C-351-835
Investigation
Public Document
III/VII: DM, SC, HH

MEMORANDUM

DATE: September 23, 2002
TO: Faryar Shirzad
Assistant Secretary
for Import Administration

FROM: Joseph A. Spetrini
Deputy Assistant Secretary
for AD/CVD Enforcement III

SUBJECT: Issues and Decision Memorandum for the Final Affirmative Countervailing Duty Determination in the Countervailing Duty Investigation of Certain Cold-Rolled Carbon Steel Flat Products from Brazil

Background

On March 4, 2002, the Department of Commerce (“the Department”) published the preliminary determination in this investigation. See Preliminary Affirmative Countervailing Duty Determination and Alignment with Final Antidumping Duty Determinations: Certain Cold- Rolled Carbon Steel Flat Products from Brazil, 67 FR 9652 (March 4, 2002) (Brazil Preliminary Determination). The “Analysis of Programs” and “Subsidies Valuation Information” sections below describe the subsidy programs and the methodologies used to calculate the benefits from these programs. We have analyzed the comments submitted by the interested parties in their case and rebuttal briefs in the “Analysis of Comments” section below. Our review of the comments has led us to change certain aspects of our preliminary determination. All changes are set forth in either the “Analysis of Programs” section or the “Analysis of Comments” section. We recommend that you approve the positions we have developed in this memorandum. Below is a complete list of the issues in this investigation for which we received comments and rebuttal comments from parties:

Comment 1: CSN, USIMINAS and COSIPA Privatization
Comment 2: PIS/COFINS - Direct Taxes v. Indirect Taxes
Comment 3: PIS/COFINS- Rebate of Prior-Stage Cumulative Indirect Taxes
Comment 4: PIS/COFINS Credit - Excessive Remission
Comment 5: FINEM Financing and Specificity
Comment 6: FINAME as an Import Substitution Program
Comment 7: FINAME Financing and Specificity
Comment 8: Integral Linkage of FINAME and FINEM
Comment 9: Financial Contribution and Benefit of BNDES Loan Programs
Comment 10: BNDES-ExIm Financing and Specificity
Comment 11: BNDESPAR Program
Comment 12: PRO-Industria- Specificity
Comment 13: Non-Use of PROEX

Company Histories

USIMINAS

As stated in the Final Affirmative Countervailing Duty Determination: Certain Cold Rolled Flat-Rolled Carbon-Quality Steel Products From Brazil, 65 FR 5536 (February 4, 2000) (Brazil Cold-Rolled Final), Usinas Siderurgicas de Minas Gerais ("USIMINAS") was founded in 1956 as a venture between the Government of Brazil (GOB), various stockholders and Nippon USIMINAS. In 1974, the majority interest in USIMINAS was transferred to SIDERBRAS, the government holding company for steel interests. The company underwent several expansions of capacity throughout the 1980s. In 1990, SIDERBRAS was put into liquidation and the GOB included SIDERBRAS’ operating companies, including USIMINAS, in its National Privatization Program (NPP). In October 1991, the privatization of USIMINAS occurred with the auction of shares to Nippon USIMINAS, Employees of USIMINAS, Companhia do Vale Rio Doce (CVRD), Previ-BB Pension Fund, Valia-CVRD Pension Fund, Bozano Simonson, Banco Economico, and other financial institutions. In 1994, the GOB disposed of additional holdings.

USIMINAS is now owned by a group of private investors, including Nippon USIMINAS with 18.39 percent of shares, Caixa de Previdencia dos Funcionarios do Banco do Brasil (Previ) (the pension fund of the Bank of Brazil) with 14.9 percent of shares, CVRD, with 22.99 percent of shares, and the USIMINAS Employee Investment Club with 9.94 percent of shares.

In January 1999, a project was implemented for the corporate, financial, equity, and operational restructuring of USIMINAS and Companhia Siderurgica Paulista (COSIPA). The result of this project was the reallocation of assets and liabilities between the two companies. According to the questionnaire responses, one result of this restructuring was a slight change in USIMINAS’s shareholdings in COSIPA, to 49.77 percent from 49.8 percent in January 1999. Another result of the restructuring was the subscription by USIMINAS to 892 million reais in convertible debentures issued by COSIPA. These debentures are not redeemable. They are convertible on demand, at a fixed price, in groups of three, to one common (voting) and two preferred shares. As of the end of the POI, USIMINAS had not converted any of these debentures to shareholdings.

One of USIMINAS’s minority shareholders is CVRD, one of the world’s largest producers of iron ore. CVRD also owns stock in Companhia Siderurgica Nacional (“CSN”). However, we have determined that CVRD does not exercise direct or indirect control of either USIMINAS or CSN. See “Cross-Ownership and Attribution of Subsidies” section below, for a complete analysis of the extent of CVRD's control over USIMINAS and CSN.

COSIPA

COSIPA was established in 1953 as a government-owned steel production company. In 1974, COSIPA was transferred to SIDERBRAS. Like USIMINAS, COSIPA was included in the NPP after SIDERBRAS was put into liquidation. In 1993, the privatization of COSIPA occurred with the sale of a majority of the shares to a consortium of investors led by USIMINAS. In 1994, additional government-held shares were sold. During the POI, USIMINAS owned 49.77 percent of the voting capital stock of the company. Other principal owners include Bozano Simonsen Asset Management, Ltd.; the COSIPA Employee Investment Club; and COSIPA's Pension Fund (FEMCO). See Brazil Cold Rolled Final, 65 FR at 5544. The President of USIMINAS is a member of COSIPA's administrative council, which operates similarly to a board of directors. As discussed in the history of USIMINAS above, COSIPA and USIMINAS underwent a major corporate restructuring in January 1999, resulting in the reallocation of assets and liabilities between the two companies and the subscription by USIMINAS to 892 million reais in convertible debentures issued by COSIPA.

CSN

CSN was established in 1941 and commenced operations in 1946 as a government-owned steel company. In 1974, CSN was transferred to SIDERBRAS. In 1990, when SIDERBRAS was put into liquidation, the GOB included CSN in its NPP. In 1993, a majority of CSN’s shares were sold to private investors. At that time, CVRD, through its subsidiary Vale do Rio Doce Navegacao, S.A. (Docenave/CVRD), also acquired 9.4 percent of the common shares. The GOB's remaining share of the firm was sold in 1994. CSN’s shareholders during the POI were Vicunha Siderurgia, with 46.48 percent of the voting shares; Previ, with 13.85 percent; Docepar/CVRD (formerly known as Docenave/CVRD), with 10.33 percent; and a consortium of private investors, including Uniao Comercio e Partipacoes, Ltda.; Textilia, S.A.; the CSN Employee Investment Club; and the CSN employee pension fund. CSN was part of the group that acquired control of CVRD at the time of CVRD’s privatization in 1997. See Brazil Cold- Rolled Final, 65 FR at 5544.

Subsidies Valuation Information

Allocation Period

Section 351.524(d)(2) of the Department's regulations states that we will presume the allocation period for non-recurring subsidies to be the average useful life (AUL) of renewable physical assets for the industry concerned, as listed in the Internal Revenue Service's (IRS) 1977 Class Life Asset Depreciation Range System, as updated by the Department of Treasury. The presumption will apply unless a party claims and establishes that these tables do not reasonably reflect the AUL of the renewable physical assets for the company or industry under investigation, and the party can establish that the difference between the company-specific or country-wide AUL for the industry under investigation and the AUL from the IRS tables is significant.

In the preliminary determination, the Department used an allocation period of 15 years, which is the AUL listed in the IRS tables for the steel industry. See Brazil Preliminary Determination, 67 FR at 9654. No party provided any information or argument concerning the use of the IRS tables. Therefore, for purposes of this final determination, we are continuing to use the 15-year AUL as reported in the IRS tables to allocate any non-recurring subsidies under investigation which were provided to CSN, USIMINAS, and COSIPA.

Cross Ownership and Attribution of Subsidies

There are three producers/exporters of the subject merchandise under investigation: USIMINAS, COSIPA, and CSN. As discussed above, during the POI, USIMINAS owned 49.77 percent of COSIPA. The Department’s regulations, at section 351.525(b)(6)(ii), provide guidance with respect to the attribution of subsidies between or among companies which have cross-ownership. Specifically, with respect to two or more corporations producing the subject merchandise which have cross-ownership, the regulations direct us to attribute the subsidies received by either or both corporations to the products manufactured by both corporations.

Further, section 351.525(b)(6)(vi) defines cross-ownership as existing “between two or more corporations where one corporation can use or direct the individual assets of the other corporation(s) in essentially the same ways it can use its own assets. Normally, this standard will be met where there is a majority voting ownership interest between two corporations through common ownership of two (or more) corporations.” The preamble to the Department’s regulations identifies situations where cross-ownership may exist even though there is less than a majority voting interest between two corporations: “{i}n certain circumstances, a large minority interest (for example, 40 percent) or a ‘golden share’ may also result in cross-ownership.” See Department of Commerce, International Trade Administration, Countervailing Duties; Final Rule, 63 FR 63548, 65401 (November 25, 1998) (Preamble).

In this investigation, we preliminarily determined that USIMINAS’s 49.77 percent ownership interest in COSIPA was sufficient to establish cross-ownership between the two companies because USIMINAS was capable of using or directing the individual assets of COSIPA in essentially the same ways it can use its own assets. In the Brazil Cold-Rolled Final, 65 FR at 5544, we found that USIMINAS’s 49.8 percent shareholding, given the distribution of shareholdings among the remaining shareholders, was sufficient to establish cross ownership of the two companies and attribution of the two companies’ subsidies to both companies.

In the instant investigation, we determine that USIMINAS’s shareholding, at 49.77 percent, is sufficient to establish control since USIMINAS is by far the largest shareholder and no other shareholder is in a position to exert control. In addition, when taken together with the COSIPA convertible debentures that USIMINAS holds, we find that USIMINAS effectively held a majority interest in COSIPA during the POI. This satisfies the definition of cross-ownership provided in section 351.525(b)(6)(vi) of the regulations. Therefore, we determine that USIMINAS’s virtual majority share in COSIPA, and the COSIPA debentures held by USIMINAS that are not redeemable and are convertible to shares in COSIPA, establish crossownership between USIMINAS and COSIPA. Thus, we will continue to calculate one subsidy rate for USIMINAS/COSIPA. For all domestic subsidies, we will follow the methodology outlined in section 351.525(b)(6)(ii) of the regulations. In the case of export subsidies for USIMINAS/COSIPA, we will determine the countervailable subsidy by following the methodology outlined in sections 351.525(b)(2) and 351.525(b)(6)(ii) of the regulations.

In the Brazil Cold-Rolled Final, the Department also examined the ownership of CSN. We note that, in the instant investigation, the same two entities, CVRD and Previ, that were found to have minority shareholdings in USIMINAS and CSN in the Brazil Cold-Rolled Final, still have minority holdings in both USIMINAS and CSN. 65 FR at 5544. As these entities both have ownership interests in and elect members to the Boards of Directors of both companies, we examined whether CSN and USIMINAS could, notwithstanding the absence of direct crossownership between them, have cross-ownership such that their interests are merged, and one company could have the ability to use or direct the assets of the other through their common investors. Since the Brazil Cold-Rolled Final, CVRD’s common shares in USIMINAS have increased from 15.48 percent to 22.99 percent, while its common shares in CSN, through its wholly-owned subsidiary Docepar/CVRD, remained unchanged at 10.33 percent at the end of the POI. For this same period, Previ’s holdings of common shares in USIMINAS fell slightly from 15 percent to 14.90 percent, and remained unchanged for its holdings in CSN at 13.85 percent. Previ is the second largest shareholder in USIMINAS.

As noted in the Brazil Cold-Rolled Final, both USIMINAS and CSN are controlled through shareholders’ agreements which require participating shareholders (who together account for more than 50 percent of the shares of the company) to pre-vote issues before the Board of Directors and to vote as a block. 65 FR at 5544. While CVRD and Previ both participate in the CSN shareholders’ agreement, and thus exercise considerable influence over the use of CSN’s assets, neither CVRD nor Previ participates in the USIMINAS shareholders’ agreement, and therefore, neither is in a position to exercise any appreciable influence (beyond their respective 22.99 and 14.90 percent USIMINAS shareholdings) over the use of USIMINAS’s assets. See Final Affirmative Countervailing Duty Determination: Certain Hot-Rolled Flat-Rolled Carbon- Quality Steel Products from Brazil, 64 FR 38741, 38744 (July 19, 1999) (Brazil Hot-Rolled Final), which noted the Department’s verification of USIMINAS’s shareholder agreement.

