This action stems from an administrative review of a countervailing duty order covering iron-metal castings from India. Plaintiffs-Appellants, Kajaria Iron Castings Pvt. Ltd. ("Kajaria"), Calcutta Ferrous Ltd., Crescent Foundry Co. Pvt. Ltd., Commex Corporation, Dinesh Brothers ("Dinesh"), Nandikeshwari Pvt. Ltd., Carnation Enterprises Pvt. Ltd., Kejriwal Iron & Steel Works, R.B. Agarwalla & Company, RSI Limited, Serampore Industries Pvt. Ltd., Tirupati International (P) Ltd., and Uma Iron & Steel Co. (collectively "Producers"), are Indian producers and exporters of iron-metal castings. Defendants-Appellees, Alhambra Foundry, Inc., Allegheny Foundry Co., Deeter Foundry, Inc., East Jordan Iron Works, Inc., Lebaron Foundry Inc., Municipal Castings, Inc., Neenah Foundry Co., U.S. Foundry & Manufacturing Co., and Vulcan Foundry, Inc. (collectively "Domestic Industry"), are United States producers of iron-metal castings. The Producers appeal the final decision of the United States Court of International Trade sustaining the determination of the International Trade Administration, United States Department of Commerce ("Commerce"), that the Producers were receiving net subsidies on iron-metal castings that were within the scope of the countervailing duty order and that were imported into the United States, and therefore imposing countervailing duties on the castings. See Kajaria Iron Castings Pvt. Ltd. v. United States, 969 F. Supp. 90 (Ct. Int'l Trade 1997) ("Kajaria II "); Kajaria Iron Castings Pvt. Ltd. v. United States, 956 F. Supp. 1023 (CIT 1997) ("Kajaria I ").

The Producers allege that Commerce's methodology double counted certain subsidies and included income from merchandise not covered by the countervailing duty order in calculating the net subsidy. They also allege that Commerce's methodology in calculating the country-wide countervailing duty rate was erroneous because it included producers with significantly different higher subsidy rates and rates based on the best information available ("BIA"). Because Commerce's methodology did double count subsidies and did include income from merchandise not covered by the countervailing duty order in determining the net subsidy, we reverse and remand for further proceedings. However, we affirm Commerce's methodology for calculating the country-wide countervailing duty rate.

The countervailing duty laws impose additional duties on merchandise that is imported, or sold or likely to be sold for importation, into the United States when the manufacture, production, or exportation of the merchandise is directly or indirectly subsidized. See 19 U.S.C. §§ 1303(a), 1671(a) (1988).1 These additional duties, called "countervailing" duties, are levied on subsidized imports to offset the unfair competitive advantages created by the foreign subsidies. See Wolff Shoe Co. v. United States, 141 F.3d 1116, 1117 (Fed.Cir.1998). The quantum of the countervailing duties imposed upon such merchandise is equal to the amount of the net subsidy. See 19 U.S.C. § 1671(a); see also 19 U.S.C. §§ 1677(6) (defining "net subsidy"), 1677(5) (defining "subsidy").

Once affirmative determinations of subsidization and material injury or threatened material injury to a domestic industry have been made, see 19 U.S.C. §§ 1671(a), 1671d(b), 1671e(a), Commerce issues a countervailing duty order that directs customs officials to assess a countervailing duty "equal to the amount of the net subsidy determined or estimated to exist." See 19 U.S.C. § 1671e(a)(1). That countervailing duty rate presumptively applies to all merchandise within the scope of the investigation. See 19 U.S.C. § 1671e(a)(2). Commerce annually reviews countervailing duty orders if properly requested to do so. See 19 U.S.C. § 1675(a)(1)(A); 19 C.F.R. § 355.22 (1992).

In 1980, Commerce issued a countervailing duty order covering "certain iron-metal castings [from India] consisting of manhole covers and frames, clean-out covers and frames, and catch basin grates and frames."2 Certain Iron Metal Castings From India: Countervailing Duty Order, 45 Fed.Reg. 68,650 (Oct. 16, 1980). On October 27, 1992, the Municipal Castings Fair Trade Council and individually-named members ("petitioners") requested an administrative review of the 1980 order pursuant to 19 U.S.C. § 1675(a)(1)(A) and 19 C.F.R. § 355.22(a). See Certain Iron-Metal Castings From India: Preliminary Results of Countervailing Duty Administrative Review, 60 Fed.Reg. 4596 (Jan. 24, 1995) ("Preliminary Results "). On November 27, 1992, Commerce initiated review for the period January 1 through December 31, 1991, involving fourteen producers/exporters ("producers") and twelve Indian programs. See Initiation of Antidumping and Countervailing Duty Administrative Reviews, 57 Fed.Reg. 56,318 (Nov. 27, 1992); Preliminary Results, 60 Fed.Reg. at 4596.

On January 24, 1995, Commerce issued its preliminary results, in which it determined the country-wide net subsidy rate with respect to the iron-metal castings at issue to be 5.54 percent ad valorem for all manufacturers and exporters, except for firms which had received significantly different aggregate benefits and were assigned individual net subsidy rates. See Preliminary Results, 60 Fed.Reg. at 4596. Three firms were assigned significantly different net subsidy rates: (1) Dinesh (a net subsidy rate of zero), (2) Super Castings (India) Pvt. Ltd. ("Super Castings") (a net subsidy rate of 41.75 percent), and (3) Kajaria (a net subsidy rate of 16.14 percent). See id. at 4599. Commerce explained its calculation methodology as follows:

Pursuant to Ceramica Regiomontana, S.A. v. United States, 853 F. Supp. 431 (CIT 1994), Commerce is required to calculate a country-wide [countervailing duty] rate, i.e., the all-other rate, by "weight averaging the benefits received by all companies by their proportion of exports to the United States, inclusive of zero rate firms and de minimis firms." Therefore, we first calculated a subsidy rate for each company subject to the administrative review. We then weight-averaged the rate received by each company using as the weight its share of total Indian exports to the United States of subject merchandise. We then summed the individual companies' weight-averaged rates to determine the subsidy rate from all programs benefitting exports of subject merchandise to the United States.

Since the country-wide rate calculated using this methodology was above de minimis, as defined by 19 CFR 355.7 (1993), we proceeded to the next step and examined the net subsidy rate calculated for each company to determine whether individual company rates differed significantly from the weighted-average country-wide rate, pursuant to 19 CFR 355.22(d)(3). Three companies ... received significantly different net subsidy rates during the review.... These companies are treated separately for assessment and cash deposit purposes. All other companies are assigned the country-wide rate.

Id. at 4596.

