IPRS rebates were consistent with Item (d) of the Illustrative List of Export Subsidies (“Illustrative List”) annexed to the Agreement on Interpretation and Application of Articles VI, XVI, and XXIII of the General Agreement on Tariffs and Trade (GATT).4 Item (d) provides that the following is considered to be a subsidy:
(d) The delivery by governments or their agencies of imported or domestic products or services for use in the production of exported goods, on terms or conditions more favourable than for delivery of like or directly competitive products or services for use in the production of goods for domestic consumption, if (in the case of products) such terms or conditions are more favourable than those commercially available on world markets to their exporters. [Emphasis ours.]
H.R.Doc. No. 158, Pt. I, 96th Cong., 1st Sess. 295 (1979).
More specifically, Commerce stated in this first determination:
We consider a government program that results in the provision of an input to exporters at a price lower than to producers of domestically-sold products to confer a subsidy within the meaning of section 771(5). It is irrelevant whether the IPRS is consistent with [I]tem (d) because we are not concerned with world market prices but with the alternative price of pig iron commercially available in the domestic market.
First Commerce Determination, 55 Fed.Reg. at 50749. Commerce believed that it need not defer to Item (d) because section 771(5) of the Tariff Act grants Commerce the authority to find subsidies outside those situations specifically set forth in the Illustrative List. The CIT disagreed.
In a decision dated January 31, 1992, the CIT overruled Commerce’s first determination, holding that “Commerce must examine the IPRS in light of the exception to counter-vailable subsidies provided in [I]tem (d).” Creswell Trading Co. v. United States, 783 F.Supp. 1418, 1419 (Ct.Int’l Trade 1992). The CIT reasoned that Commerce must consider Item (d) because Item (d) explicitly addresses the type of situation at issue in this case. The CIT logically concluded that Commerce’s authority to find subsidies in situations not specifically listed in the Illustrative List extends only to those situations which do not fall into one of the specific situations listed. Thus, the CIT refused to allow Commerce to rely on its alleged broad authority under the Tariff Act to circumvent the explicit provisions of Item (d). Accordingly, the CIT remanded with a specific order to Commerce to “examine the IPRS in light of [I]tem (d) and determine whether the Indian government’s provision of pig iron was on terms more favorable than on world markets.” Creswell, 783 F.Supp. at 1420-21.
On February 26, 1992, Commerce issued a Draft Redetermination on Remand, Final Results of Countervailing Duty Administrative Review (“Draft Remand Determination”), which was released to all parties for comment, and in which Commerce stated that it “had found enough information to try to determine whether the IPRS program was consistent with the [I]tem (d) exception.” Final Remand Determination at 3. Commerce described therein its methodology for analyzing the IPRS payments for counter-vailability upon remand as follows:
In determining whether IPRS payments qualify for the [I]tem (d) exception as interpreted by the Court [CIT], the Department attributed a countervailable benefit from the IPRS equal to the amount of ,the overrebate of the differential between the Indian domestic price of pig iron and the international price of pig iron.
Draft Remand Determination at 2. Although the foregoing adequately describes the appropriate calculation to be carried out, Commerce inadvertently miscalculated the “overrebate” due to Commerce’s understandable confusion regarding the pricing data *1058that the Indian government provided it. Commerce’s confusion resulted from the Indian government’s use of the terminology “international price” in its August 26, 1987 questionnaire response to refer to two different items, namely, the price at which the Indian government believed pig iron could be purchased internationally, i.e., at terms and conditions available on world markets, and the post-rebate price of domestically-produced pig iron established by the Indian government.