No new information has been submitted on the record of this investigation to indicate any changes in the terms of USIMINAS’s shareholders’ agreement since the Department’s verification in the Brazil Hot-Rolled Final. Although CVRD has increased its holdings in USIMINAS, there are also other significant shareholders in USIMINAS. Thus, CVRD is not in a position to control USIMINAS. Therefore, consistent with our finding in the Brazil Cold-Rolled Final and the Brazil Hot-Rolled Final, we determine that CVRD’s and Previ’s shareholdings in both USIMINAS and CSN are not sufficient to establish cross-ownership between those two companies under our regulatory standard. This absence of common majority or significant minority shareholders leads us to determine that USIMINAS’s and CSN’s interests have not merged, i.e., one company is not able to use or direct the individual assets of the other as though the assets were their own. Thus, for the purposes of this determination, we have calculated a separate countervailing duty rate for CSN.

Equityworthiness

In accordance with section 351.507(a)(1) of the Department’s regulations, a government provided equity infusion confers a benefit to the extent that the investment decision is inconsistent with the usual investment practice of private investors, including the practice regarding the provision of risk capital, in the country in which the equity infusion is made. See also section 771(5)(E)(i) of the Act. In past investigations, we determined that COSIPA was unequityworthy from 1977 through 1989, and 1992 through 1993; USIMINAS was unequityworthy from 1980 through 1988; and CSN was unequityworthy from 1977 through 1992. See Brazil Cold-Rolled Final, 65 FR at 5545, citing to Affirmative Countervailing Duty Determinations: Certain Steel Products from Brazil, 58 FR 37295, 37297 (July 9, 1993) (Brazil Certain Steel Final); Brazil Hot-Rolled Final, 64 FR at 38746. We note that, because the Department determined that it is appropriate to use a 15-year allocation period for non-recurring subsidies, equity infusions provided prior to 1986 no longer provide benefits in the POI. None of the parties have submitted information or argument, nor is there evidence of changed circumstances, which would cause us to reconsider these determinations.

Equity Methodology

Section 351.507(a)(3) of the Department’s regulations provides that a determination that a firm is unequityworthy constitutes a determination that the equity infusion was inconsistent with usual investment practices of private investors. The applicable methodology is described in section 351.507(a)(6) of the regulations. The Department will treat the equity infusion as a grant for equity infusions into an unequityworthy company based on the premise that an unequityworthiness finding by the Department is equivalent to saying that the company could not have attracted investment capital from a reasonable investor in the infusion year based on the available information. See Preamble, 63 FR at 65375. For purposes of this investigation, we have converted all equity infusion and swap amounts to U.S. dollars using the exchange rate information used in the final determination of the prior investigation. See Brazil Cold-Rolled Final, 65 FR at 5547.

Creditworthiness

To determine whether a company is uncreditworthy, the Department must examine whether the firm could have obtained long-term loans from conventional commercial sources based on information available at the time of the government-provided loan. See section 351.505(a)(4) of the Department’s regulations. In this context, the term "commercial" refers to bank loans from a commercial lending institution. See section 351.505(a)(2)(ii) of the Department’s regulations.

The Department has previously determined that respondents were uncreditworthy in the following years: USIMINAS, 1983-1988; COSIPA, 1983-1989 and 1991-1993; and CSN 1983- 1992. See Brazil Cold-Rolled Final, 65 FR at 5546, citing to Brazil Certain Steel Final, 58 FR at 37298 and Brazil Hot-Rolled Final, 64 FR at 38747. No new information or evidence of changed circumstances has been presented in this investigation that would lead us to reconsider these findings.

Discount Rates

From 1984 through 1994, Brazil experienced persistent high inflation. There were no long-term fixed-rate commercial loans made in domestic currencies during those years that could be used as discount rates. As in the Brazil Certain Steel Final, 58 FR at 37298, the Brazil Hot- Rolled Final, 64 FR at 38745-38746 and the Brazil Cold-Rolled Final, 65 FR at 5546, we have determined that the most reasonable way to account for the high inflation in the Brazilian economy through 1994, given the lack of an appropriate Brazilian currency discount rate, is to convert values of the equity infusions provided in Brazilian currency into U.S. dollars. If the date of receipt of the equity infusion was provided, we applied the exchange rate applicable on the day the subsidies were received, or, if that date was unavailable, the average exchange rate in the month the subsidies were received. Then we applied, as the discount rate, a long-term dollar lending rate in Brazil. Therefore, for our discount rate, we used data for U.S. dollar lendings in Brazil for long-term, non-guaranteed loans from private lenders, as published in the World Bank Debt Tables: External Finance for Developing Countries. This conforms with the methodology applied in the Brazil Certain Steel Final, and in the Brazil Hot-Rolled Final.

As discussed above, we determine that USIMINAS, COSIPA, and CSN were uncreditworthy in all the years in which they received equity infusions. Section 351.505 (a)(3)(iii) of the Department’s regulations directs us regarding the calculation of the benchmark interest rate for purposes of calculating the benefits for uncreditworthy companies: to calculate the appropriate rate for uncreditworthy companies, the Department must identify values for the probability of default by uncreditworthy and creditworthy companies. For the probability of default by an uncreditworthy company, we normally rely on the average cumulative default rates reported for the Caa to C-rated category of companies as published in Moody’s Investors Service, Historical Default Rates of Corporate Bond Issuers, 1920 - 1997 (February 1998). See 19 CFR 351.505(a)(3)(iii). For the probability of default by a creditworthy company, we used the cumulative default rates for Investment Grade bonds as reported by Moody's. We established that this figure represents a weighted average of the cumulative default rates for Aaa to Baa-rated companies. The use of the weighted average is appropriate because the data reported by Moody’s for the Caa to C-rated companies are also weighted averages. For non-recurring subsidies, we used the average cumulative default rates for both uncreditworthy and creditworthy companies based on a 15-year term, because all of the non-recurring subsidies examined were allocated over a 15-year period.

Benchmarks for Loans

a. Long-Term Benchmarks

There are three BNDES loan programs determined to be countervailable: Financing for the Acquisition or Lease of Machinery and Equipment through the Special Agency for Industrial Financing (FINAME); BNDES-ExIm; and BNDES Participacoes S.A. (BNDESPAR). All of these programs provide long-term loans at variable interest rates. Some of these loans are denominated in reais and some in U.S. dollars. Pursuant to section 351.505(a)(5) of the Department’s regulations regarding long-term variable rate loans, we will use comparable long-term variable-rate commercial loans approved in the same year, when available, as the loan benchmark. Section 351.505(a)(2) of the Department’s regulations defines a comparable commercial loan as one that, when compared to the government-provided loan in question, has similarities in the structure of the loan (e.g., fixed or variable interest rate), the maturity of the loan (e.g., short- or long-term), and the currency in which the loan is denominated.

All of the long-term, variable rate commercial lending reported by respondent companies is denominated in U.S. dollars rather than reais. We were unable to find any information on fixed or variable-rate long-term commercial financing in reais. Therefore, where available, we are using long-term variable rate commercial loans denominated in dollars as benchmarks for both real- and dollar-denominated loans because they satisfy two of the three requirements set forth under section 351.505(a)(2) of the Department’s regulations concerning the structure and the maturity of the loan (i.e. variable rate and long-term). However, in reviewing the annual reports and financial statements of respondent companies, we found that in addition to interest payments, these companies are also subject to exchange rate variations when using dollar financing. See exhibits 2, 5 and 7 of respondent companies’ June 3, 2002 submission and relevant notes in the loan and financing section of their annual reports which identify an interest spread plus an exchange rate variation in their costs for dollar financing. Therefore, we are adjusting the benchmark to reflect this additional cost associated with the exchange rate variation.

We have adjusted these dollar-denominated benchmark rates to account for the exchange rate variation that occurs when translating it into a Brazilian currency, by subtracting the year-end exchange rate from the January end-of-month exchange rate listed in the same year in which the commercial loans were approved. This exchange rate variation was not applied to any commercial loans approved in the period 1994 through 1998 when the Brazilian real was pegged to the U.S. dollar and the resulting currency fluctuations were minimal.

In instances where no comparable commercial loans were available for the years in which the loans were taken out, section 351.505(a)(5)(ii) of the Department’s regulations allows us to modify the method noted above for purposes of selecting an appropriate benchmark for comparison purposes. Accordingly, in instances where no comparable commercial loans were available for the years in which FINAME, BNDES-ExIm and BNDESPAR loans were approved, we used the commercial loans reported for the POI as our benchmark, adjusting for the exchange rate variation that occurred in 2000.

We determined that a benefit was conferred through these loan programs when the government loan rates were lower than the interest rates on the companies’ own comparable commercial loans. However, in cases where no comparable commercial loans were available in the year in which the government loans were approved, we did not compare the government loan interest rates to the comparable commercial loan rates in the POI in order to determine whether a benefit exists. For all loans, we calculated the amount of the benefit by comparing the actual interest paid during the POI on those FINAME, BNDES-ExIm, or BNDESPAR loans, to the amount of interest that would have been paid on these loans using the comparable commercial dollar benchmark rates noted above.

b. Short-Term Benchmarks

As discussed below, the Department found the Program for the Industrial Integration and Diversification (Pro-Industria) to be countervailable as a one year tax-deferral program. Section 351.505(a)(3)(ii) of the Department’s regulations state that, in instances where the firm in question did not take out any comparable commercial loans, the Department may use a national average interest rate for comparable commercial loans. Since respondents did not provide any short-term commercial loans, we have used the Brazilian lending rate reported in the IMF International Financial Statistics Yearbook for 2000 as our short-term interest rate benchmark. This Brazilian bank rate is the interest rate charged on short- and medium-term financing to the private sector. We measured the benefit by multiplying the amount of the tax deferral due in the POI by this IMF lending rate. This entire amount constituted the benefit because no interest charges were assessed on these tax deferrals during the POI.

Trading Companies

Section 351.525(c) of the regulations requires that the benefits from subsidies provided to a trading company which exports subject merchandise be cumulated with the benefits from subsidies provided to the firm which is producing the subject merchandise that is sold through the trading company, regardless of their affiliation. In its questionnaire response, the GOB indicated that seven trading companies exported cold-rolled steel to the United States during the POI. The GOB, however, did not identify these trading companies, nor did the GOB provide any quantity and value information, explaining that it was unable to determine whether any of the steel products exported by these trading companies to the United States consisted of subject merchandise. We issued supplemental questionnaires to the GOB and USIMINAS, COSIPA, and CSN, and requested that they identify these trading companies and provide the quantity and value of subject merchandise shipped by them during the POI and that they provide information concerning the use by the trading companies of any of the non-company-specific subsidy programs during the POI. This information was provided by the parties on February 22, 2002.

At verification we established that during the POI, the trading companies did not benefit from any subsidy programs under investigation. See Countervailing Duty Investigation of Certain Cold-Rolled Carbon Steel Flat Products from Brazil: Verification of the Questionnaire Responses Submitted by the Government of Brazil (GOB) (GOB Verification Report) (August 9, 2002) at 14-15.

All subsidies conferred on the production and exportation of subject merchandise benefit the subject merchandise even if it is exported to the United States by an unaffiliated trading company rather than by the producer itself. Therefore, the Department calculates countervailable subsidy rates on the subject merchandise by cumulating subsidies provided to the producer with those provided to the exporter. See section 351.525(c) of the Department’s regulations. However, because we verified that none of the trading companies received benefits under any of the countervailable subsidy programs discussed below, it is not necessary to calculate separate combination rates for the producers and their exporters. Accordingly, we have calculated rates for the producers of the subject merchandise which will be applicable to all U.S. exports, whether direct or through trading companies, by these companies.

Changes in Ownership

On February 2, 2000, the U.S. Court of Appeals for the Federal Circuit (CAFC) in Delverde Srl v. United States, 202 F.3d 1360, (Fed. Cir. 2000), reh’g granted in part (June 20, 2000) (Delverde III), rejected the Department’s change-in-ownership methodology as explained in the General Issues Appendix of the Final Affirmative Countervailing Duty Determination: Certain Steel Products from Austria, 58 FR 37217, 37225 (July 9, 1993). The CAFC held that “the Tariff Act, as amended, does not allow Commerce to presume conclusively that the subsidies granted to the former owner of Delverde’s corporate assets automatically ‘passed through’ to Delverde following the sale. Rather, the Tariff Act requires that Commerce make such a determination by examining the particular facts and circumstances of the sale and determining whether Delverde directly or indirectly received both a financial contribution and benefit from the government.” Delverde III, 202 F.3d at 1364.