In determining the subsidies received by each producer, Commerce found that the producers had received countervailable benefits under section 80HHC of India's Income Tax Act ("section 80HHC"). See id. at 4597. Pursuant to this provision, the government of India "allows exporters to deduct profits derived from the export of goods and merchandise from taxable income." Id. Commerce had determined in prior administrative reviews that the program provided a countervailable subsidy because the "receipt of benefits under this program [was] contingent upon export performance." Id. Commerce determined in this review that no new information altered its prior determination. See id. Commerce calculated the countervailable subsidy attributable to section 80HHC by subtracting the total amount of income tax a producer actually paid from the amount of income tax the producer would have paid absent the section 80HHC deduction. See id. at 4598.

Commerce also determined that the producers had received over-rebates under India's Cash Compensatory Support ("CCS") Program.3 See id. "The CCS [P]rogram rebates indirect taxes and import duties borne by inputs physically incorporated into an exported product." Kajaria I, 956 F. Supp. at 1025. Through the CCS Program, producers receive a tax rebate upon export, calculated as a percentage of the invoice price of the exported merchandise. See Preliminary Results, 60 Fed.Reg. at 4598. In prior administrative reviews, Commerce had determined that the CCS Program did not provide a countervailable benefit because "(1) [t]he program operate[d] for the purpose of rebating prior stage cumulative indirect taxes and/or import charges; (2) the government [of India] accurately ascertained the level of the rebate; and (3) the government [of India] reexamine[d] its schedules periodically to reflect the amount of actual indirect taxes and/or import charges paid." Id. However, Commerce noted that it must determine on a case-by-case basis whether there is an over-rebate, i.e., "whether the rebate for the subject merchandise exceeds the total amount of indirect taxes and import duties borne by inputs that are physically incorporated into the exported product." Id. If an over-rebate exists, the difference between the allowable rebate and the actual rebate is a countervailable subsidy. See id.

During the review period, the producers switched from domestic pig iron to imported pig iron as the basic raw material used in the production of castings exported to the United States. See id. The producers listed port and harbor taxes associated with the imported pig iron as part of the indirect tax incidence on inputs of the subject castings. See id. During the verification of the 1990 administrative review, Commerce determined that the port tax is a wharfage charge and the harbor tax included berthage, port duties, pilotage, and towing charges. See id. During the current review, the government of India failed to show that the port and harbor taxes were indirect taxes. See id. Commerce determined that those taxes were actually service charges and therefore disallowed those items in the calculation of the indirect tax incidence. See id. Commerce recalculated the indirect tax incidence incurred on items physically incorporated in the manufacture of the exported castings, excluding the service charges, and compared the recalculated tax incidence rate to the CCS rebates. See id. Commerce then determined that the CCS Program provided an over-rebate of indirect taxes and that the over-rebates were countervailable subsidies. See id.

Commerce also found that the producers did not apply for or receive benefits under India's International Price Reimbursement Scheme ("IPRS") with respect to subject castings. See id. at 4599. The IPRS reimburses producers for the difference in price between higher-priced domestic pig iron and imported pig iron. See Kajaria I, 956 F. Supp. at 1025; see generally Creswell Trading Co. v. United States, 15 F.3d 1054 (Fed.Cir.1994), and Creswell Trading Co. v. Allegheny Foundry Co., 141 F.3d 1471 (Fed.Cir.1998) (explaining the IPRS program and its history).

On August 29, 1995, Commerce issued its final results, in which it adjusted the countervailing duty rate for all producers, other than those assigned individual rates, to 5.53 percent ad valorem. See Certain Iron-Metal Castings From India: Final Results of Countervailing Duty Administrative Review, 60 Fed.Reg. 44,843, 44,849 (Aug. 29, 1995) ("Final Results "). In the final results, Commerce also maintained the significantly different rates for Dinesh, Super Castings, and Kajaria that were established in the preliminary results. See id. At the same time, Commerce reaffirmed the calculation methodology used in arriving at the preliminary results. See id. at 44,843-44.

Commerce rejected the producers' claim that it had erroneously classified the port and harbor taxes as service charges rather than indirect taxes. See id. at 44,845. Commerce reiterated that the over-rebates under the CCS Program were countervailable under the countervailing duty laws. See id. Based on the 1990 administrative review, Commerce determined that the port and harbor charges were not taxes on items physically incorporated into exported castings, but rather were service charges that led to over-rebates under the CCS Program that are countervailable. See id. at 44,845-46.

In response to the petitioners' claim that Commerce should have countervailed CCS rebates on non-subject castings, Commerce responded:

Where the Secretary determines that a countervailable benefit is tied to the production or sale of a particular product or products, the Secretary will allocate the benefit solely to that product or products. If the Secretary determines that a countervailable benefit is tied to a product other than the [subject] merchandise, the Secretary will not find a countervailable subsidy on the merchandise. This practice of tying benefits to specific products is an established tenet of the Department's administration of the countervailing duty law.

Id. at 44,845 (internal quotation marks and citations omitted). Using this same reasoning, Commerce also rejected the petitioners' contention that it should have countervailed IPRS rebates on non-subject castings. See id. at 44,847. Commerce reiterated that "[g]rants that are tied to the production or export of only non-subject merchandise do not provide a countervailable benefit to the subject merchandise." Id.

The producers argued that Commerce erred in countervailing the portion of the section 80HHC tax deduction that was based on the CCS over-rebates and the IPRS rebates on non-subject castings. See id. at 44,848. They argued that under the section 80HHC program CCS rebates are considered export income, which may be deducted from taxable income in determining the tax payable by the producer. See id. Since Commerce fully countervailed the CCS over-rebates, the producers claimed that countervailing the portion of the section 80HHC tax deduction attributable to those over-rebates double counted the benefit of the over-rebates. See id. The producers also argued that Commerce erred in countervailing the portion of the section 80HHC deduction attributable to IPRS rebates on non-subject castings because those rebates did not benefit castings subject to the countervailing duty order. See id.

Commerce rejected these arguments. Commerce determined that the full amount of the section 80HHC tax deduction was a countervailable subsidy and that the offsets requested by the producers did not fit within any of the statutorily-permitted offset categories. See id. Commerce noted:

Adjusting the benefit conferred by the 80HHC tax program to account for the CCS and IPRS rebates is not a permissible offset under section 771(6) of the Act. In addition, we also note that, with respect to respondents' CCS argument, that it is the Department's established policy to disregard the secondary tax effects of countervailable subsidies. See, e.g., Certain Fresh Atlantic Groundfish From Canada, 51 FR 10041 (March 24, 1986) and Fresh and Chilled Atlantic Salmon From Norway, 56 FR 7678 (February 25, 1991).

Id.