The Indian government provided a table in this questionnaire response which, pursuant to a header therein, purported to provide “international price” information for pig iron during 1985 on a monthly basis. Following this table, the Indian government provided an explanation as to how it arrived at these 1985 prices as follows:
Since no price quotation is available in any international journal or metal bulletins for determining international price, of pig iron, it was decided in 1983 that the price quoted to Steel Authority of India, Ltd. [SAIL] (which is a public sector unit and through which imports of steel and pig iron was canalised at that time) should be the basis for calculating the international ■prices. On the basis of contracts entered into by SAIL for import of pig iron in 1983, the international price was determined at $126 per metric ton FOB. The international prices for pig iron for the period January to October, 1985 were based on the rate of $126 per metric ton FOB and an element on account of- excise duty was added to those prices. The variations in the prices were on account of exchange rate fluctuations between the US dollar and the Indian rupee. The international prices were reviewed by the Standing Committee headed by the Chief Controller of Imports and Exports and it was found that the international price at the rate of $126 per metric ton FOB was in fact on the higher side as compared to the price quoted to the tender floated by MMTC [Minerals and Metals Trading Company] and opened in October, 1985. The Committee, however, decided to maintain the international price at US$ 126 per metric ton FOB and not to lower it to the price quoted to MMTC viz. US$ 113 per metric ton at higher rates under the IPRS in respect of pig iron used for exports of engineering goods. For the period from November to December, 1985, the international price for iron was fixed on the basis of quotations received by MMTC viz. US$ 113 per metric ton.
Questionnaire Response at pages 14-15 (emphasis added).
The foregoing illustrates the Indian government’s loose usage of the language “international price.” Upon close analysis, it is evident that the $126 and $113 price quotes in the excerpt represent the prices at which the Indian government believed it could purchase pig iron internationally. It is equally evident that the “international price” information provided in the table represents the post-rebate price of domestically-produced pig iron established by the Indian government. Unfortunately, due to the Indian government’s loose usage of the terminology “international price,” Commerce believed the opposite, namely that the pricing information in the table represented the price at which the Indian government believed pig iron could be purchased internationally and that the $126 and $113 price quotes represented the post-rebate price established by the Indian government.5 Consequently, in calculat*1059ing whether a subsidy existed,. Commerce mistakenly subtracted the price at which pig iron could be purchased internationally from the post-rebate price of pig iron established by the Indian government, instead of vice versa. Following issuance of its draft remand determination, Commerce received several letters advising it that it had misinterpreted the pricing data that the Indian government had provided.
On March 16, 1992, after reviewing the record already established and without requesting any additional information from the Indian government or Appellants, Commerce issued a final, second determination in which Commerce found that the record evidence was “ambiguous” and “woefully deficient to support any claim ... that the IPRS program is consistent with the so-called world-market price exception under [I]tem (d).” Redetermination on Remand/Final Results of Countervailing Duty Administrative Review Pursuant to Court Remand/Certain Iron-Metal Castings from India (“Final Determination") at 4. Consequently, because the alleged ambiguity and deficiency of the record evidence made it impossible for Commerce “to determine whether the IPRS program is consistent with the world market price exception in [I]tem (d),” Commerce ruled that the IPRS rebates made in 1985 were direct export subsidies which were countervailable in their entirety. Final Remand Determination at 10.
On August 28, 1992, the CIT affirmed Commerce’s final remand determination that the 1985 IPRS rebates were countervailable in their entirety and dismissed the proceeding. In reaching this decision, the CIT accepted Commerce’s interpretation that Item (d) provides an “exception” to countervaila-bility for goods that are not provided on terms or conditions that are more favorable than those commercially available to exporters on world markets, and that, as a result, Item (d) imposes a burden on an exporter to “establish” a singular “world market price” before that exporter may claim the benefit of the Item (d) exception. Creswell, 797 F.Supp. at 1040-41. Going forward from this premise, the CIT held that there was substantial evidence of record to support Commerce’s determination that Appellants ■ had “failed to carry their burden of proof’- regarding the “world market price” during the relevant period. Creswell, 797 F.Supp. at 1041.6
III.
A. Burdens of Proof and Production
(1) Applicable Burdens
At the heart of this ease lie the intertwining issues of who bears the ultimate burden of proof of establishing the existence of a countervailable subsidy pursuant to Item (d) and of what the appropriate procedural burdens of production are during the course of a Commerce investigation to determine if such a countervailable subsidy exists. By “burden of proof,” we mean the ultimate burden of persuasion in convincing a fact finder of the existence of a countervailable subsidy. By “burden of production,” we mean the burden of going forward. Because there is no statutory or regulatory guidance, the appropriate standard of proof in an Item (d) analysis is by a preponderance of the evidence. The proper allocation of the burdens of proof and production are important procedural rights that may have substantive consequences, and therefore we find it necessary to consider Commerce’s assignment of these burdens as part of our review to determine whether Commerce’s ruling in this ease was “in accordance with law.”7
*1060Neither Appellants nor Appellees have pointed to any explicit statutory or regulatory allocation of the burdens of proof or production with respect to establishing the existence of a countervailable subsidy under Item (d). However, Item (d) itself is enlightening on this point. As noted supra, Item (d) provides that the following is considered to be a subsidy:
(d) The delivery by governments or their agencies of imported or domestic products or services for use in the production of exported goods, on terms or conditions more favourable than for delivery of like or directly competitive products or services for use in the production of goods for domestic consumption, if (in the case of products) such terms or conditions are more favourable than those commercially available on world markets to their exporters. [Emphasis added.]