Pursuant to the CAFC finding, the Department developed a new change-in-ownership methodology. This new methodology was first announced in a remand determination on December 4, 2000, and was also applied in Grain-Oriented Electrical Steel from Italy; Final Results of Countervailing Duty Administrative Review, 66 FR 2885 (January 12, 2001) (remanded on other grounds in Acciai Speciali Terni S.pA. And Acciai Speciali Terni USA v. United States, 206 F. Supp. 2d 1344 (CIT 2002), affd., Slip. Op. 2002-82 (CIT 2002) (ASTGOES)). We have applied this methodology in analyzing the changes in ownership in this determination. See Comment 1, below.

Methodology

The first step under this new methodology is to determine whether the legal person to which the subsidies were given is, in fact, distinct from the legal person that produced the subject merchandise exported to the United States. If we determine the two persons are distinct, we then analyze whether a subsidy has been provided to the purchasing entity as a result of the change-inownership transaction. If we find, however, that the original subsidy recipient and the current producer/exporter are the same person, then that person continues to benefit from the original subsidies, and its exports are subject to countervailing duties to offset those subsidies. In other words, we will determine that a “financial contribution” and a “benefit” have been received by the “person” under investigation. Assuming that the original subsidy has not been fully amortized under the Department’s normal allocation methodology as of the beginning of the POI, the Department would then continue to countervail the remaining benefits of that subsidy. See, Final Affirmative Countervailing Duty Determination: Pure Magnesium From Israel, 66 FR 49351, (September 27, 2001), and accompanying Issues and Decision Memorandum at “Comment 1: The Department failed to take into account the effects of the privatization of ICL.” In making the “person” determination, where appropriate and applicable, we analyze factors such as (1) continuity of general business operations, including whether the successor holds itself out as the continuation of the previous enterprise, as may be indicated, for example, by use of the same name, (2) continuity of production facilities, (3) continuity of assets and liabilities, and (4) retention of personnel. No single factor will necessarily provide a dispositive indication of any change in the entity under analysis. Instead, the Department will generally consider the post-sale person to be the same person as the pre-sale person if, based on the totality of the factors considered, we determine that the entity in question can be considered a continuous business entity because it was operated in substantially the same manner before and after the change in ownership. See id.

Analysis

Using the approach described above, we have analyzed the information provided by the GOB and USIMINAS, COSIPA, and CSN to determine whether the pre-sale and post-sale entities of each company can be considered the same person. As stated in their questionnaire responses, since their initial privatization auctions of common shares, USIMINAS, COSIPA, and CSN have operated as independent entities. The Department finds that the information on the record of this investigation supports respondents' statements, and therefore we find that USIMINAS was privatized in 1991, and COSIPA and CSN were privatized in 1993.

Continuity of General Business Operations

Although respondents state that there have been numerous changes in the operations of USIMINAS, COSIPA, and CSN since their privatizations, respondents have also noted that these changes were made as part of their ongoing operations and business decisions. See USIMINAS’s, COSIPA's, and CSN's December 17, 2001 questionnaire response at 79.

According to respondents, since their privatizations, all of these companies have acquired interests in steel distributors or service centers; have initiated new management techniques or sales strategies; and, have focused on developing new product lines and value-added products. However, respondents add that none of these changes was directly related to their privatizations. Id. at 79. Additionally, none of the respondents changed their name or corporate identity as a result of the change in ownership.

Continuity of Production Facilities

Respondents note that, since their privatizations, USIMINAS, COSIPA, and CSN have all added and shut down facilities and equipment in order to upgrade their production processes. According to respondents, all the companies have upgraded their blast furnaces in order to increase production capacities; USIMINAS and CSN have also added coating facilities in an effort to expand their product lines. Again, respondents note that these changes were not directly related to their privatizations, but were part of the companies’ ongoing operations and business decisions.

Our review of USIMINAS’s production information indicates little change in the quantity and composition of its production following its privatization. The comparative production data provided at pages 4-5 of USIMINAS’s 1992-1993 financial statement (exhibit 34 of the GOB's December 17, 2001 response) indicates that USIMINAS’s production totals declined slightly, by 1.6 percent, from 1991 to 1992, and that its product mix remained essentially unchanged for this period. In addition, there was only a slight change in its labor productivity ratio of 386 tons/man/year in 1992 (an increase of 3 tons over 1991). A similar review of COSIPA's 1993 financial statement at pages 5 and 11, indicated that annual production of uncoated flat-rolled steel products remained steady, declining slightly from 2.6 in 1992 to 2.5 million tons in 1993. However, COSIPA's labor productivity ratio in 1993 did increase to 223.9 tons/man/year from 208.6 tons/man/year in 1992. No specific information was provided about the continuity of CSN’s production facilities after its change in ownership in 1993. Continuity of Assets and Liabilities

The privatizations of USIMINAS, COSIPA, and CSN were accomplished through the sale of the GOB's shares to private investors, and did not involve the transfer of any of the corporate assets of the companies in question. According to respondents, the privatizations of these companies involved the purchasing of shares of an ongoing corporation. As a result, the new shareholders of these companies continued to maintain an ownership interest that included both the assets and liabilities of the privatized companies. Therefore, the assets and liabilities of USIMINAS, COSIPA, and CSN remained intact throughout the privatization process. See GOB's December 17, 2001 questionnaire response at 56.

Retention of Personnel

Respondents state that after the privatizations of USIMINAS, COSIPA, and CSN, management began to reorganize the personnel of these companies in order to adjust to the private sector and improve production efficiencies. Specifically, USIMINAS revised its sales strategy by establishing closer customer relationships and additional customer services that required a modest increase in its sales staff and a reduction in the number of sales managers. This is supported by information provided at page 9 of USIMINAS’s 1992-1993 financial statement, indicating that the number of USIMINAS’s hired personnel in 1992 was 2.7 percent below the number of its personnel in 1991. COSIPA also experienced a 16.8 percent reduction in personnel between December 1992 and December 1993, as reflected on page 11 of COSIPA's 1993 financial statement. This period encompasses four months from the time of COSIPA’s initial privatization auction in August 1993. No specific information was provided about CSN's personnel adjustments made as a result of its change in ownership in 1993.

Summary

Based on the analysis above, we determine that the vast majority of the business aspects of USIMINAS, COSIPA, and CSN were unchanged by their respective privatizations. All of these companies still operate in a manner similar to that characterizing their operations prior to privatization. As respondents themselves noted, the legal status of these businesses did not change as a result of their privatizations. Instead, the GOB's privatization process involved the sale of shares of ongoing corporations that resulted in the transfer of control and ownership, and in the assumption of each company's existing assets and liabilities. Any changes made in the business operations of USIMINAS, COSIPA, and CSN can be attributed to the ongoing operations and business decisions of these companies, as stated by respondents themselves. In addition, the production levels and product mix of each company remained essentially the same after its change in ownership. While there is information that indicates that the management and personnel of these companies may have been altered as a result of their privatizations, on balance, we do not consider these changes to be sufficient to find that USIMINAS, COSIPA, and CSN were different business entities after privatization. As respondents themselves have noted, most of the changes were due to ongoing business decisions and were not directly related to privatization itself. Accordingly, our analysis leads us to determine USIMINAS, COSIPA, and CSN to be the same persons which benefitted from subsidies bestowed by the GOB prior to their privatizations.

I. Programs Determined to Confer Subsidies

A. Equity Infusions

Petitioners alleged that the GOB provided equity infusions during the following periods: to CSN from 1986 through 1992; to USIMINAS from 1986 through 1988; and to COSIPA from 1986 through 1993. In our past investigations of hot-rolled steel from Brazil and cold-rolled steel from Brazil, we found that the GOB, through SIDERBRAS, provided equity infusions to USIMINAS, CSN and COSIPA. See Brazil Hot-Rolled Final, 64 FR at 38747, 38748 and Brazil Cold-Rolled Final, 65 FR at 5546, 5547. For the reasons cited in the last cold-rolled investigation by the Department (see id.), and because none of the parties have provided new information or argument which would lead us to reconsider this determination, we are continuing to find, under section 771(5)(D) of the Act, that these equity infusions constitute a financial contribution by the GOB to CSN from 1986 through 1992, to USIMINAS from 1986 through 1988, and to COSIPA from 1986 through 1993. The equity infusions into CSN in 1992, and into COSIPA in 1992 and 1993, were made through debt-for-equity swaps and are discussed in more detail below.

As in the previous cold-rolled investigation, we also find that these equity infusions provided a benefit pursuant to section 771(5)(E)(i) of the Act because the companies were unequityworthy at the time the equity infusions were made. These equity infusions are specific within the meaning of section 771(5A)(D)(i) of the Act because they were provided specifically to each company. Accordingly, we determine that the pre-1992 equity infusions are countervailable subsidies within the meaning of section 771(5) of the Act.

As explained in the “Equity Methodology” section above, we treat equity infusions into unequityworthy companies as grants given in the year the infusion was received. These infusions are non-recurring subsidies in accordance with section 351.524(c)(1) of the Department’s regulations. Consistent with section 351.524(d)(3)(ii) of the Department’s regulations, because USIMINAS, COSIPA and CSN were uncreditworthy in the relevant years (the years the equity infusions were received), we applied an uncreditworthy discount rate, as discussed in the “Discount Rates” section above.

Additionally, we find, as in the last cold-rolled investigation, that the GOB provided debtfor- equity swaps to CSN in 1992 and COSIPA in 1992 and 1993. See Brazil Cold-Rolled Final, 65 FR at 5547, 5548. Prior to CSN’s and COSIPA’s privatization, and on the recommendation of consultants who examined CSN and COSIPA, the GOB made a debt-for-equity swap for CSN in 1992 and two debt-for-equity swaps for COSIPA in 1992 and 1993. We previously examined these swaps and determined that they were not consistent with the usual investment practices of private investors. As such, they constitute a financial contribution by the GOB within the meaning of section 771(5)(D)(i) of the Act. We also determined that a benefit in the amount of each conversion was conferred in accordance with section 771(5)(E)(i) of the Act. See id., citing to Brazil Hot-Rolled Final, 64 FR at 38747, 38748. These debt-for-equity swaps are specific within the meaning of section 771(5A)(D)(i) of the Act because they were limited to CSN and COSIPA. Accordingly, we determine that the GOB debt-for-equity swaps provided to CSN in 1992 and COSIPA in 1992 and 1993 are countervailable subsidies within the meaning of section 771(5) of the Act. No party has provided any new information or argument which would lead us to reconsider this determination.

Each debt-for-equity swap constitutes an equity infusion in the year in which the swap was made. As such, we have treated each debt-for-equity swap as a grant given in the year the swap was made, in accordance with section 351.507(b) of the regulations. Further, these swaps, as equity infusions, are non-recurring in accordance with section 351.524(c)(1) of the regulations. Because CSN and COSIPA were uncreditworthy in the years of receipt, we applied a discount rate consistent with section 351.524(d)(3)(ii) of the regulations, as discussed in the “Discount Rates” section above.

As a result of our privatization approach outlined in the “Changes in Ownership” section above, we find that USIMINAS, CSN and COSIPA continue to benefit from subsidies received prior to their privatization, and therefore, the full value of the benefits allocable to the POI from these equity infusions and debt-for-equity swaps is being used in the calculation of USIMINAS’s, CSN’s and COSIPA's subsidy rates. For USIMINAS/COSIPA, we summed the benefits allocable to the POI from each equity infusion and swap, and divided this total by the combined total sales of USIMINAS/COSIPA during the POI. On this basis, we determine the net subsidy to be 10.31 percent ad valorem for USIMINAS/COSIPA. For CSN, we summed the benefits allocable to the POI from each equity infusion and swap, and divided this total by CSN’s total sales during the POI. On this basis, we determine the net subsidy to be 7.08 percent ad valorem for CSN.

B. Presumed” Tax Credit for the Program of Social Integration and the Social Contributions of Billings on Inputs Used in Exports (“PIS/COFINS”)

In 1970, through Supplementary Law No. 7, the GOB established PIS which is “intended to bring about integration of employees in the life and growth of their companies.” Essentially, companies make PIS contributions to a fund which is “a means of creating wealth for ... employees.” In 1991, through Supplementary Law No. 70, the GOB established COFINS as a contribution for the financing of social insurance “intended solely to defray the cost of health care and social security and assistance work.” PIS and COFINS taxes are assessed on all products purchased domestically but do not apply to the sale of products that are exported. During the POI, the PIS and COFINS rates were 0.65 percent and 3.0 percent, respectively. Each company is responsible for making monthly payments of PIS and COFINS based on the total value of its domestic sales of goods and services.

In 1996, through Law 9363, the GOB established the PIS and COFINS tax credit program to provide a rebate of PIS and COFINS contributions assessed on the purchase of raw materials, intermediate products, and packing materials used in the production of exports. The PIS and COFINS “presumed” tax credit was established to prevent the cascading effect of these taxes which accrue at each point in the chain of production. A company calculates its own PIS and COFINS credit, on a monthly basis, using a standard formula established by Law 9363, and claims the credit by making deductions from the Industrial Products Tax (“IPI”) due.