The producers also commented that it was inappropriate for Commerce to include the net subsidy rates of producers that were assigned individual, significantly higher net subsidy rates in calculating the country-wide rate applicable to all other producers. See id. In addition, they claimed that it was inappropriate for Commerce to include BIA-based net subsidy rates in the calculation of the country-wide rate.4 See id. The producers argued that including the significantly different higher rates and the BIA-based rates in calculating the country-wide rate unfairly inflated the country-wide rate. See id. at 44,849. Commerce rejected these arguments, stating:

In [Ipsco, Inc. v. United States, 899 F.2d 1192 (Fed.Cir.1990) and Ceramica Regiomontana, S.A. v. United States, 853 F. Supp. 431 (CIT 1994) ], the Department excluded the zero and de minimis company-specific rates that were calculated before calculating the country-wide rate. The court in Ceramica, however, rejected this calculation methodology. Based upon the Federal Circuit's opinion in Ipsco the court held that the Department is required to calculate a country-wide [countervailing duty] rate applicable to non-de minimis firms by "weight averaging the benefits received by all companies by their proportion of exports to the United States, inclusive of zero rate firms and de minimis firms." Ceramica, 853 F. Supp. at 439 (emphasis on "all" added).

Thus, the court held that the rates of all firms must be taken into account in determining the country-wide rate. As a result of Ceramica, Commerce no longer calculates, as it formerly did, an "all others" country-wide rate. Instead, it now calculates a single country-wide rate at the outset, and then determines, based on that rate, which of the company-specific rates are "significantly" different.

Given that the courts in both Ipsco and Ceramica state that the Department should include all company rates, both de minimis and non de minimis, there is no legal basis for excluding "significantly different" higher rates, including BIA rates. To exclude these higher rates, while at the same time including zero and de minimis rates, would result in a similar type of country-wide rates bias of which the courts were critical when the Department excluded zero and de minimis rates under its former calculation methodology.

Id.

On September 27, 1995, the Producers sought review of Commerce's final results in the Court of International Trade. The Producers did not contest Commerce's determination that the over-rebates under the CCS Program provided a countervailable subsidy. See Kajaria I, 956 F. Supp. at 1025. The Producers did argue, however, that Commerce had double counted the subsidy provided by the CCS over-rebates by countervailing both the over-rebates and the portion of the section 80HHC deduction attributable to those over-rebates. See id. The Producers also argued that Commerce had erred in countervailing the portion of the section 80HHC deduction attributable to the IPRS rebates on non-subject castings. See id. In addition, the Producers urged that Commerce had erred in including the individual subsidy rates of firms assigned significantly different higher rates and firms with BIA-based rates in calculating the country-wide countervailing duty rate applicable to all other producers. See id. at 1025-26. In a January 29, 1997 opinion, the Court of International Trade remanded to Commerce for further explanation of the methodology used in calculating the subsidy from the section 80HHC deduction and affirmed Commerce's methodology for calculating the country-wide countervailing duty rate. See id. at 1028.

The court based its decision on the issue of whether countervailing the CCS over-rebates and the portion of the section 80HHC deduction attributable to those over-rebates was improper on its prior decision in Crescent Foundry Co. Pvt. Ltd. v. United States, 951 F. Supp. 252 (CIT 1996) ("Crescent I ").5 See Kajaria I, 956 F. Supp. at 1026. In Crescent I, the producers argued that "countervailing a tax deduction [i.e., the section 80HHC deduction] while at the same time also countervailing revenues that make up that deduction [i.e., the CCS over-rebates] double counts the subsidy." 951 F. Supp. at 260. In Crescent I, the court found that Commerce's methodology was inconsistent with its prior reasoning in Carbon Steel Wire Rod From Argentina; Suspension of Investigation, 47 Fed.Reg. 42,393 (Sept. 27, 1982) ("Argentine Wire Rod "), and Final Affirmative Countervailing Duty Determinations and Countervailing Duty Orders; Certain Welded Carbon Steel Pipe and Tube Products From Argentina, 53 Fed.Reg. 37,619 (Sept. 27, 1988) (collectively "the Argentine cases "). See Crescent I, 951 F. Supp. at 260. In the Argentine cases, Commerce declined to separately countervail the tax exemption of a subsidy that already had been fully countervailed. See id. The court also noted Commerce's policy of disregarding the secondary income tax consequences of non-income tax subsidy programs, i.e., "when companies pay higher taxes as a result of receiving a subsidy, Commerce does not subtract the added taxes from the amount of the subsidy when calculating the benefit conferred." See id. at 261. The court found that Commerce's policy of disregarding secondary tax consequences and Commerce's position in the Argentine cases seemed to dictate that Commerce should not have countervailed the section 80HHC tax exemption on the CCS over-rebates. See id. Noting the apparent inconsistency between Commerce's action in this case and its previous policies and noting further that Commerce had not stated whether it had considered and rejected the reasoning in the Argentine cases, the court remanded for Commerce to (1) explain whether the Argentine cases reflected current policy and (2) reexamine whether countervailing the portion of the section 80HHC deduction attributable to the CCS over-rebates double counted the subsidy from those over-rebates. See id. at 261-62. Based on Crescent I, the court in Kajaria I similarly remanded for an explanation of the Argentine cases and reconsideration. See 956 F. Supp. at 1026.

Based on its decision in Crescent I, the court also remanded on the issue of whether Commerce had impermissibly countervailed the portion of the section 80HHC deduction attributable to IPRS rebates on non-subject castings. See id. In Crescent I, the court held that "[s]ince the IPRS program did not benefit subject merchandise ... Commerce should not have countervailed the portion of the § 80HHC tax deduction attributable to that program." See 951 F. Supp. at 262. The court stated:

Commerce may only exercise its authority to set countervailing duties through a countervailing duty order on specific merchandise. 19 U.S.C. §§ 1671(a), 1671e (1988). Its decision to countervail a tax benefit attributable to a subsidy of nonsubject merchandise exceeded its statutory authority and was not in accordance with law. Accordingly, the court remands this issue for recalculation of the benefit received through § 80HHC after subtracting the value of IPRS payments received from each company's taxable income.

Id.

The court in Kajaria I, however, sustained Commerce's inclusion of producers receiving individual, significantly different higher rates in calculating the country-wide rate based on its decision in Crescent I. See Kajaria I, 956 F. Supp. at 1026. In Crescent I, the court determined that "[i]t would be inconsistent with the concept of a country-wide rate to exclude companies receiving a significantly different benefit, whether higher or lower, when determining the amount of the weighted-average net subsidy." 951 F. Supp. at 258. The court noted that Commerce's methodology resulted in a higher net subsidy rate than applying individual rates to each company or using the country-wide average rate for all producers would have, but the court stated:

This "double-counting" is a consequence of the then-existing statutory exception for significantly different rates. It will not occur in investigations initiated after January 1, 1995 because Congress has revised the countervailing duty law, directing Commerce to use either a country-wide average or individual rates, but not both. 19 U.S.C. §§ 1671b(d), 1671d(c)(5) (1994).

Id. at 259. The Crescent I court held that Commerce's methodology was reasonable and that under Ceramica and Ipsco Commerce was required to include all companies, even those receiving significantly different benefits, in calculating the country-wide rate. See id.