The “if’ clause of Item (d) sets forth on its face a statutory condition that Commerce must establish before it may exercise its right to levy a countervailing duty against an investigated party, as opposed to an exception into which that party must prove its actions fall. The ultimate burden of proof is thus upon Commerce to establish by a preponderance of the evidence that a government, or an agency thereof, has delivered to an exporter products or services for use in the production of exported goods on terms or conditions more favorable than those commercially available to the exporter on world markets.
Typically, the burden of production is initially upon the party who bears the ultimate burden of proof and generally requires that a party produce sufficient evidence to support a finding in favor of that party. The burden of production then shifts to the other party, who must in turn produce enough evidence to raise a question of material fact. However, in the context of an Item (d) investigation, Commerce necessarily requires information that is within the knowledge and control of the government of the exporting country, its exporters, or both, which information Commerce cannot obtain independently. For example, Commerce has need of information regarding the terms and conditions at which the government provided the exporter -with the services or products under investigation, how the government arrived at these terms and conditions, and how these terms or conditions compared to the terms and conditions at which a domestic seller of the exported goods could obtain these services or products. Thus, it is only logical that some burden be placed on the government or exporter to come forward with such information. See Zenith Elec. Corp. v. United States, 988 F.2d 1573, 1583 (Fed.Cir.1993) (“The burden of production should belong to the party in possession of the necessary information.”)
To this end, we hold that the existence of a program wherein a government, or an agency thereof, delivers to an exporter products or services for use in the production of exported goods on terms or conditions more favorable than for delivery of like or directly competitive products or services for use in the production of goods for domestic consumption is, standing alone, presumptive evidence that the program also provides the products or services under investigation to that exporter on terms or conditions more favorable than the terms and conditions available on world markets. Commerce’s initial burden of production is thus satisfied by way of this presumption. The burden of production accordingly shifts to the exporter to come forward with evidence that the services or products were not provided on terms or conditions more favorable than available on world markets.
We do not hold that the ultimate burden of proof is shifted by this presumption. Rather, we merely hold that this presumption creates a prima facie case that shifts the burden of production to the exporter to come forward with sufficient evidence to rebut this presumption. See Fed.R.Evid. 301. The totality of the evidence must then be weighed to determine whether a countervailable subsidy exists. Commerce, which retains the *1061ultimate burden of persuasion, may prevail only if it establishes the existence of a eoun-tervailable subsidy by a preponderance of the evidence.
(2) Allocation of Burdens in This Case
It is uncontroverted that, by way of the IPRS program, the government of India provided Appellants with pig iron, a product for use in the production of exported goods, on terms and conditions more favorable than those available to manufacturers who produced iron castings for domestic consumption. This creates a presumption that the pig iron was provided on terms and conditions which were also more favorable than those commercially available on world markets. The burden of production thus shifted to Appellants to come forward with sufficient evidence to establish that such favorability did not exist. Appellants met this burden in so far as they relied on the information that the Indian government provided in its August 26, 1987 questionnaire response discussed previously.
The Indian government explained in this questionnaire response, and Appellants have admitted in these proceedings, that it was not aware of any published, international reference price for pig iron when the 1985 IPRS rebates were given. Because it knew of no such published international price, the Indian government did the next best thing — it looked to actual market experience. The record establishes that the Indian government based the $126 per ton international benchmark price used from January to October of 1985 on a 1988 SAIL contract and the $113 per ton international benchmark price used from November to December of 1985 on a price quotation in a tender floated by MMTC in 1985. An article in The Mining Journal, Ltd.; Mining Annual Review (June, 1985) corroborates the substantial accuracy of the international benchmark prices used by the Indian government for the 1985 time period, in that it states that, as of 1984, “Australians were looking for a pig iron price substantially above the prevailing world level of around $.US120/t.” Mining Journal at 86.