The “presumed” tax credit rate for PIS and COFINS is 5.37 percent and applies to all industries. According to the GOB, this percentage was calculated using the PIS and COFINS rate in effect at the inception of Law 9363, which was 2.65 percent. In establishing this credit rate, the GOB assumed two stages of processing prior to exportation and, thus, two prior stages of PIS and COFINS tax incidence. As stated by the GOB in its February 5, 2002 questionnaire response, in establishing the credit rate, the GOB did not determine the value, quantity or type of inputs consumed in the production of any particular producer of subject merchandise, nor did the GOB take into account any yield factors. According to the GOB, this tax credit rate was arbitrarily chosen for administrative convenience. To calculate its credit, a company divides its export revenues, accumulated through the prior month, by its total sales revenues for the same period. See GOB December 26, 2001 Questionnaire Response at 6. This export revenue ratio is then multiplied by the company’s total value of purchases as reflected in the supplier’s sale invoices for raw materials, semi-finished products, and packaging materials used in the production process. Articles 1 through 3 of Law 9363 provide the methodology for calculating a company’s tax credit. This amount is then multiplied by the tax credit rate of 5.37 percent to yield the year-to-date accumulated tax credit. In order to calculate the credit for the current month, the credit used through the prior month is deducted from this accumulated tax credit. Both the IPI tax and presumed credit are electronically filed quarterly and annually with the Secretariat of Federal Revenue under a company’s tax identification number. Those companies claiming presumed credits under PIS/COFINS are subject to audits which require them to provide evidence of the types and amounts of inputs used to calculate the credit, and the costs of such inputs.

Section 351.102(b) of the Department’s regulations defines an indirect tax as a “sales, excise, turnover, value added, franchise, stamp, transfer, inventory, or equipment tax, border tax, or any other tax other than a direct tax or an import charge.” As noted in the PIS and COFINS legislation, these taxes are derived from the “monthly invoicing” or “invoicing” originating from the sale of goods and services. Therefore, we find that the manner in which these taxes are assessed is characteristic of an indirect tax, and we are treating PIS and COFINS taxes as indirect taxes. (For further discussion, see Comment 2, infra.)

Further, the Department’s regulations define both “cumulative indirect tax” and “priorstage indirect tax.” See section 351.102(b) of the Department’s regulations. A “cumulative indirect tax” is a “multi-staged tax levied where there is no mechanism for subsequent crediting of the tax if the goods or services subject to tax at one stage of production are used in a succeeding stage of production.” A “prior-stage indirect tax” is defined as “an indirect tax levied on goods or services used directly or indirectly in making a product.” Therefore, because PIS and COFINS are charged on inputs used to make cold-rolled steel, they are charged on goods at one stage of production that are used in a succeeding stage of production, thus falling within the definitions of “cumulative indirect tax” and “prior-stage indirect tax.”

Based on our determination that PIS and COFINS are prior-stage cumulative indirect taxes, we examined whether the GOB has a system or procedure in place within the meaning of section 351.518(a)(4)(i) of the Department’s regulations. First, we note that this system was established as a simplified and streamlined methodology to determine the amount of the tax rebate for all companies in Brazil. The only limitation imposed on companies making rebate claims is that the claims be limited to those inputs defined under the PIS and COFINS rebate law, which is broader than the “consumed in production” standard provided for in the Department’s regulations at section 351.518(a)(1). In fact, evidence on the record of this investigation shows that companies include in their claims purchases that are not consumed in production. See e.g. Exhibit 3 of February 5, 2002 Companies’ Response. Companies report their purchases of inputs based on the assumption that all goods purchased are consumed equally in exported and domestically sold goods. Further confirmation is not conducted by the government. As such, we find that this system does not permit the GOB to confirm which inputs are being consumed in the consumption of exported goods and in what amounts.

We also have additional concerns that this system does not operate in accordance with the requirements under the regulations. The system does not account for the fact that domestic and export sales may include imported inputs. These imports may be included by varying degrees in either export or domestic sales, thus distorting the ratio. Either way, the GOB does not account for, and does not require Brazilian companies to account for, any such potential distortions. Further, in determining the actual amounts of inputs consumed in final products, the government does not make due allowance for waste, thereby raising the concern that the claim amount is overstated. This is an important element that the Department has addressed in countervailing another rebate program in the past. See Final Affirmative Countervailing Duty Determination: Certain Hot- Rolled Carbon Steel Flat Products From Thailand, 66 FR 50410 (October 3, 2001), and accompanying Issues and Decision Memorandum (Thai Hot-Rolled Final) at “Comment 3: Countervailability of Section 36(1) Benefits.” Because we found that the GOB has not met the requirements under section 351.518(a)(4)(i) of the Department’s regulations, we have determined that the entire amount of the PIS and COFINS remission confers a benefit to respondent companies. According to section 771(5)(D)(ii) of the Act, granting tax credits constitutes a financial contribution. Furthermore, because PIS and COFINS rebates are calculated based on a company’s export revenue, i.e., are available only to exporters, we find that this program is de jure specific according to section 771(5A)(B) of the Act.

In calculating the benefit, consistent with section 351.524(c)(1) of the Department’s regulations, we treated the tax rebate as a recurring benefit and divided the total tax credit claimed by each respondent company during the POI, by its export sales during the POI. For CSN, we have calculated the ad valorem rate in accordance with section 351.525(b)(2) of the Department’s regulations by dividing the total tax credit claimed during the POI by CSN’s total export sales during the POI. On this basis, we determine the net subsidy rate for CSN to be 0.79 percent ad valorem. In calculating a combined rate for USIMINAS/COSIPA, we calculated the benefit by first combining the tax credits claimed by both USIMINAS and COSIPA during the POI, and then dividing this total benefit amount by their combined export sales during the POI.

This is consistent with the calculation methodology outlined under section 351.525(b)(6)(ii) of the Department’s regulations for corporations with cross-ownership. On this basis, we determine the net subsidy to be 1.21 percent ad valorem for USIMINAS/COSIPA.

C. National Bank for Economic and Social Development (BNDES) Loan Programs BNDES administers several programs and is the main source of medium- and long-term financing in Brazil. BNDES works with commercial banks that act as agents for BNDES. BNDES’s goals include improving productivity and competitiveness in Brazil by financing new players in investment and in growing sectors of the economy, and by strengthening capital markets in Brazil. BNDES’s main credit activities are to finance fixed investments, finance the export of goods and services, finance the acquisition of leased machinery and equipment, and to support investments.

There are two types of BNDES financing, direct and indirect. Under direct financing, the company receives financing directly from BNDES. In these cases, BNDES bears the credit risk. In indirect financing, the company receives financing from an agent bank that is working with BNDES financing, e.g., using funds made available to the agent bank by BNDES. In these cases, the agent bank bears the credit risk.

For all BNDES financing programs, the interest rate used is composed of three variables: the financial cost; a basic spread; and a risk spread. The average basic spread is 1.91 percent, which is used to pay the operational costs of the financing. BNDES also includes a risk spread that ranges from zero to 2.5 percent that is applied to all direct financing operations in which BNDES assumes the risk of default. For all indirect operations, a portion of the basic spread and the risk spread are negotiated between the agent financial institutions and the companies obtaining the financing. The financial cost is determined based on a percentage of the total amount of financing.

BNDES can make loans in foreign currency using U.S. dollars and the UMBNDES, a mechanism that is used to value a basket of currencies. BNDES dollar financing is normally charged an interest rate of LIBOR plus a spread. The UMBNDES currency is determined daily. BNDES publicizes the exchange rate data for the UMBNDES, showing the daily exchange rate for Brazilian reais. This currency is not publicly traded. This mechanism allows BNDES to avoid its exchange rate risk by transferring it to the borrower.

When UMBNDES is used as the loan currency for a BNDES loan, the value of the UMBNDES at the date of disbursement is the value that will be applied to the loan. Although the loan funds that BNDES obtains on the international market for purposes of onward lending may actually be in any one of a number of foreign currencies, the company receives the loan disbursements in reais. The repayment amount is based on the value of the UMBNDES at the time of repayment. The repayment amount includes principal, plus interest based on a basic spread, plus a risk spread.

UMBNDES provides the weighted-average of the exchange variations among the various currencies in the BNDES currency baskets. BNDES applies a variable interest rate that is determined in part by the weighted average of all the fees and expenses, excluding taxation, incurred by BNDES in the process of raising funds abroad. This variable interest rate is calculated quarterly. Clients who agree to use a currency basket in their transactions with BNDES are obligated to pay a basic spread and a risk spread.

Another means of long-term BNDES lending in reais is done using the BNDES long-term interest rate (TJLP) which is set quarterly by the Central Bank of Brazil. The TJLP is the basic cost of financing granted by the BNDES. This long-term rate incorporates both a risk premium and an inflation target based on the Broad Consumer Price Index (IPCA). This type of lending also applies a capitalization factor when the TJLP interest rate is above six percent which allows borrowers to capitalize the portion of interest above six percent and amortize it over the remaining loan term. The TJLP interest rate was applied to some of the loans that were outstanding during the POI for the FINAME, FINEM and BNDES-ExIm financing programs. At verification, we confirmed the interest capitalization feature of these loans.

We are investigating several BNDES’s financing programs. Each of the programs the Department finds countervailable is analyzed below.
1. FINAME

The FINAME financing program is administered by BNDES; its purpose is to finance machinery and equipment purchases. This type of financing is primarily made on an indirect basis through agent banks. The FINAME financing program began in the 1990s to finance purchases of Brazilian-produced equipment. At that time, there was a separate guarantee under the FINAME program for foreign-made equipment over domestic-made equipment. Essentially, financing was provided for Brazilian-made equipment and for imported equipment only if it could not be obtained in Brazil. Financing was not provided for foreign-made equipment if the equipment was produced in Brazil.

In 1997, a second line of FINAME financing became available. This line of FINAME financing is available to borrower’s purchasing foreign equipment regardless of the availability of Brazilian-made equipment. However, our review of the provisions regulating this second line indicates that it is available only to “micro, small and medium size companies” in Brazil. See GOB’s Submission of Verification Exhibit dated July 3, 2002, at exhibit BNDES-5. Since respondent companies do not qualify as micro, small or medium companies, any FINAME financing which they received was granted under the provisions of the original line of FINAME financing as noted above, which is limited in that it provides financing for imported equipment only when domestically manufactured equipment is unavailable in Brazil.

For FINAME transactions up to 7 million reais, the program has become known as FINAME Automatico, because the process is standardized and is used frequently by many different companies in Brazil. In these transactions, the project must fall within the purview of BNDES, however BNDES will not directly assess the credit risk. This is the responsibility of the agent banks. If the loan request is for an amount above 7 million reais, the program is known as FINAME Especial. In these cases, the agent bank must first consult and obtain permission from BNDES to finance the transaction.

The terms applied to FINAME loans vary depending on whether the financing is for imported or domestically-produced equipment. For domestically-produced equipment, FINAME finances up to 90 percent of the purchase for a small business and up to 80 percent of the purchase for a large company. If the equipment is imported, or less than 60 percent Brazilian content, the financing must be made from a basket of foreign currencies. For imported equipment, a maximum financing term of five years is applied, and financing is available for 85 percent of the value of the equipment for small businesses and for 80 percent of the value for large businesses. During the POI, respondents had outstanding FINAME loans denominated in dollars, reais, and cruzeiros. See Benchmark Rates for BNDES Loans, section, above.

The respondent companies did not use the line of financing available for imported equipment, regardless of the availability of domestic equipment, as this line of financing is only available to micro, small, and medium businesses. Therefore, we are only examining the specificity of the original FINAME line of financing, which they did use. We find this program to be de jure specific as an import substitution program because it is only available to finance the purchase of domestically-produced equipment. See section 771(5A)(C) of the Act. We further find that there is a financial contribution, through the provision of loans, under section 771(5)(D)(i) of the Act.

We calculated the benefit in accordance with sections 351.505(a)(5)(i) and 351.505(a)(5)(ii) of the Department’s regulations, by comparing the actual interest paid on the outstanding FINAME loans during the POI, to the amount of interest that would have been paid on these loans using the comparable commercial dollar benchmark rates noted in the Benchmarks for BNDES Loans section above. Based on this comparison, we determine that USIMINAS/COSIPA and CSN received countervailable subsidies under the FINAME financing program during the POI. For CSN, we calculated the subsidy amount by dividing the benefit received during the POI from these loans by CSN’s total sales during the POI, and determined a countervailable benefit of 0.03 percent ad valorem for the POI. For USIMINAS/COSIPA, we summed the benefits allocable to the POI, and divided this total by the combined total sales of USIMINAS/COSIPA during the POI. We thus determine the countervailable subsidy from this line of credit for USIMINAS/COSIPA to be less than 0.005 percent ad valorem.