Finally, the Kajaria I court determined that Commerce's decision to include a significantly different BIA-based individual rate in calculating the country-wide rate also was acceptable. See 956 F. Supp. at 1026. The court noted that Commerce's BIA methodology is not necessarily punitive, but is used to reflect the current situation as accurately as possible. See id. at 1027. Given the Producers' concession that the choice of BIA used was not unfair as to Super Castings, the court reasoned that the inclusion of Super Castings' individual rate in the country-wide rate did not necessarily result in a more unfavorable country-wide rate than would have resulted if Super Castings had supplied the requested loan information. See id. The court concluded that the Producers had failed to show that the BIA-based rate chosen was not the most accurate estimate of actual current conditions and that pursuant to Ceramica and Ipsco Commerce did not err in including Super Castings' BIA-based rate in calculating the country-wide rate. See id. at 1027-28.B.

On February 24, 1997, Commerce issued remand results. See Final Results of Redetermination on Remand Pursuant to Kajaria Iron Castings Pvt. Ltd. et. al. v. United States, Slip. Op. 97-10 (Ct. Int'l Trade, January 29, 1997) ("Remand Results").6 On remand, Commerce decided not to revise its calculation of the section 80HHC subsidy to account for the CCS over-rebates. See id. at 1. Commerce stated that "[i]n calculating the amount of a countervailable benefit, the Secretary will ignore the secondary tax consequences of the benefit." Id. at 3 (quoting Countervailing Duties; Notice of Proposed Rulemaking and Request for Public Comments, 54 Fed.Reg. 23,383 (May 31, 1989)). Commerce noted that it "would not take into account the secondary effects of [a] rebate program on the calculation of the benefit conferred by an income tax deduction program." See id. at 4. In addressing Argentine Wire Rod, Commerce stated:

If Argentine Wire Rod is interpreted as suggesting that the Department would not investigate and calculate separate benefits for a rebate program and a tax deduction program, then Argentine Wire Rod must be considered an anomaly and not reflective of current Department policy or of Department policy in all of the case precedents that we have found.

See id. Commerce summarized its position as follows:

In our view, there is no double-counting of the CCS subsidy in the case before the Court. The benefit from the CCS program and the benefit from the 80 HHC program are two separate and distinct subsidies. The countervailable portion of the rebates provided to the exporters from the CCS program is considered a grant to the company, and the full amount of the grant constitutes the benefit. The countervailable portion of the 80 HHC program is the amount of taxes on all export sales (both of subject and non-subject merchandise) that is exempted and that otherwise would have been paid absent the tax deduction. Just as the Department does not consider the income tax effect on the amount of a grant to be countervailed (i.e., by deducting from the grant the amount of taxes that may have been due on the grant), it does not consider the secondary effect of other direct subsidy programs on the amount of the tax deduction because both programs provide separate and distinct countervailable benefits....

Id. at 5-6. Commerce determined that exempting the CCS over-rebates from tax, through the section 80HHC deduction, provided an additional benefit in the amount of the tax exemption. See id. at 6.

Commerce also decided not to adjust the section 80HHC deduction to account for the IPRS rebates. See id. at 10. Commerce determined that the section 80HHC subsidy, including the portion based on the IPRS rebates, was "untied" in that it was not particular to specific subject or non-subject merchandise. See id. at 10-11. Therefore, Commerce divided the amount of each producer's section 80HHC deduction by the producer's total exports to allocate the untied benefit between subject and non-subject exports. See id. at 13.

In a June 26, 1997 opinion, the Court of International Trade sustained Commerce's remand results and sustained Commerce's calculation of the section 80HHC subsidy in the final results. See Kajaria II, 969 F. Supp. at 91. The court vacated the portion of its previous decision ordering recalculation of the section 80HHC subsidy. See id. The court sustained Commerce's reasoning "that when a company receives a grant such as a CCS or IPRS payment, and then receives a tax exemption for that grant, it has received two separate benefits which may both be countervailed without double-counting." Id. In reaching that decision, the court relied on its decision of the same day in Crescent Foundry Co. Pvt. Ltd. v. United States, 969 F. Supp. 1341 (CIT 1997) ("Crescent II "). See Kajaria II, 969 F. Supp. at 91.

The court in Crescent II vacated the portion of Crescent I ordering Commerce to recalculate the section 80HHC subsidy and sustained Commerce's final results concerning the section 80HHC subsidy. See 969 F. Supp. at 1345. The court determined that the key to the double-counting issue was whether the CCS over-rebates and the portion of the section 80HHC deduction based on those over-rebates were a single subsidy or two separate subsidies conferring distinct benefits. See id. at 1342. The court held that the two programs provided separate benefits, which were each countervailable. See id. at 1343. The court also reconsidered its prior holding concerning countervailing the portion of the section 80HHC deduction attributable to the IPRS rebates. See id. at 1344. The court determined that the section 80HHC deduction, including the portion based on the IPRS payments, was an untied subsidy and that Commerce had correctly countervailed the subsidy by dividing the total benefit of the subsidy by the total exports of the producer, thereby "diluting" the countervailing duty rate to reflect that some of the benefit was attributable to non-subject merchandise. See id.

The Producers appealed to this court, and we have jurisdiction pursuant to 28 U.S.C. § 1295(a)(5).7

We review a decision of the Court of International Trade affirming or reversing the final results of an administrative review de novo. See Aimcor v. United States, 141 F.3d 1098, 1108 (Fed.Cir.1998). That is, we "apply anew" the Court of International Trade's statutorily-mandated standard of review to the administrative review. See Creswell, 141 F.3d at 1475; Torrington Co. v. United States, 82 F.3d 1039, 1044 (Fed.Cir.1996). Therefore, we uphold Commerce's final results unless they are "unsupported by substantial evidence on the record, or otherwise not in accordance with law." 19 U.S.C. § 1516a(b)(1)(B) (1994). "Substantial evidence is more than a mere scintilla. It means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion." Matsushita Elec. Indus. Co. v. United States, 750 F.2d 927, 933, 3 Fed. Cir. (T) 44, 51 (Fed.Cir.1984) (internal quotations and citations omitted). This court decides de novo the proper interpretation of the governing statutory provisions. See Creswell, 141 F.3d at 1475; Fujitsu Gen. Ltd. v. United States, 88 F.3d 1034, 1038 (Fed.Cir.1996).

On appeal, the Producers argue that Commerce erroneously double counted the subsidy from the CCS over-rebates by countervailing both the over-rebates and the portion of the section 80HHC tax deduction attributable to those over-rebates. The Producers state that they do not seek an offset to the section 80HHC subsidy based on the CCS over-rebates, but instead seek not to have the subsidy based on the CCS over-rebates counted twice. According to the Producers, Commerce countervailed a greater benefit than was actually received. The government responds that the Producers are seeking an offset that does not fit within the allowable offsets set forth in 19 U.S.C. § 1677(6). It urges that the CCS over-rebates and the portion of the section 80HHC deduction based on those over-rebates are separate subsidies which provide distinct countervailable benefits.