The foregoing sufficiently rebutted the presumption of countervailability attached to the IPRS program, and thus the burden of production shifted back to Commerce to come forward with proof that this evidence was neither accurate nor sufficient to establish that the IPRS rebates paid to exporters did not exceed the difference between the domestic price of pig iron and the price at which it could have been purchased internationally. Commerce failed to meet this burden.
In its final remand determination, Commerce did nothing more than allege that the Indian government’s method of establishing international benchmark prices for calculating the 1985 IPRS rebates was inaccurate. Indeed, Commerce failed to set forth any evidence at all that these international benchmark prices were in fact inaccurate. The record before us clearly indicates that Commerce made no attempt whatsoever to determine any price at which pig iron could be purchased on world markets during the relevant time period.8
In this appeal, Appellees suggest that the Indian government’s international benchmark prices were inherently inaccurate, as they were allegedly based on outdated and inconclusive sources, and therefore Commerce did not need to come forward with any evidence of their inaccuracy. Specifically, Appellees argue that there was no reasonable basis for the Indian government to believe that the price of pig iron had not changed since the 1983 SAIL contract, that the single price quoted in the tender floated by MMTC in 1985 did not alone establish what the world market price was in 1985, and that the reporting of “a” price for pig iron in the Mining Journal does not establish that this was the prevailing world market price.
*1062We are unconvinced by these arguments. Viewed in combination, the 1983 SAIL contract price, the 1985 MMTC price quote, and the Mining Journal article represent more than a bare allegation of terms and conditions on world markets, and they support the substantial accuracy of the Indian government’s international benchmark prices. Nothing in the record casts doubt on the validity of these prices or suggests that these prices were not in accord with prices in markets outside India.9
Commerce’s misunderstanding of its burden of coming forward with evidence that the Indian government’s international benchmark prices were in fact inaccurate was a procedural error that went to the heart of its administrative determination justifying reversal. In addition, because there is no evidence of record indicating that the international benchmark prices used by the Indian government to calculate the 1985 IPRS rebates were inaccurate, Commerce’s ruling that the 1985 IPRS rebates resulted in India providing its exporters with pig iron on terms or conditions more favorable than those commercially available on world markets was not based on substantial evidence. The CIT thus erred as a matter of law when it failed to reverse Commerce’s final remand determination. For the foregoing reasons, reversal and remand is necessary so that Commerce may carry out a countervailing duty investigation pursuant to a proper allocation of burdens of proof and production.
B. Inherent Procedural Unfairness
The foregoing having been said, we are compelled to address in addition the inherent unfairness as to the manner in which Commerce handled its Item (d) investigation in this case.
During the course of Commerce’s investigation leading up to its first determination, Commerce considered Item (d) to be “irrelevant” and thus never suggested to the Indian government or Appellants the need to establish a “world market price” for pig iron in 1985. City Pipe contends in its Reply Brief, and Appellees do not contest, that all information requested by Commerce during its investigation leading up to the first determination was produced. Commerce also never suggested that the Indian government or Appellants needed to provide any further information as to the 1985 international price of pig iron during the course of its investigation following remand by the CIT. On the contrary, Commerce’s draft remand determination indicated that it had all of the information that it required to determine whether a countervailable subsidy existed.
Commerce is presumably in the best position to know what it means by its own requirements and what evidence will satisfy these requirements, and therefore, given the unique circumstances of this case, Commerce at a minimum bore a burden of requesting any additional information that it required when it came to its conclusion during the course of its remand determination, contrary to earlier indications, that the information of record was insufficient. It was inherently unfair for Commerce to take an affirmative position during the course of its initial investigation that “world market price” information was “irrelevant” and then, when the time had long passed for submitting such information, to declare the missing information so relevant that it could not render a determination in its absence. In this same vein, it also was unfair for Commerce to take an affirmative position during its remand investigation that it had all of the information that it needed and then declare in its final determination that the evidence was deficient without giving the Indian government or Appellants an opportunity to supplement the record. Appellant Crescent appropriately summarized the unfairness at page 8 of its reply brief when it stated: “India somehow had the burden to prove the ‘international price’ to the satisfaction of [Commerce], even though what [Commerce] required of India was never known or communicated to India and was never even an issue at the administrative level.”