2. BNDES ExIm

BNDES provides ExIm financing to support the production and commercialization of exports. There are three types of BNDES ExIm Financing: pre-shipment financing, special preshipment financing; and, post-shipment financing. Pre-shipment financing is a working capital loan that is for the production of goods and can be for a term of up to 12 years. Special preshipment financing is provided when the good has already been produced, but the buyer needs to obtain financing for commercialization of the product, and can be granted for a term of up to 30 months. This financing uses interest rates tied to TJLP, which provides for the capitalization of interest payments when the rate exceeds six percent. Post-shipment financing is a buyer’s or supplier’s credit that can be granted for a term of up to 30 months. All of the terms applied to BNDES ExIm financing include a finance cost consisting of a minimum of a one percent spread from BNDES based on the total amount of the financing, plus a spread determined by the agent. Most export transactions are conducted in U.S. dollars, but BNDES disburses funds in reais. Eligibility for any BNDES ExIm financing must be approved by an agent bank and by BNDES. However, the relationship concerning the loans remains between the agent bank and the borrower. All ExIm financing is guaranteed by the exports being financed. This financing is generally used by companies that are exporting regularly.

We determine that the BNDES ExIm financing program is de jure specific in accordance with section 771(5A)(B) of the Act because eligibility for the financing is contingent upon export performance; there is a financial contribution through the provision of loans, under section 771(5)(D)(i) of the Act. Furthermore, there is a benefit to the extent that the interest rates on these loans are lower than the interest rates on comparable commercial loans. See section 771 (5)(E)(ii) of the Act.

We calculated the benefit in accordance with sections 351.505(a)(5)(i) and 351.505(a)(5)(ii) of the Department’s regulations, by comparing the actual interest paid on the outstanding BNDES-ExIm loans during the POI, to the amount of interest that would have been paid on these loans using the comparable commercial dollar benchmark rates noted in the Benchmarks for BNDES Loans section above. Based on this comparison, we determine that COSIPA received countervailable subsidies under the BNDES-ExIm financing program during the POI. In calculating a combined rate for USIMINAS/COSIPA, we divided COSIPA’s total benefit amount by the total combined export sales of USIMINAS and COSIPA during the POI. This is consistent with the calculation methodology outlined under section 351.525(b)(6)(ii) of the Department’s regulations for corporations with cross-ownership. On this basis, we determine the net subsidy to be 2.14 percent ad valorem for USIMINAS/COSIPA.

3. BNDESPAR

In 1992, BNDES established BNDESPAR to develop and promote the Brazilian capital markets, and to encourage the sale of BNDES’s financial assets of convertible debentures and shares (including privatization currencies) in public companies to outside investors as part of Brazil’s National Privatization Program (“NPP”). See GOB Verification Report at 28. These assets were sold to third-party investors residing in Brazil; to legal entities directly or indirectly controlled by individuals residing in Brazil; to closed, private security foundations “constituted” in Brazil; or to security investment funds managed in Brazil whose “quota-holders” also reside in Brazil. See Resolution No. 786/92, Articles 1 and 2 of verification exhibit BNDES-10 in respondents’ July 8, 2002 submission. Under the NPP, BNDES authorized financial agents to acquire and resell these privatization currencies to themselves and/or their clients via installment payments through individual or collective financing facilities provided by BNDES. Id at Article 3 through Article 5. BNDESPAR was responsible for administering and implementing procedures associated with all transactions occurring within this program. Id at Article 20.

Finally, any type of operation that involved the transfer or assigning of assets to another party involved the approval of BNDES. Id. at Articles 8 through 12; see also August 9, 2002 GOB Verification Report at 28. During the POI, only USIMINAS had an outstanding BNDESPAR loan from BNDES for the purchase of privatization currencies used to purchase COSIPA. USIMINAS took over this loan when it purchased shares in COSIPA, and it refinanced this loan directly with BNDES.

At verification, we requested information on the distribution of these loans prior to and in 1993; we were only able to obtain information that covered 1993, the year that the original BNDESPAR loan was provided to another company to purchase shares in COSIPA. In 1993, BNDESPAR funds went to a limited number of companies in a few industries for investments in a small number of state-owned enterprises. See BNDES verification exhibit 8 submitted July 3, 2002; also Calculations Memorandum for the Final Determination: Certain Cold-Rolled Flat- Rolled Carbon Quality Steel Products from Brazil which contains detailed information on the actual number of industries and companies involved. We also have information on loan approvals in 1995, the year in which USIMINAS refinanced this loan directly with BNDES when it purchased shares in COSIPA. In 1995, although 43 industries were identified by BNDES in exhibit 4 of the GOB’s April 9, 2002 submission, BNDESPAR loans were approved in only 13 industries for investments in only state-owned enterprises. Furthermore, we identified the top five recipients of these loan approvals which includes the steel industry, and found that these five industries represent over 80 percent of the BNDESPAR loans approved in 1995. Accordingly, we find this program to be de facto specific under section 771(5A)(D) because it is limited to a group of enterprises or industries.

We further find that because USIMINAS’s BNDESPAR loan has been refinanced directly through BNDES rather than through a financial agent, this program provided a financial contribution pursuant to section 771(5)(D)(i) of the Act, in the form of a direct transfer of funds in the form of loans from the GOB. We find that a benefit has been conferred in accordance with section 771 (5)(E)(ii) of the Act to the extent that the interest rates on this loan are lower than the interest rates on comparable commercial loans.

Since USIMINAS did not have comparable commercial loans taken out in the same year in which it received this BNDESPAR loan, we calculated the benefit in accordance with section 351.505(a)(5)(i) and 351.505(a)(5)(ii) of the Department’s regulations, by comparing the actual interest paid on the outstanding BNDESPAR loan during the POI, to the amount of interest that would have been paid on this loan using the comparable commercial dollar benchmark rate noted in the Benchmarks for BNDES Loans section above. Based on this comparison, we determine that USIMINAS received a countervailable subsidy under this program during the POI. To calculate the POI subsidy amount, we divided USIMINAS’s POI benefit received under this program in the POI, by the combined total sales of USIMINAS and COSIPA during the POI. Accordingly, we find a countervailable benefit of 0.20 percent ad valorem for USIMINAS/COSIPA during the POI.

D. Provincial Government Program

PRO-INDUSTRIA

Pro-Industria is a tax-deferral program administered by the Development Bank of Minas Gerais (BDMG) that was set up by the state government of Minas Gerais with two purposes: (1) to increase the Tax for Operations Relating to the Circulation of Merchandise and Services of Interstate and Intercity Transport and Communication (ICMS), a value-added tax revenue, and (2) to encourage all industries in the state of Minas Gerais to expand their current production facilities. Additional ICMS revenue is generated by expanding an existing plant or by establishing a new one in Minas Gerais. When an industry decides to expand or to set up a new plant, the company calculates how much ICMS taxes are due. The increase in ICMS that must be paid is equivalent to the amount of the new investment. The Pro-Industria program only works with the increased ICMS resulting from the increase of ICMS taxes a company must pay due to its expanded operations. All companies that expand their operations by at least one percent must pay an increased ICMS amount. The Pro-Industria program works almost exclusively for production of goods sold to local markets, because for exports, no ICMS taxes are paid. Every time a corporation pays ICMS taxes, the amount paid is compared to previous ICMS paid before the expansion. Under this program, approved companies have the opportunity to defer, for up to one year, payment of up to fifty percent of the increase in the ICMS.

This program can be used by a company for a maximum of 60 months. The program operates as follows: the companies pay the total amount of the tax due each month, including the increased ICMS to the BDMG. BDMG then pays back fifty percent of the increase in ICMS to the company five days after the ICMS tax is paid by the company, with a 2.5 percent processing fee deducted by BDMG based on the total amount of money paid back to the company. The amount of money that is paid back by BDMG to the companies plus the processing fee is the total tax deferral. One year later, the company must pay back to BDMG the monthly installment received in the same month in the prior year.

In order to qualify for benefits under the Pro-Industria program, a company must first provide to BDMG a formal outline explaining either the new project or the expansion of an existing project, which is then analyzed by BDMG. The analysis that is undertaken by BDMG involves a study of the overall financial standing of the company, the company’s history, the legal situation of the company, and the market for the increased production. If the company is only increasing production for export, no ICMS would be due, therefore the company would not be eligible for participation in the Pro-Industria program. Additionally, only one qualifying Pro- Industria project can be undertaken by any company at one time.

The calculation of the ICMS eligible for Pro-Industria benefits is based on a tax unit where the tax is calculated per unit per month. If the company pays taxes based on more than one tax unit, the ICMS eligible for Pro-Industria benefits only works with the tax unit to which the project applies.

Of the companies under investigation, only USIMINAS was eligible to use the Pro- Industria program during the POI. USIMINAS was approved for the Pro-Industria program for the maximum time allowable, 60 months, and approval was formally granted on October 14, 1997. The schedule that was applied to USIMINAS was for 60 monthly installments, and corresponded to fifty percent of ICMS tax paid compared to an average year of reference.

In examining the specificity of Pro-Industria, we note that there is no limitation on eligibility within the state of Minas Gerais. The law, Decree 38.106 of July 1, 1996, makes the program available to any business undertaking a new project or a project expansion. Therefore, we find no basis for determining that Pro-Industria is de jure specific under section 771(5A)(D)(i) of the Act. In order to determine whether Pro-Industria is provided to a specific group of enterprises or industries on a de facto basis under section 771 (5A)(D)( iii) of the Act, we have examined the actual distribution of benefits under this program since USIMINAS had outstanding one-year tax deferrals due in the POI. From 1997 through 2000, a wide range of industries received approval under the program. However, information on the record shows that the automobile and steel industries have received a disproportionate share of the tax deferrals under this program. The tax deferrals granted to these two combined industries, from 1997 through 2000, amount to more than sixty percent of total tax deferrals provided. In this same period, virtually no other industry received tax deferrals of more than 1 to 2 percent of total tax deferrals. Accordingly, we find this program limited on a de facto basis to a group of industries in accordance with section 751(5A)(D)(iii)(II) of the Act. We further find that a financial contribution has been provided in accordance with section 771(5)(D)(ii) of the Act because the government is not collecting tax revenue that is otherwise due by allowing for these tax deferrals.

Because the Pro-Industria program provides a one-year tax deferral, a benefit exists to the extent that appropriate interest charges are not collected. See section 771(5)(E) of the Act. According to section 351.509(a)(2) of the Department’s regulations, a deferral of direct taxes will normally be treated as a government-provided loan in the amount of the tax deferred, according to the methodology described in section 351.505 of the Department’s regulations. Section 351.505(a)(3)(ii) of the Department’s regulations provides that in the case where a firm has no comparable commercial loans, the Department may use a national average interest rate as a shortterm benchmark interest rate in determining if the amount a firm pays on the governmentprovided loan (tax deferral) is less than the amount the firm would have paid using this benchmark interest rate.

Because the entire amount of USIMINAS’s tax deferral is interest free, we multiplied the total amount of the taxes deferred and due in the POI by the average lending rate in Brazil for short- and medium-term financing in 2000 to determine the total benefit received by USIMINAS during the POI. See Short-Term Benchmark for BNDES Loans section noted above for a complete discussion on the selection of a short-term benchmark interest rate for the Pro-Industria program. To calculate the subsidy amount, we divided USIMINAS’s total POI benefit received by the combined total sales of USIMINAS and COSIPA during the POI. The countervailable subsidy amount under this program for USIMINAS/COSIPA for the POI is 0.13 percent ad valorem.

II. Program Determined Not to Confer a Subsidy

FINEM

FINEM is a loan program administered by BNDES that supports fixed asset investments for projects in the industrial and agricultural sector in excess of seven million reais. Projects supported under this program include construction projects, engineering projects, projects for the acquisition of equipment, studies of special projects, and working capital projects. In order to be eligible for working capital financing, the project must either be new or must be an expansion of an existing project.

Most FINEM financing is direct financing from BNDES, however, the project is always analyzed by BNDES because of the large amount of funding involved. The board of directors of BNDES must approve the credit risk before financing is given. A company may only obtain financing under the FINEM program once per year.

In order for a company to receive FINEM financing, the company must fill out a preliminary application that is included in the “BNDES Guide for Preliminary Application,” that is used by BNDES to track information on each company and its eligibility for financing. If a project is determined to meet all of the rules of eligibility for BNDES financing, then the project is sent to the credit committee of the bank, which is made up of the superintendents of all of the different areas of the bank. Every superintendent may participate in the decision of whether to grant FINEM financing to the requesting company. BNDES will often make another internal evaluation of how the project will affect the market, what the expected benefit of the project will be, how many new jobs will result from the project, and the amount of funding that the company requires and the terms of this funding.