It is undisputed that the CCS over-rebates and the section 80HHC deduction both provided countervailable subsidies, as delineated in 19 U.S.C. §§ 1671(a), 1677(5), and that Commerce fully countervailed both subsidies. However, we must determine whether countervailing the full section 80HHC deduction impermissibly double counted the subsidy from the CCS over-rebates. For the following reasons, we agree with the Producers that Commerce's methodology erroneously double counted the subsidy provided by the CCS over-rebates.

Under the Indian tax system, the CCS rebates are considered taxable export income. Therefore, a producer's tax base is increased by the amount of the CCS rebate, and the producer is required to pay tax on the rebate. Absent the section 80HHC deduction, the producer would pay tax on the CCS rebate and would receive a benefit from the rebate equal to the amount of the rebate minus the tax. However, the CCS rebates are also considered export-related income. Under the section 80HHC program, a producer can deduct export profits from taxable income. The inclusion of the CCS rebates in export income raises export profits by an amount no greater than the amount of the CCS rebates. Including the CCS rebates in export income increases the amount of the section 80HHC export profit deduction by no more than the amount of the CCS rebates. Thus, the portion of the section 80HHC deduction attributable to the CCS rebates did no more than offset the inclusion of the CCS rebates in taxable income, thereby eliminating the tax a producer would have paid on the CCS rebate absent the section 80HHC deduction. In effect, the Producers received the CCS rebates tax-free because of the section 80HHC deduction.

The government asserts that the Producers are seeking an offset, based on the countervailing of the CCS over-rebates, to the subsidy provided by the section 80HHC deduction. The government claims that such an offset is not permitted under 19 U.S.C. § 1677(6), which provides an exclusive list of permissible offsets.8 While we agree that 19 U.S.C. § 1677(6) provides the exclusive list of permissible offsets, see Geneva Steel v. United States, 914 F. Supp. 563, 609-10 (CIT 1996); IPSCO, Inc. v. United States, 687 F. Supp. 614, 621-22 (CIT 1988), we do not agree that the Producers are seeking an offset in this case. The Producers seek to avoid having the same subsidy countervailed twice; they do not want Commerce to countervail a greater benefit than that actually received. The Producers simply seek to have the amount of the subsidies they received accurately valued.

The government also argues that the CCS over-rebates and the section 80HHC deduction are distinct subsidies that provide separate benefits. The government claims that there was no double counting under Commerce's methodology. Commerce stated in the remand results:

Because all companies' profits are taxable at the corporate tax rate, any exemption of payment of the corporate tax that is specific (the exemption is for exporters only) constitutes a countervailable subsidy. The amount of the benefit is equal to the amount of the exemption. Countervailable grants may or may not have contributed to the taxable profits, but those grants do not change the amount of the exemption that the government provided, and countervailing the tax exemption and the grants separately does not in any way result in double-counting.

Remand Results, at 8. While we agree with the general proposition that the CCS Program and the section 80HHC program are distinct and provide separate benefits, we cannot accept Commerce's calculations in this case. Countervailing the portion of the section 80HHC deduction attributable to the CCS over-rebates countervails the tax the Producers would have paid on the CCS over-rebates as a result of their inclusion in taxable income. In effect, Commerce fully countervailed the CCS over-rebates and the tax that would have been paid on the over-rebates. However, the Producers only received a benefit equal to the full amount of the CCS over-rebates, which Commerce fully countervailed. Commerce overstated the subsidies received by double-counting the CCS over-rebates.

Commerce's stated policy is that it will not take into account the secondary tax effects of grants. See id. at 3-4 (citing various administrative reviews and Countervailing Duties; Notice of Proposed Rulemaking and Request for Public Comments, 54 Fed.Reg. 23,366, 23,383 (May 31, 1989) (proposed rule 19 C.F.R. § 355.46(b))). This means that when a producer receives a grant and must pay taxes on the grant Commerce will countervail the full amount of the grant, despite the net benefit to the producer being the amount of the grant minus the tax paid on the grant. See id. at 5-6. Commerce stated that "[j]ust because we do not allow firms to discount grants by the amount of the tax payable on the grants, it does not follow that firms that receive grants should be exempted from paying taxes on those grants." Id. at 8. Commerce explained that "all firms are expected to pay taxes[, and] [i]f they do not, they are receiving a benefit that is distinct from any grants they may have received." Id. Commerce's policy of not accounting for the secondary tax consequences of subsidies is inapplicable to the facts of this case, however.

Commerce countervailed the full value of the CCS over-rebates and then countervailed the non-taxation of the CCS subsidy by countervailing the portion of the section 80HHC deduction attributable to the over-rebates. Commerce's policy of discounting secondary tax consequences cannot mean that if a producer receives a subsidy that is taxed Commerce will countervail the pre-tax subsidy, but that if a producer receives a subsidy that is not taxed Commerce will countervail the subsidy and the tax that should have been paid if the subsidy were taxed. The circumstances in this case are akin to the latter situation, in that the Producers in effect received the CCS over-rebates tax-free because of the section 80HHC deduction. It was improper for Commerce to countervail both the CCS over-rebates and the tax that would have been paid on those over-rebates but for the section 80HHC deduction. The reason is that, in so doing, Commerce imposed a countervailing duty that was not "equal to the amount of the net subsidy" in contravention of 19 U.S.C. § 1671(a). In the circumstances of this case, we see no meaningful difference between a tax-free subsidy and a subsidy that is in effect tax-free because of a corresponding tax deduction. We therefore reverse the Court of International Trade's decision sustaining Commerce's final results on this issue.

On remand, Commerce should recalculate the subsidy provided by the section 80HHC deduction in a manner that eliminates the double-counting of the CCS over-rebates. Mindful of the burden on Commerce, our decision does not mean that in every administrative review or investigation Commerce must trace the tax treatment of subsidies to determine if two independent subsidies partially include the same benefit. However, Commerce must avoid double-counting subsidies, i.e., countervailing both the full amount of a subsidy and the nontaxation of that subsidy, when the party under investigation provides documentation that allows Commerce to separate the tax deduction based on the fully countervailed subsidy from the otherwise countervailable portion of the tax deduction.

The Producers also challenge Commerce's decision to countervail the portion of the section 80HHC deduction attributable to the IPRS rebates on non-subject castings, claiming that Commerce acted outside its statutory authority. The Producers argue that they are not seeking an offset to the section 80HHC subsidy, but rather are seeking not to have Commerce countervail the IPRS rebates, which were tied to non-subject castings. For its part, the government claims that the Producers are seeking an impermissible offset and that Commerce correctly allocated the portion of the section 80HHC deduction attributable to the IPRS rebates, as an untied subsidy, over all exports.