*1063CONCLUSION
For the foregoing reasons, we hold that Commerce’s ruling is neither supported by substantial evidence on the record nor in accordance with law. Accordingly, to allow Commerce to correct the errors discussed above, the August 28, 1992 judgment of the CIT is reversed, and the case is remanded for action consistent with this opinion.
REVERSED AND REMANDED.
. City Pipe & Foundry, Inc., Capitol Foundry of Virginia, Inc., Crescent Foundry Co. P. Ltd., Select Steels, Ltd., R.B. Agarwalla & Co., Seram-pore Industries P., Ltd., Super Castings (INDIA), Carnation Enterprises P., Ltd., Uma Iron & Steel Co., and Commex Corporation (collectively "Appellants”) are Indian exporters of iron-metal *1056castings who participated as plaintiffs-interve-nors before the CIT.
. The IPRS was introduced in 1981 by the Government of India. Under the IPRS, users of domestically produced steel, pig iron, or scrap, as raw materials, are entitled upon export of a finished product to a refund equal to the difference between the higher cost of the Indian-produced raw material and the cost of that same raw material if purchased internationally.
. For purposes of this opinion, we refer to the products under investigation as "castings." Commerce described these products generally in a previous decision, not under review here, as follows:
Imports covered by the review are shipments of Indian manhole covers and frames, clean-out covers and frames, and catch basin grates and frames. These articles are commonly called municipal or public works castings and are used for access or for drainage for public utility, water, and sanitary systems.
Certain Iron-Metal Castings from India; Final Results of Countervailing Duty Administrative Review (“First Commerce Determination "), 55 Fed. Reg. 50747 (Dec. 10, 1990).
. "The list in question is a list prepared under the [GATT] which, by 19 U.S.C. §§ 2502(1) and 2503(c)(5), the Congress incorporated into United States Law.” RSI (India) Pvt., Ltd. v. United States, 876 F.2d 1571, 1572 (Fed.Cir.1989).
. The following description of the calculation that Commerce carried out in its draft remand determination to determine whether a counter-vailable subsidy existed evidences Commerce's error:
1) For each month of 1985, the record contains an "international” metric ton price of pig iron. The basis for these "international" pig iron prices are arm's-length quotes solicited by the Steel Authority of India Ltd. (SAIL), which is a public agency that oversees the import of pig iron. (These price quotes were provided by SAIL on page 14 of the questionnaire response.)
2) For each month of 1985, the record contains the Engineering and Export Council's (EEPC) reference price in metric tons. This is the price that the EEPC uses to calculate the amount of IPRS reimbursement. (These reference prices were provided by the EEPC on page 15 of the questionnaire response and are cited on page 2 of the verification report.)
*10593) For each month, we subtracted the EEPC reference price from the international price. If the difference is positive, an overrebate exists. That is, if the EEPC reference price is less than the international price, the net cost of pig iron to the exporter is less than the international price of pig iron and hence preferential.
Draft Remand Determination at 2.
. Of interest, this court previously affirmed a CIT decision upholding a Commerce determination that IPRS rebates made in 1984 were counter-vailable subsidies based on the record established in that 1984 review. RSI, note 4, supra.
. Appellees suggest that this court should not involve itself in a burden of proof analysis because Appellants’ allegedly did not raise arguments on this point below. We disagree. The issue of who bears the burdens of proof and *1060production has been at the heart of this dispute 1 all along, and we can not turn a blind eye to this issue.
. Given that Commerce had sufficient resources to send a team to India to ensure that the Indian government did indeed base its 1985 IPRS rebates on the $126 and $113 benchmark prices indicated in its questionnaire response, any arguments that Commerce did not have the necessary resources to determine whether pig iron could be purchased on international markets in 1985 at any price substantially different from that relied upon by the Indian government seem disingenuous at best.
. Moreover, to the extent that the price of pig iron on world markets had changed since the 1983 SAIL contract, the 1985 MMTC price quote evidences that the price had gone down, and thus the Indian government's international benchmark price of $126 in fact represented a conservative estimate which kept IPRS rebates lower than allowed.