FINEM financing terms are typical of the terms applied to all BNDES loans. This usually consists of an interest rate of TJLP, plus a basic spread, plus a risk spread. This financing can also be made up of a basket of currencies or U.S. dollars. If the loan is in U.S. dollars, the interest rate would be LIBOR plus a basic spread and a risk spread. The repayment schedule is negotiated between the company and BNDES or the agent bank. Typical repayment terms are 6 months beyond the conclusion of the project on interest and principal, with a grace period of principal only. Interest payments are paid throughout the life of the loan. Loan funds are disbursed quarterly according to the expenditures of the project. With each installment, interest begins accruing. Interest must be paid quarterly, including during the grace period. After the grace period ends, the repayment period starts. The terms of the repayment are constant throughout the life of the loan. Also, the amount of money actually disbursed may be more or less than the amount originally contracted by the company because the company may end up spending more or less on the financed project than was originally projected. Repayment periods typically last about seven years, one and a half for disbursement and construction, with a six month grace period, and a five year repayment period. However, repayment terms can differ based on negotiations between the company and BNDES or the agent bank.

Our review of the GOB-provided breakdown of FINEM financing shows that the financing was provided to a wide variety of industries ranging from paper to electricity to farming products. We found nothing which suggested that this program was de jure specific under 771(5A)(D)(i). Further, the breakdown of FINEM financing by industry indicates that the steel industry was neither a predominant user nor disproportionate recipient of FINEM financing each year in which respondent companies received approval for loans that were still outstanding during the POI and the three years before each relevant year of loan approval. We therefore find that this program is not de facto specific since the loans are widely available to numerous industries in Brazil and are actually used by a wide distribution of industries. See section 771(5A)(D)(iii) of the Act.

III. Programs Determined to be Not Used A. Programa de Financiamento as Exportacoes (PROEX)

The PROEX program is administered by the Banco do Brasil. PROEX funding is available to Brazilian exporters or foreign importers of Brazilian goods and services. PROEX funds are available in two forms:

(1) PROEX Financing, which involves the direct financing of a company’s exports and
(2) PROEX Equalization, which reimburses certain interest costs to Brazilian and foreign financing agencies.

Under the PROEX Financing program, the GOB provides financing to the foreign importer for up to 85 percent of the value of the exported merchandise. The Brazilian exporter receives a cash payment from the Bank of Brazil after presenting its export records for approval. The importer then makes installment payments to the Bank of Brazil at the interest rate negotiated between the importer and the exporter. If this interest rates includes a spread in excess of the spread established in the PROEX program, this additional amount will be passed through the Bank of Brazil to the exporter. See GOB Verification Report dated August 9, 2002 at 13; see also GOB Verification exhibit GOB-8, dated July 3, 2002.

The PROEX Equalization program provides funds to Brazilian and foreign international banks that provide credit to Brazilian exporters or foreign importers of Brazilian goods and services at international market rates. GOB officials explained at verification that absent this program, the cost of export financing in Brazil would be more expensive due to the higher cost of securing funding from abroad. Under this program, PROEX is responsible for assuming these additional financing costs that are incurred by these international banks for financing exports in Brazil. None of the three respondent companies or relevant trading companies received loans under the PROEX loan program during the POI, nor did any of them owe interest or principal during the POI on such loans. Therefore, we determine that this program is not used for purposes of this investigation.

B. Provincial Program: Program to Induce Industrial Modernization of the State of Minas Gerais (PROIM)

PROIM is a long-term financing program offered by the state of Minas Gerais and administered by the Development Bank of Minas Gerais (BDMG). The PROIM program provides loans for up to eight years to finance investments such as production, with the goal of improving the quality of companies’ technology in the State of Minas Gerais. The PROIM program is funded by the state treasury of Minas Gerais.

The application process for approval for PROIM loans is similar to that of the Pro- Industria program. After the company files a request, BDMG analyzes whether there are available funds to finance the company’s project.

Of the three companies under investigation, only USIMINAS has facilities located in the state of Minas Gerais; therefore it is the only company eligible to use the PROIM program. USIMINAS did not receive loans under this program during the POI, nor did it owe interest or principal during the POI on such loans. Therefore, we determine that this program is not used for purposes of this investigation.

Analysis of Comments

Comment 1: CSN’s, USIMINAS’s and COSIPA’s Privatization

Respondent’s Argument: Respondents argue that the Department must discontinue using its current privatization methodology in determining whether the current producer/exporter is the same person after a change-in-ownership transaction, and therefore is benefitting from preprivatization subsidies. Respondents argue that the Department’s use of the “same person” test contradicts the ruling of the Court of Appeals in Delverde SrL v. United States, 202 F.3d 1360 (2000) (Delverde III). Respondents specifically argued that the Delverde III Court required Commerce to find that the purchaser indirectly received subsidies from the government by looking at the facts of the purchase transaction to determine whether the new owner “paid full value for the asset and thus received no benefit from the prior owner’s subsidies.” The respondents argue that because of this holding, the Department’s “same person” test is not permissible as a legal matter, and therefore, cannot determine whether the new owners have received a countervailable benefit from the pre-privatization subsidies. Respondents argue that in order to comply with the Delverde III test, the Department must inquire into whether the new owners of CSN, USIMINAS, and COSIPA paid fair market value for their acquisition of the previously state-owned companies.

Respondents further cite to four separate Court of International Trade decisions from 2002, all of which hold that the Department’s “same person” test applied by the Department in privatization analysis violates the requirements of Delverde III by assuming that benefits have passed through to the purchasers of the company so long as the company’s productive assets remain the same. Respondents argue that these four CIT cases hold that the Department’s privatization analysis must satisfy three criteria. First, the Department must not use a per se rule that assumes a passthrough of benefits without examining the facts. Second, the factual examination must focus on the transaction that resulted in the “change of ownership,” and third, the Department must determine whether the purchaser received a subsidy for which it did not pay adequate compensation. Respondents argue that the Department’s use of the “same person” test violates these rules.

Respondents next argue that a recent WTO panel also determined that the Department’s use of the “same person” methodology contradicts the WTO Agreement on Subsidies and Countervailing Measures because this methodology essentially fails to examine whether the benefit has passed through to the privatized company. Respondents argue that the WTO panel held that there should be no distinction made between a company and its shareholders, as together they constitute a producer, a natural or legal person that may be the recipient of the benefit to be assessed. As a result, respondents argue that the Department’s methodology is contrary to the SCM agreement, as discussed by the WTO, because the Department’s methodology fails “to determine whether the privatized producer has received any benefit from a prior financial contribution bestowed on the state-owned producer.”

In the alternative, respondents argue that in applying the Department’s “same person” test, the record demonstrates that USIMINAS, COSIPA, and CSN were all fundamentally different “persons” after privatization. For example, respondents point out that all three mills adopted different management teams, structures and sales strategies after privatization; all three companies radically revamped their physical assets as a result of privatization; and, all companies cut their personnel since privatization. Respondents argue that these facts show that three of the four criteria used by the Department to measure whether the post-privatization entity is the “same person” as the pre-privatization entity show that these companies are not the same entities after privatization.

Petitioner’s Arguments: Petitioners argue that the equity infusions into the companies by the government of Brazil are countervailable because the Department properly applied its new change in ownership methodology to determine that the companies are the “same person” after as before privatization. Petitioners argue that the fact that the companies have made certain investments, or reduced their workforce, shows that they have striven to become more efficient. However, petitioners argue that these changes were made over time and were not part of the privatization process, but were in fact, part of the ongoing operations and business decisions of the companies.

Petitioners further argue that each of the factors that the Department considers in its “same person” privatization methodology argue in favor of finding that the companies are the same legal persons as they were before they were privatized. For example, petitioners argue that the companies’ actions to acquire interests in steel distributors or service centers, upgrade their steelmaking facilities, and increase and decrease certain parts of their operations, were part of the ongoing business operations of the companies, and not directly tied to the privatizations.

Petitioners further argue that there was essentially no change in the continuity of the production facilities of the companies, as indicated by the fact that the production of USIMINAS and COSIPA remained virtually unchanged in the year following their privatizations. According to petitioners, the same is true of CSN. Petitioners further argue that the companies’ responses show that the companies’ assets and liabilities did not change as a result of privatization, and that changes in the number of personnel was not a result of the privatization process.

Finally, petitioners argue that the Department’s “same person” methodology properly interprets U.S. law, and consequently, should be applied in the final determination as it was applied in the preliminary determination. Petitioners first argue that respondents failed to note that the most recent CIT decision addressing the legality of the Department’s methodology upheld that the “same person” test is consistent with Delverde III. As a result, because there are conflicting holdings from the CIT on the issue of the “same person” methodology, petitioners argue that the Department should continue to use its privatization methodology to determine whether benefits have passed through to the purchaser, as it did in the recent final decision in the Final Affirmative Countervailing Duty Determination and Final Negative Critical Circumstances Determination: Carbon and Certain Alloy Steel Wire Rod from Brazil, 67 FR 55805 (August 30, 2002) (Brazil Wire-Rod), as explained in the Issues and Decision Memorandum for the Final Determination in the Countervailing Duty Investigation of Carbon and Certain Alloy Steel Wire Rod from Brazil from Richard W. Moreland, Deputy Assistant Secretary, Import Administration, to Faryar Shirzad, Assistant Secretary, Import Administration,” dated August 23, 2002 at Comment 1: “Usiba and Cosinor Privatizations” (Brazil Wire-Rod Decision Memo).

Petitioners further argue that respondents’ argument that the Department must adhere to the WTO panel decision in United States-Countervailing Duty Measures Concerning Certain Products from the European Communities is misguided because this decision is subject to an appeal to the WTO appellate body. According to petitioners, 19 USC § 3533(g)(1) provides that “where a Panel or Appellate Body finds that a regulation or practice of the United States is inconsistent with any of the Uruguay Round Agreements, ‘that regulation or practice may not be amended, rescinded, or otherwise modified in the implementation of the report unless and until’ every one of six separate actions is taken.” Petitioners note that these actions include such things as consulting with the relevant private sector advisory committees, consulting with appropriate congressional committees, submitting a report to such congressional committees, and providing an opportunity for public comment on the proposed modification.

Department’s Position: We disagree with the respondents that our “same person” methodology is not in accordance with law or in conformance with the CAFC’s decision in Delverde III. In several recent cases, various judges of the CIT have ruled on the Department’s “same person” test. Some found that this methodology was not in accordance with the law and consequently remanded several cases to the Department for further consideration. See Allegheny Ludlum 182 F Supp. 2d 1357 (CIT 2002); GTS Industries S.A. v. United States, 182 F. Supp 2d 1369 (CIT 2002); AST- Stainless Steel Plate; ILVA Lamiere E Tubi S.R.L. and ILVA S.p.A. v. United States, Slip Op. 2002-32 (CIT 2002). However, in a more recent case, see Acciai Speciali Terni S.p.A., et al.,, Plaintiffs, v. United States of America, Defendant, and Allegheny Ludlum Corp., et al., Defendant-Intervenors, 2002 Ct. Intl. Trade Lexis 82; Slip Op. 2002-82, (August 6, 2002), the CIT affirmed the Department’s “same person” methodology.

All of these cases are subject to further appeal. Therefore, notwithstanding respondents’ arguments regarding the appropriateness of our “same person” methodology, until there is a final and conclusive decision regarding the legality of the Department’s change-in-ownership methodology, we have continued to apply it for purposes of this final determination.

We also disagree with the respondents that USIMINAS, COSIPA and CSN were not the same person following the change in ownership. As discussed above in the Changes in Ownership section, the first step in our change in ownership methodology is to determine whether the legal person, or the business entity to which the subsidies were given, is distinct from the business entity that produced the subject merchandise exported to the United States. In examining the continuity of general business operations, although respondents argued that there have been numerous changes made to the business operations of USIMINAS, COSIPA, and CSN, respondents also noted in their December 17, 2002 response that these changes were not a result of the privatization process.

We also examined the second and third criteria, continuity of production facilities and continuity of assets and liabilities. Respondents argued that all of the companies changed their production facilities and their assets and liabilities after privatization. Our review of the production information for all three of the companies under investigation indicates that there was little change in the quantity and composition of their production following privatization. For example, the comparative production data provided at pages 4-5 of USIMINAS’s 1992-1993 financial statement (exhibit 34 of the GOB's December 17, 2001 response) indicates that USIMINAS’s production totals declined slightly, by 1.6 percent, from 1991 to 1992, and that its product mix remained essentially unchanged for this period. In addition, there was only a slight change in its labor productivity ratio of 386 tons/man/year in 1992 (an increase of 3 over 1991). A similar review of COSIPA's 1993 financial statement, at pages 5 and 11, indicated that annual production of uncoated flat-rolled steel products remained steady, declining slightly from 2.6 in 1992 to 2.5 million tons in 1993.