As with the CCS over-rebates, IPRS rebates are included in taxable income. Therefore, absent the section 80HHC deduction, the Producers would have paid taxes on the IPRS rebates. Because the IPRS rebates are also considered export revenue, the Producers' export profits increased by no more than the amount of the IPRS rebates. Since export profit is tax deductible under the section 80HHC program, the producers did not pay taxes on the IPRS rebates. In effect, the Producers received the IPRS rebates tax-free because of the section 80HHC deduction. Unlike the CCS over-rebates, the IPRS rebates are not a countervailable subsidy because the rebates were related to non-subject castings. See Preliminary Results, 60 Fed.Reg. at 4599 (IPRS rebates were not received on subject castings during the period of review). However, we must determine whether the portion of the section 80HHC deduction based on the IPRS rebates was a countervailable subsidy.

Commerce determined that the portion of the section 80HHC deduction based on the IPRS rebates was an "untied" subsidy because it "was not particular to only specific merchandise (either subject or non-subject merchandise)." Remand Results, at 10-11. Commerce therefore allocated the benefit of the subsidy over all of the producers' exports in accordance with policy. See id. at 11. We cannot agree with Commerce's decision to treat the portion of the section 80HHC deduction attributable to the IPRS rebates as an untied countervailable subsidy. The portion of the section 80HHC deduction based on the IPRS rebates was not an untied subsidy, but rather a subsidy tied to non-subject castings because the rebates were particular to non-subject castings. The IPRS rebates, coupled with the portion of the section 80HHC deduction based thereon, produced the same results as if the IPRS rebates were a tax-free subsidy on non-subject castings. The entire benefit received, i.e., the subsidy and the nontaxation of the subsidy, was related solely to non-subject castings. It cannot be that a tax-free subsidy on non-subject merchandise produces no countervailable benefit, but that a taxable subsidy on non-subject merchandise coupled with a tax deduction equal to the amount of the tax payable produces an untied countervailable subsidy. We agree that the Producers are not seeking an impermissible offset, but rather are seeking to have a subsidy related to non-subject merchandise eliminated because it was beyond Commerce's statutory authority in this review.

Commerce erred in countervailing the portion of the section 80HHC deduction based on the IPRS rebates because the rebates involved were tied to merchandise not within the scope of the review. Therefore, we reverse the decision of the Court of International Trade sustaining Commerce's final results on this issue. On remand, Commerce should eliminate the IPRS rebates in calculating the subsidy received on subject castings through the section 80HHC deduction. As in the case of the CCS over-rebates and the portion of the section 80HHC deduction based thereon, we are mindful of the government's argument that Commerce does not engage in subsidy tracing because of the burden involved in sorting the tax treatment of subsidies. Again, our decision does not mean that in every review or investigation Commerce must trace the tax treatment of subsidies on non-subject merchandise when a tax deduction results in a countervailable subsidy to determine if the deduction is partially based on the subsidy on non-subject merchandise. However, when the party under investigation provides documentation that allows Commerce to separate the portion of the tax deduction based on rebates related to non-subject merchandise from the remainder of a countervailable tax deduction, Commerce should not countervail the portion of the tax deduction subsidy tied to non-subject merchandise. Since the Producers provided such data, Commerce should eliminate the IPRS rebates from the calculation of the subsidy provided by the section 80HHC deduction.

After calculating the subsidy rate of each producer, Commerce calculated the country-wide countervailing duty rate by weight averaging the subsidy rate of each producer, using as the weight the producer's share of total exports to the United States. See Final Results, 60 Fed.Reg. at 44,843. Commerce then assigned individual countervailing duty rates to the producers receiving significantly different net subsidies. See id. The Producers argue that this method of calculating the country-wide countervailing duty rate conflicts with United States law, conflicts with international obligations, and violates substantive due process by depriving United States importers of property. The Producers assert that Commerce erred in including producers that were assigned significantly different rates in calculating the country-wide rate. The government responds that Commerce is required to include all producers in calculating the country-wide rate and that Commerce's methodology was a reasonable interpretation of the governing statutory provisions.

Pursuant to 19 U.S.C. § 1671e, after Commerce determines that a net subsidy exists, it "shall publish a countervailing duty order which ... shall presumptively apply to all merchandise of such class or kind exported from the country investigated, except that if ... [Commerce] determines there is a significant differential between companies receiving subsidy benefits ... the order may provide for differing countervailing duties." 19 U.S.C. § 1671e(a)(2)(A). While the statute expresses a presumption in favor of country-wide countervailing duty rates, see Ipsco, Inc. v. United States, 899 F.2d 1192, 1197, 8 Fed. Cir. (T) 80, 86 (Fed.Cir.1990); Saha Thai Steel Pipe Co. v. United States, 828 F. Supp. 57, 64 (CIT 1993), it permits differing countervailing duty rates when companies receive significantly different subsidy benefits, see 19 U.S.C. § 1671e(a)(2)(A). The regulations also require Commerce to specify the net subsidy on a country-wide basis. See 19 C.F.R. § 355.22(c)(7)(ii) (1992). However, Commerce may assign producers receiving significantly different subsidies individual countervailing duty rates:

(1) If a producer or exporter is government-owned, the Secretary will, and to the extent practicable for other producers or exporters the Secretary may, review whether a significant differential existed, during the period under review, between the net subsidy received by an individual producer or exporter of the merchandise and the weighted-average net subsidy calculated on a country-wide basis.

(2) If the Secretary decides that an individual (including government-owned) producer or exporter received a significantly different net subsidy during the period, the Secretary will state in the final results an individual rate for that person, and that rate will be the basis for the assessment of countervailing duties....

(3) A significant differential is:

(i) A difference of the greater of at least five percentage points, or 25 percent, from the weighted-average net subsidy calculated on a country-wide basis; or

(ii) The difference between a net subsidy of zero (or de minimis ) and any rate greater than de minimis.

The regulations require Commerce to determine, to the extent practicable, if a producer received a significantly different net subsidy, by comparing the net subsidy received by the producer with the weighted-average net subsidy calculated on a country-wide basis. See 19 C.F.R. § 355.22(d)(1). This comparison is not possible unless Commerce first calculates a country-wide rate based on all companies investigated. The Producers do not challenge the assignment of individual, significantly different rates to Dinesh, Super Castings, and Kajaria, but they claim that Commerce should have recalculated the country-wide rate excluding the significantly different higher rates of Super Castings and Kajaria. We find no support in the statute or the regulations for such an approach.

In IPSCO, this court rejected a country-wide rate that excluded companies receiving no benefit or a de minimis benefit. See 899 F.2d at 1194, 8 Fed. Cir. (T) at 82-83. The court stated:

Where the overall level of subsidization provided to a particular industry by a foreign government is de minimis, no countervailing duty should be assessed. And, if there is a non-de minimis subsidy being provided, the countervailing duty should not exceed the weighted-average benefit received by all firms that produce or export the subject goods, including those firms that receive little or no subsidy.

Id. at 1197, 899 F.2d 1192, 8 Fed. Cir. (T) at 86 (emphasis added). In Ceramica Regiomontana, S.A. v. United States, 16 Ct. Int'l Trade 358, 1992 WL 107338 (1992), the Court of International Trade rejected Commerce's methodology as inconsistent with IPSCO. See 16 C.I.T at 365-66. In Ceramica, Commerce had calculated a country-wide rate based on all companies being reviewed and then had calculated individual company rates and compared them to the country-wide rate, to determine if the individual rates were significantly different from the country-wide rate. See id. at 362-63. Commerce assigned individual rates to companies with significantly different net subsidies and then calculated an "all other" rate based on the remaining companies and applied this "all other" rate as the country-wide rate. See id. at 363. The court noted that "[t]he 'all other' rate is different from a country-wide rate because it is not based upon all companies, but, rather, those companies being reviewed whose rates are not significantly different from the country-wide rate." Id. (internal quotations omitted). The court held that "it is anomalous for [Commerce] to first determine that there is no significant differential between the net subsidy rates for 'all other' companies and the country-wide rate, and then assess countervailing duties at a rate different from the country-wide rate." See id. at 365-66. In Ceramica Regiomontana, S.A. v. United States, 853 F. Supp. 431 (CIT 1994), rev'd on other grounds, 64 F.3d 1579 (Fed.Cir.1995), the court rejected Commerce's methodology of calculating the country-wide rate "by weight averaging the benefits received by non-de minimis companies by their proportion of United States exports." 853 F. Supp. at 439 (footnote omitted). The court held that this methodology was inconsistent with IPSCO because Commerce must calculate the country-wide rate based on all companies, even those with zero and de minimis subsidy rates. See id.

While these cases expressly require Commerce to include zero and de minimis rates in calculating the country-wide rate, the cases also support Commerce's methodology in this case. Just as it would be inconsistent to disregard significantly different lower rates in the calculation of the country-wide rate, it would be inconsistent to disregard significantly different higher rates in the calculation, which is precisely what the Producers seek. See Ceramica, 16 C.I.T. at 365-66 (rejecting recalculation of the country-wide rate after excluding companies with significantly different net subsidies). As stated in IPSCO, "the countervailing duty should not exceed the weighted-average benefit received by all firms that produce or export the subject goods." 899 F.2d at 1197, 8 Fed. Cir. (T) at 86 (emphasis added). The court stated that "[t]he country-wide countervailing duty rate that applies to imports from the investigated country must bear some relation to the approximate average rate of subsidization of the subject goods." Id. Just as including zero and de minimis rates leads to a more accurate country-wide rate, the same is true for including significantly different higher rates.

The countervailing duty imposed by Commerce must be "equal to the amount of the net subsidy." 19 U.S.C. § 1671(a). A countervailing duty order must assess "a countervailing duty equal to the amount of the net subsidy determined or estimated to exist." 19 U.S.C. § 1671e(a)(1). We reject the Producers' claim that Commerce's methodology assessed a countervailing duty greater than the net subsidy. Commerce's methodology, i.e., including all producers' rates in determining the country-wide rate, is a reasonable interpretation of the statute and leads to the most accurate country-wide rate. See British Steel PLC v. United States, 127 F.3d 1471, 1475 (Fed.Cir.1997) ("Courts must sustain an agency's interpretation of a statute if it falls within the range of permissible construction." (internal quotations omitted)); Ipsco, 899 F.2d at 1194-95, 8 Fed. Cir. (T) at 83 ("We give due weight to the agency's interpretation of the statute it administers, and we accept that interpretation if it is 'sufficiently reasonable.' " (citations omitted)). We find no conflict between Commerce's approach and the requirement that the countervailing duty equal the net subsidy.

The Producers also argue that Commerce's methodology conflicts with The Agreement on Interpretation and Application of Articles VI, XVI, and XXIII of the General Agreement on Tariffs and Trade,10 which provides in Article 4.2 that "[n]o countervailing duty shall be levied on any imported product in excess of the amount of the subsidy found to exist, calculated in terms of subsidization per unit of the subsidized and exported product." (footnotes omitted). This is substantially the same requirement set forth in 19 U.S.C. §§ 1671(a), 1671(e). Commerce's methodology assessed a countervailing duty on the imported castings based on the overall level of subsidization determined to exist. Commerce countervailed the imported castings at the overall rate of subsidization determined based on all producers. Commerce's methodology comported with the cited provision, and we therefore reject the Producers' claim that Commerce's methodology conflicts with international obligations. We also find the Producers' claim that Commerce's methodology denies United States importers substantive due process equally unavailing.

While Commerce's methodology is based on a reasonable interpretation of 19 U.S.C. § 1671e(a)(2) and was proper in this case, we note that the issue presented in this case will not arise under the revised statutory provisions. The revised statute requires Commerce to "determine an individual countervailable subsidy rate for each known exporter or producer of the subject merchandise." 19 U.S.C. § 1677f-1(e)(1) (1994). Under the revised statutes, Commerce may only use a country-wide rate when it is impracticable to use individual rates, and in such a case the country-wide rate will apply to all exporters and producers. See 19 U.S.C. §§ 1677f-1(e)(2), 1671d(c)(5)(B) (1994).11

Finally, the Producers challenge Commerce's inclusion of Super Castings' subsidy rate, based on BIA, in calculating the country-wide rate. They claim that it was improper for Commerce to include Super Castings' BIA-based rate in the country-wide rate because it was a significantly different higher rate and because it unfairly assigned a BIA component to cooperating producers. The government argues that BIA-based rates are not necessarily punitive and that it was proper for Commerce to include Super Castings' BIA-based rate in the country-wide rate because it is presumed to be the most accurate rate available.

Pursuant to 19 U.S.C. § 1677e(c), Commerce may use the "best information otherwise available" "whenever a party or any other person refuses or is unable to produce information requested in a timely manner and in the form required." 19 U.S.C. § 1677e(c); see Cemex, S.A. v. United States, 133 F.3d 897, 904 (Fed.Cir.1998). As a result of Super Castings' failure to respond to questionnaires regarding pre- and post-shipment loans, Commerce assumed, as BIA, that all reported loans were pre-shipment loans, which were subject to a higher interest rate. See Preliminary Results, 60 Fed.Reg. at 4597. The Producers do not dispute the propriety of the BIA assumption used to calculate Super Castings' subsidy rate.

The issue before us is whether Commerce erred in including Super Castings' BIA-based rate in calculating the country-wide rate. To the extent that the Producers' argument against including Super Castings' rate in the country-wide rate is based on their claim that significantly different higher rates cannot be included in the country-wide rate, we need not further address this argument, which is fully discussed above.