Furthermore, although the privatization of USIMINAS, COSIPA and CSN were accomplished through the sale of the GOB’s shares to private investors, according to the respondents, the privatization of these companies involved the purchasing of shares of an ongoing corporation. As a result, the new shareholders of these companies continued to maintain an ownership interest that included both the assets and liabilities of the privatized companies. Therefore, the assets and liabilities of USIMINAS, COSIPA, and CSN remained intact after privatization. Finally, we examined the fourth criterion under our “same person” methodology, retention of personnel, and we found that, while there is information that indicates that the management and personnel of these companies may have been altered as a result of their privatizations, we do not consider these changes to be sufficient to find that USIMINAS, COSIPA, and CSN were different entities after privatization.

Based on the totality of the factors considered, we determine that the pre- and post-sale USIMINAS, COSIPA, and CSN were continuous business entities because they were operated in substantially the same manner before and after the change in ownership. Although it is evident that long-term changes are being carried out by all three companies, a comparison of the companies prior to and subsequent to the sale indicated that the three entities were the same person. Because USIMINAS, COSIPA, and CSN are the same legal person after their privatizations as they were before, all of the requirements for countervailing pre-privatization subsidies continue to be met. Thus, to the extent that subsidies received by USIMINAS, COSIPA, and CSN prior to the POI are allocable to the POI, we are attributing them to USIMINAS’s, COSIPA’s, and CSN’s sales during the POI.

Comment 2: PIS/COFINS - Direct Taxes Versus Indirect Taxes

Respondents’ Argument: Respondents argue that, as in the preliminary determination, the Department should continue to treat the PIS/COFINS taxes as indirect taxes in the final determination. Respondents argue that simply because some of the revenue generated from these taxes is used for social programs does not alter the classification of these taxes, i.e., they are still indirect taxes. Respondents point out that in every case concerning treatment of prior-stage cumulative indirect taxes, the Department has evaluated indirect taxes, such as sales, production, or distribution taxes, but never once questioned how the tax revenues were used in determining whether the taxes were indirect. Respondents argue that in this case if the PIS/COFINS taxes are properly evaluated on the basis of how the tax is levied, the Department must find that the PIS/COFINS taxes are indirect taxes.

Respondents further argue that even if the Department decides to evaluate the treatment of PIS/COFINS taxes based on how the revenues are spent, the record demonstrates that these taxes are used for general programs and not strictly as social welfare charges. Additionally, respondents point out that technically the PIS and COFINS are “social contributions,” and not “social welfare charges,” and therefore do not fall withing the meaning of “social welfare charges” in the Department’s regulations.

Petitioner’s Argument: Petitioners argue that the information learned by the Department at verification shows that COFINS and PIS taxes should be treated as social welfare charges because GOB officials said that the PIS and COFINS taxes are designated by law as “social contributions.” See GOB Verification Report at 3 (August 9, 2002). Petitioners further argue that GOB officials explained at verification that social contributions are “every form of tax revenue raised via either a payroll tax or a consumption tax that is allocated and intended to be used for social security assistance, welfare, or health programs for the lower income sectors of the population.” See Id at 2. Petitioners lastly argue that the fact that 20 percent of PIS and COFINS taxes are unearmarked and are available for uses outside of the social security budget indicates that some portion of these taxes are used for social welfare purposes, and PIS and COFINS should therefore be treated as direct taxes within the meaning of the Department’s regulations.

Department’s Position: Based on our analysis of all the information on the record, we find that the PIS and COFINS taxes should be treated as indirect taxes. The Department’s regulations at section 351.102(b) provide several specific definitions which are applicable to PIS and COFINS. According to section 351.102(b), “indirect tax means a sales, excise, turnover, value added, franchise, stamp, transfer, inventory, or equipment tax, a border tax, or any other tax other than a direct tax or an import charge.” Although the information on the record indicates that the purpose of these taxes is for social welfare, PIS and COFINS taxes are calculated and assessed upon the gross revenue of a company, and therefore, PIS and COFINS fall within the definition of an indirect tax.

The Department’s regulations further define both “cumulative indirect tax” and “prior-stage indirect tax.” See section 351.102(b) of the Department’s regulations. A “cumulative indirect tax” is a “multi-staged tax levied where there is no mechanism for subsequent crediting of the tax if the goods or services subject to tax at one stage of production are used in a succeeding stage of production.” A “prior-stage indirect tax” is defined as “an indirect tax levied on goods or services used directly or indirectly in making a product.” Therefore, because PIS and COFINS are charged on inputs used to make cold-rolled steel, they are charged on goods at one stage of production that are used in a succeeding stage of production, thus falling within the definitions of “cumulative indirect tax” and “prior-stage indirect tax.” PIS and COFINS are even described as indirect taxes by independent accounting firms. In a publication on Doing Business in Brazil, the accounting firm of Price Waterhouse Coopers lists PIS and COFINS as two of a number of indirect taxes levied in Brazil. See Doing Business Guide: Brazil, Price Waterhouse Coopers, in “Appendix on Taxes in Brazil” (2001). As such, we find that PIS and COFINS should be treated as indirect taxes and not as direct taxes.

Comment 3: PIS/COFINS- Rebate of Prior-Stage Cumulative Indirect Taxes

Respondents’ Arguments: Respondents argue that the GOB monitoring system is reasonable and effective in overseeing the application of claimed PIS and COFINS tax credits. Respondents note that in the preliminary determination, the Department found that the GOB did not have an adequate system in place because it was “not effective for calculating the credit corresponding to the actual inputs consumed in the production process.” However, respondents note that since the preliminary determination, the Department has gathered and verified additional information that the GOB monitoring system was reasonable and effective in overseeing the application of the tax credits.

Respondents argue that the Revenue Secretariat closely monitors claimed tax credits consistent with its goal to minimize credits and maximize tax revenues. Respondents note that the Secretariat’s monitoring involves both on-going, real-time monitoring of claimed credits, and onsite audits. Respondents argue that in the present case, the system that is used by the GOB to track the PIS/COFINS credits, and the manner in which exporters can claim the credit are similar to the system that was examined in Final Countervailing Duty Determination: Polyethylene Terephthalate Film, Sheet, and Strip from India, 67 FR 34905, (May 16, 2002) (PET Film from India). Respondents note that in PET Film from India, the Department found that the burden of showing eligibility for the tax exemption lies with the exporter to demonstrate its eligibility for the tax exemption, but the government tracks exempted indirect taxes through quarterly reporting by exporters.

Respondents argue that the GOB’s system is “reasonable and effective for the purposes intended.” Respondents state that, because the Secretariat’s objective is to ensure compliance with the laws and regulations of the presumed credit in order to minimize credits and maximize revenues, the GOB’s main goal is to avoid excessive credits.

Respondents further argue that, although the GOB’s system for monitoring the presumed credit is reasonable and effective for the purposes intended, the GOB has also satisfied the requirement, at section 351.518(a)(4)(ii), that the “government has carried out an examination of actual inputs involved to confirm which inputs are consumed in the production of the exported product, in what amounts, and which indirect taxes are imposed on inputs.” Respondents note that if either of these two prongs are met, the Department may not consider the entire amount of the exemption to confer a benefit. Respondents argue that the GOB has several methods of examining the consumption of inputs in the production of exported product and the indirect taxes imposed on those inputs. First, respondents argue that the Secretariat’s reviews of actual inputs used in the exported products, along with the Secretariat’s audits of company annual tax returns and special production audits, qualify as an adequate examination of the consumption of inputs that are used in the production of the exported product and in what amounts.

Respondents further argue that in Certain Cold-Rolled Carbon Steel Flat Products From Argentina, exporters submitted to the administering authority documentation of their tax incidence on a regular basis, and this was used to evaluate the allowable exemption. Respondents note that this constituted “an appropriate examination of actual inputs to confirm which inputs are consumed in the production of the exported products.” See section 351.518(a)(4)(ii) of the Department’s regulations. Respondents argue that the similar requirements of Brazilian exporters in claiming the PIS/COFINS credits shows that the GOB carries out an appropriate examination of the inputs consumed in the production of the exported product, in accordance with section 351.518(a)(4)(ii) of the Department’s regulations.

Petitioners’ Arguments: Petitioners argue that Brazilian Federal Revenue Secretariat’s monitoring of claimed credits and on-site audits to ensure that the inputs and outputs reported to the GOB are correct, do not provide a reasonable system to examine the inputs used in the production of exported merchandise, and do not assess the amount of PIS/COFINS tax that is paid on the purchase of inputs. Petitioners further argue that the GOB has not examined the actual inputs for which the tax credits are claimed to confirm which inputs are consumed in the production of the exported product, and in what amounts. Petitioners point out that the GOB acknowledged in a questionnaire response that it makes no attempt to identify these inputs, because it would be impossible to examine every input, and because these inputs vary from industry to industry.

Petitioners further argue that the GOB’s description of the PIS/COFINS taxes at two stages of processing as a “rule of administrative convenience,” is not effective for the purposes of calculating the appropriate credit for taxes incurred on inputs consumed in the production of exported merchandise. Petitioners further argue that the Department should reach the same conclusion as it did in the Brazil Wire Rod decision by determining that the Brazilian system is not reasonable and effective for the purposes intended because there is no accounting of PIS/COFINS taxes for inputs that are not consumed in the production of the imported inputs on which these taxes were never paid, and the system does not make due allowances for waste.

Department’s Position: As discussed in Comment 2, supra, we have determined that PIS and COFINS are indirect taxes and that the PIS/COFINS rebate should be analyzed under the requirements of section 351.518 of the Department’s regulations. Pursuant to that section, we must find the entire amount of the PIS and COFINS credit countervailable unless we find that the GOB “has in place and applies a system or procedure to confirm which inputs are consumed in the production of the exported products and in what amounts, and to confirm which indirect taxes are imposed on these inputs, and the system is reasonable, effective for the purposes intended, and is based on generally accepted commercial practices in the country of export.” See section 351.518(a)(4)(i) of the Department’s regulations. These provisions establish specific criteria and standards, all of which must be met to find the credit of prior-stage cumulative indirect taxes not to be a subsidy. Record evidence establishes that the GOB has not demonstrated in this investigation that its system for rebating PIS and COFINS taxes meets all of these criteria, or that the information submitted to show that the actual tax incidence is lower than the rebated amount effectively does so for the producers of the subject merchandise.

The GOB has claimed that the system it applies to rebate PIS/COFINS taxes to Brazilian exporters meets the system requirements identified in the Department’s regulations. To determine the monthly PIS/COFINS tax rebate under this program, Brazilian companies first establish their total monthly purchases of eligible inputs used to produce all products produced by the firm. “Eligible” inputs for which rebates may be claimed are defined by the PIS/COFINS law as packing materials, intermediate products and raw materials. The value of these total input purchases is multiplied by the company’s ratio of export revenue to total revenue. The resulting amount is then multiplied by 5.37 percent, the “presumed” PIS/COFINS tax incidence, to determine the rebate amount that can be claimed. Brazil has also reported that companies are subject to production audits and must demonstrate that these inputs are being properly allocated in accordance with the law. See GOB Verification Report at 5.

We disagree that the system requirements in the Department’s regulations are in fact being met by the GOB in this case. First, this system and the rebate rate were established as a simplified and streamlined methodology to determine the amount of the tax rebate for all companies in Brazil. The only limitation imposed on companies making rebate claims is that the claims be limited to those inputs defined under the PIS and COFINS rebate law, which is broader than the “consumed in production” standard provided for in the Department’s regulations at section 351.518(a)(1). In fact, evidence on the record of this investigation shows that companies include on their claims purchases that are not consumed in production. E.g. see Exhibit 3 of February 5, 2002 Companies’ Response. Companies report their purchases of inputs based on the assumption that all goods purchased are consumed equally in exported and domestically sold goods. Further confirmation is not conducted by the government. As such, we find that this system does not permit the GOB to confirm which inputs are being consumed in the consumption of exported goods and in what amounts.

We also have additional concerns that this system does not operate in accordance with the requirements under the regulations. The system does not account for the fact that domestic and export sales may include imported inputs. These imports may be included by varying degrees in either export or domestic sales, thus distorting the ratio. Either way, the GOB does not account for, and does not require Brazilian companies to account for, any such potential distortions. Further, in determining the actual amounts of inputs consumed in final products, the government does not make due allowance for waste, thereby raising the concern that the claim amount is overstated. This is an important element that the Department has addressed in countervailing another rebate program in the past. See, e.g., Thai Hot-Rolled Final at “Comment 3: Countervailability of Section 36(1) Benefits.”