We also reject the Producers' claim that by including Super Castings' rate in the country-wide rate Commerce unfairly assigned an inflated BIA component to cooperating producers. This court has stated that the use of BIA is not punitive and often leads to the most accurate approximation of the situation. See Rhone Poulenc, Inc. v. United States, 899 F.2d 1185, 1190-91, 8 Fed. Cir. (T) 61, 67 (Fed.Cir.1990); see also Floral Trade Council v. United States, 799 F. Supp. 116, 120 (CIT 1992) ("BIA may be very close to reality."). Commerce's use of BIA is discretionary, and its methodology need only be consistent with its statutory authority. See Allied-Signal Aerospace Co. v. United States, 28 F.3d 1188, 1191 (Fed.Cir.1994). In Saha Thai, the court upheld Commerce's decision to apply a country-wide countervailing duty rate, based in part on BIA, to a cooperating company that had been shown to have a lower, but not significantly different, subsidy rate. See 828 F. Supp. at 59, 65. While the court suggested that Commerce should apply the lower individual rate to the cooperating company, the court held that Commerce's decision to apply the country-wide rate, based on BIA, to the company was within the law and the regulations. See id. at 65.

Commerce's decision in the present case is similar to that of Saha Thai in that the parties under investigation were assigned a country-wide rate, based on BIA, which was higher than their individual subsidy rates. However, as did the court in Saha Thai, we find no error in Commerce's decision to include the BIA-based rate in determining the country-wide rate. Section 1671e(a) of Title 19 expresses a preference for a country-wide rate, and Commerce's methodology for calculating this rate was not contrary to the statute. No statutory provision or regulation suggests that BIA-based rates should be excluded from the country-wide rate. Indeed, 19 U.S.C. § 1671e(a) directs Commerce to assess a countervailing duty, which "shall presumptively apply to all merchandise of [the] class" investigated, that is "equal to the amount of the net subsidy determined or estimated to exist." 19 U.S.C. § 1671e(a). The use of BIA is an acceptable method of estimating net subsidy rates, and we see nothing that prohibits Commerce from including such rates in the calculation of the country-wide rate. We therefore sustain Commerce's inclusion of Super Castings' subsidy rate in the country-wide rate.

Because Commerce's methodology double counted the subsidies the Producers received from the CCS over-rebates, by countervailing both the over-rebates and the section 80HHC deduction attributable to those over-rebates, we reverse the Court of International Trade's decision sustaining Commerce's results on this issue. We also reverse the court's decision sustaining Commerce's decision to countervail the portion of the section 80HHC deduction attributable to the IPRS rebates on non-subject castings, because Commerce acted beyond its statutory authority in countervailing benefits received on non-subject merchandise. On these two issues, we remand to the Court of International Trade for further proceedings consistent with this opinion. However, we uphold the court's decision sustaining Commerce's methodology for calculating the country-wide countervailing duty rate, i.e., including significantly different higher rates and BIA-based rates in calculating the country-wide rate, because Commerce's methodology was in accordance with law and a permissible interpretation of the governing statutory provisions.

Each party shall bear its own costs.

AFFIRMED-IN-PART, REVERSED-IN-PART, and REMANDED.

RADER, Circuit Judge, dissenting-in-part.

Because one aspect of Commerce's methodology in this case imposes on the producers a countervailing duty in excess of any subsidy, I must respectfully dissent. Commerce must publish a rate that "shall presumptively apply to all merchandise of such class or kind exported from the country investigated, except that ... if there is a significant differential between companies receiving subsidy benefits ... the order may provide for differing countervailing duties." 19 U.S.C. § 1671e(a)(2)(A). However, title 19 only authorizes Commerce to impose a countervailing duty that is "equal to the amount of the net subsidy." 19 U.S.C. § 1671(a) (emphasis added). The statute therefore sets a limit on the amount of duty Commerce may collect. In this case, Commerce's presumptive rate imposes duties in excess of the net subsidy. Therefore, Commerce's regulations, as applied in this case, are an impermissible application of the statute.

In calculating a presumptive rate of 5.53%, Commerce included the net subsidy rates of three firms (Super Castings' 41.75% rate, Kajaria's 16.14% rate, and Dinesh Brothers' 0% rate) that Commerce had determined had received significantly different subsidy benefits. In its order, Commerce assigned individual rates to the three significantly different firms and assigned the 5.53% presumptive rate to the other producers. Because, however, it included the significantly different rates in the calculation of the presumptive rate, Commerce imposed countervailing duties not equal to the amount of the subsidy. Rather, Commerce's duties rise to approximately 130% of the net subsidy actually received from the government of India by the producers. Because the statute limits the duty to an amount equal to the subsidy, Commerce's calculation of the presumptive rate is not reasonable.

Commerce cites Ipsco, Inc. v. United States, 899 F.2d 1192 (Fed.Cir.1990), as support for the averaging method used to calculate the presumptive rate in this case. In Ipsco, this court rejected Commerce's presumptive rate because it excluded producers receiving no subsidy or only a de minimis subsidy. See id. at 1194. If this court had allowed Commerce to exclude producers with de minimis subsidization rates from the calculation, the resulting presumptive rate would likely have exceeded the actual net subsidy. This court stated that Commerce must calculate a presumptive rate that "bear[s] some relation to the approximate average rate of subsidization of the subject goods." Id. at 1197. Therefore, this court reversed Commerce's calculation in Ipsco.

Commerce argues that, because Ipsco requires it to include significantly lower rates in its calculations, it must also include significantly higher rates. I would agree-for the facts in Ipsco. This case, however, is not Ipsco. Here, subsidy rates are known for every producer potentially affected by the countervailing duty order. Therefore, the actual net subsidy can be determined. In this case, Commerce's methodology, which is usually reasonable when the actual net subsidy is not known, results in the imposition of countervailing duty that is not equal to the net subsidy. Therefore, the principles of the statute and this court's Ipsco opinion require a different calculation. The statute admittedly gives Commerce some latitude in choosing a method to calculate rates. It does not, however, excuse Commerce from the requirement that the countervailing duty must be "equal to the amount of the net subsidy." 19 U.S.C. § 1671(a).

With respect to the other issues in this case, I join the court in stating that, under these circumstances, Commerce calculated the subsidy from the section 80HHC tax deduction in a manner that is inconsistent with the controlling statutes. Because a portion of the section 80HHC deduction is attributable to the CCS over-rebates and to the IPRS rebates, and because the Producers have identified the amounts so attributable, Commerce must take those amounts into account in calculating the "subsidy." See 19 U.S.C. § 1677(5). Commerce's failure to take those amounts into account results in the imposition of countervailing duties that exceed the net subsidy actually provided and is therefore contrary to the statutory mandate. I therefore join parts I, II, and III of the opinion. It is also well established that Commerce may use the best information available to calculate the subsidy received by an uncooperative producer. Therefore, I join part V of the opinion. I dissent from part IV to emphasize the statutory limit on Commerce's power to impose countervailing duties.