Respondents have argued that the facts of PET Film from India, 67 FR at 34905, are similar to the instant case, and that the monitoring system which identifies and confirms the inputs used in the tax credit calculations exceeds the standards found acceptable in that case. See also PET Film From India Decision Memorandum at “SUP Programs” (May 16, 2002) (PET Film Decision Memo). We disagree. As explained in the Brazil Wire-Rod Decision Memo at Comment 9: “PIS and COFINS-Excessive Remission,” the system examined in PET Film from India is distinguishable from the GOB’s system. The system applied by the Government of the State of Uttar Pradesh (“SUP”) establishes a more rigorous standard for tracking both the inputs consumed in the production of exported products and the indirect taxes from which those inputs are exempted. First, unlike the PIS/COFINS program which rebates taxes levied on inputs, Indian exporters are exempt from paying taxes at the time they purchase their inputs. Exporters are required to track, on a specific form, all of their raw material purchases that are destined for consumption in exported products, and for which they have not paid sales taxes. Further, exporters file monthly state tax returns with the state tax authorities detailing the amount of taxes a company collects on its sales of subject merchandise and the amounts it pays on purchases of inputs. These product-specific tax returns are then audited by the tax authorities. Finally, the SUP conducts on-site inspections, during which exporters must demonstrate that the inputs exempted from taxes were used in the exported products. See, PET Film Decision Memo, at Comment 10: “Benefit of Sales Tax Incentives for Exports Under Section 4-B of the Uttar Pradesh Trade Tax Act.” This rigor of detail and audit is not applied by the GOB in its remission of PIS/COFINS taxes to Brazilian exporters.

We note that respondents have submitted two studies on the actual incidence of PIS and COFINS within various Brazilian industries, including “steelmaking.” Respondents have stated that these studies were submitted for the purpose of confirming that the remission of PIS and COFINS does not exceed the amount of PIS and COFINS paid on inputs that are consumed in the production of the exported product, making normal allowance for waste, under section 351.518(a)(2) of the Department’s regulations. These studies do not alter our conclusion that the PIS/COFINS rebates are countervailable. First, because we have found that the system requirement of 19 CFR 351.518(a)(4)(i) is not met, we do not reach as a legal matter the question of whether the remission of PIS and COFINS is excessive. Second, even if we were to consider the studies at issue, they raise the same fundamental concerns that are found in our analysis of the system requirement. These concerns are the over-inclusion of inputs, i.e., inputs not consumed in the production of exported products, and whether the system provides for a normal allowance for waste exists. In establishing the non-excessive remission or prior-stage cumulative indirect taxes, both of these elements must be accounted for. The GOB has not demonstrated in this investigation that these requirements are in fact sufficiently accounted for in the two studies.

Comment 4: PIS/COFINS Credit - Excessive Remission

Respondents’ Argument: Respondents argue that the facts on the record of this case demonstrate that the presumed credit is not excessive generally and as applied by the companies under investigation. Respondents argue that under section 351.518(a)(2) of the Department’s regulations, only excessive credits, or the difference between the amount remitted and the actual indirect taxes paid, are countervailable. Respondents state that the maximum amount that could be construed as excessive is the difference between the credit rate of 5.37 percent and the combined PIS/COFINS rate of 3.65 percent, which results in a difference of 1.72 percent, because it is indisputable that PIS and COFINS taxes are paid at least once on every input purchased domestically. Respondents further argue, however, that to the extent that any input passes through more than one stage of processing prior to its consumption in steel, the actual incidence of taxation on the input will be higher than the 5.37 percent credit. Respondents state that most inputs used in the production of the subject merchandise pass through more than one stage of processing.

Respondents argue that the two independent studies submitted in this case confirm that the PIS/COFINS credit is not excessive. Respondents note that the goal of each study was to analyze and quantify the negative impact of cascading indirect taxes, like PIS and COFINS. Respondents argue that the study that was done from data collected by the Brazilian Institute of Geography and Statistics (“IBGE”), by using data on company-specific purchasing, production, and sales data for various sectors of the economy, permit a sector-specific analysis of the various stages through which inputs pass before incorporation into the finished good, as well as the associated indirect tax incidence at each stage of the production chain.

Respondents further argue that the study prepared by the Instituto de Pesquisa Economica Aplicada (IPEA, or the “Applied Economic Research Institute”) determined that the hidden tax incidence on exported steel products is 8.82 percent, which leaves 3.53 percent of the taxes paid uncompensated by the 5.37 percent PIS/COFINS credit. Respondents argue that the FIESP study found that the actual tax incidence for exported steel products is 9.8 percent, which leaves 2 percent uncompensated by the PIS/COFINS credit.

Respondents argue that even if the credit were not excessive, at verification the Department found that two of the companies under investigation claim less than the full amount of the credit allowable under Brazilian law. Respondents therefore argue that whether the Department examines the credit calculation itself, empirical studies by independent researchers, or actual application of the credit in this case, the Department should find that the credit did not exceed the actual taxes paid.

Petitioners’ Arguments: Petitioners argue that respondents’ claim that the PIS/COFINS tax credits are not excessive is irrelevant, because, as discussed in the Brazil Wire Rod decision, once the Department determined that an effective system for calculating a tax credit has not been met, the Department need not reach the question of whether the remission of the PIS/COFINS taxes were excessive.

Petitioners further argue that respondents’ arguments that the taxes are paid on at least two stages of production is flawed because some of the inputs that are used may be imported or internallyproduced and are therefore not subject to PIS/COFINS taxes. Petitioners further argue that the two studies submitted on the record by respondents that purport to show that PIS/COFINS credits are not excessive do not analyze the inputs used in the production of cold-rolled steel, and these studies do not contain any information relating specifically to any of the companies. As a result, petitioners argue that the full amount of the tax credits given to the companies for their export sales must be countervailed.

Department’s Position: According to section 351.518(a)(4) of the Department’s regulations, if we find that the system requirement is not met, as a legal matter we do not reach the question of whether the remission of PIS and COFINS is excessive. Because we have found that the system requirement is not met, we are not addressing respondents’ argument that the PIS/COFINS tax credit is not excessive.

Comment 5: FINEM Financing and Specificity

Respondents’ Argument: Respondents contend that the FINEM program is not de facto specific to the steel industry because the steel industry was not the predominant user, nor did it receive a disproportionate share of the FINEM financing. Respondents argue that data provided by the GOB showing FINEM loan distribution from 1986 through 2000 indicate that a significant number of other industries received this financing without any one industry receiving a dominant share. Respondents state that FINEM provided broad financing in over thirty different industry sectors, and that in each year respondents received financing, the steel industry accounted for only a small share of the total financing as compared to other sectors which received larger sums from BNDES.

Department’s Position: Our review of the GOB-provided breakdown of FINEM financing by industry indicates that the steel industry was neither a predominant user nor disproportionate recipient of FINEM financing in the years in which respondent companies received approval for loans that were still outstanding during the POI. We therefore, agree with respondents that this financing is not de facto specific, and therefore, is not countervailable as a domestic subsidy under section 771(5A)(D)(iii) of the Act. This financing is available to all industries in Brazil on a de jure basis, and is, in fact, actually used by a wide array of industries. Moreover, the steel industry is neither a predominant nor a disproportionate recipient.

Comment 6: FINAME as an Import Substitution Program

Petitioners’ Argument: According to petitioners, FINAME is an import substitution program even if it allows for the financing of foreign or domestically-produced equipment because BNDES provides separate financing lines, under different terms and sources of funds which discriminate between domestic and imported equipment based on the domestic content of the merchandise. Furthermore, petitioners note that in Brazil Wire Rod, the Department found FINAME to be an import substitution program because this financing is overwhelmingly used for the purchase of domestically-produced products; it is available for the purchase of imported equipment only when domestic equipment is unavailable; and, in the case of leasing, is only available for domestic equipment.

Respondents’ Argument: Respondents argue that in order for FINAME to be considered an import substitution program, the statute requires a program to impose, as a condition for obtaining the subsidy, that domestic goods be used. FINAME loans are not a subsidy, according to respondents, because these loans are not contingent on the purchase of domestic goods, as FINAME financing was applied equally to foreign and domestic equipment through two different aspects of the program. Respondents state that FINAME can finance the acquisition of imported equipment if the loan recipient demonstrated domestic unavailability of the equipment. Further, respondents note that this program was broadened in the mid 1990's to allow for loans to support the acquisition of both domestic and imported equipment regardless of domestic availability.

Department’s Position: We determine that FINAME loans are specific because they constitute an import substitution subsidy within the meaning of section 771(5A)(C) of the Act. See Discussion on FINAME loans in Programs Determined To Confer Subsidies section, supra. Our finding is consistent with those made in Brazil Wire Rod, 67 FR at 55805. In Brazil Wire Rod, the Department found that almost all FINAME financing is provided for new machinery and equipment manufactured in Brazil, and financing for equipment leasing is only available for leasing of domestic equipment. See Brazil Wire-Rod Decision Memo in the “Analysis of Programs” section. We also verified that when this program was implemented, there was a separate provision which only allowed for the financing of imported equipment if domestically manufactured equipment is unavailable in Brazil. See GOB Verification Report at 23.

In 1997, a second line of FINAME financing specifically for imported equipment became available which no longer required a prospective borrower to demonstrate the unavailability of Brazilian-made equipment. However, as discussed above in the Programs Determined To Confer Subsidies section, supra, this financing is only available to “micro, small and medium size companies” in Brazil, and therefore is not available to large companies like the respondents. Thus, the respondent companies are only eligible for the original line of FINAME financing, which is available for imported equipment only when domestic equipment is unavailable in Brazil. Therefore, we continue to find this program to be specific as an import substitution subsidy for large companies in Brazil.

Comment 7: FINAME Financing and Specificity

Respondents’ Argument: Respondents argue that FINAME financing is not countervailable since it is widely distributed and therefore, not de facto specific. Respondents argue that data provided by the GOB which shows the breakdown of FINAME financing by industry sector for the years 1986 through 2000 for the years in which respondent companies received financing, indicates that the steel industry was not the predominant nor a disproportionate recipient of FINAME financing. Specifically, respondents note that the steel industry’s share of the total funding never exceeded 6.79 percent of the annual total, ranking it from the fourth to the twelfth largest recipient depending on the year. In all, thirty to forty industry sectors obtained financing which was distributed among all sectors over this period.

Department’s Position: Because the Department finds FINAME financing to be de jure specific as an import substitution subsidy under section 771(5A)(C) of the Act, we have not addressed respondents’ arguments on whether FINAME loans are de facto specific.

Comment 8: Integral Linkage of FINAME and FINEM

Respondents’ Argument: Respondents argue that the Department should find FINAME and FINEM financing under BNDES to be integrally linked as one program, and should examine their specificity as though they are one program, in accordance with section 351.502(c) of the Department’s regulations. According to respondents, FINEM complements FINAME to ensure that all sizes of business in all types of industries can receive financing regardless of whether the equipment is domestically produced. FINEM provides financing for fixed asset investments in excess of seven million reais, whereas FINAME provides loans for machinery and equipment, and allows for loans less than seven million reais under FINAME Automatico program.

Respondents contend that the Department verified this linkage, as shown in the verification report which states that “BNDES often suggests to the borrow{er} that FINEM may be a better means of obtaining financing.” In addition, respondents state that FINAME loans were designed to pick up where FINEM loans left off by offering financing for equipment and machinery less than or equal to seven million reais.

Respondents point to similarities between the programs which demonstrate that they should be considered one program. Specifically, respondents note that both programs provide the same type of benefit in financing equipment and machinery; require similar qualifications in that the companies must operate in Brazil; and, adopt the same credit policies and procedures. As such, respondents argue that these programs are linked and provide companies complete access to loans whether those loans are below or above seven million reais.

Finally, respondents state that the Department should find that the provision of loans under FINAME and FINEM together does not meet the statutory criteria for de facto specificity. According to respondents, the record evidence indicates that participation in this program is not and has never been limited to only a few industries, with as many as thirty to forty sectors participating in any given year. Respondents further add that the steel industry was not the predominant or disproportionate user of these programs as reflected in the broad dispersion of financing among all sectors. Specifically, respondents noted that the steel industry’s total funding for these programs combined ranged from 2.82 percent to 10.74 percent, ranking it from fourth to thirteenth largest recipient in any given year, with other sectors receiving significantly larger shares of all FINAME and FINEM loans, representing as much as 22 to 38 percent of the total.

Petitioners’ Argument: In examining the factors listed under section 351.502(c) of the Department’s regulations, petitioners contend that the evidence on the record of this investigation does not meet the requirements necessary for a finding of integral linkage between the FINEM and FINAME programs. According to petitioners, the purpose of the FINEM program is fundamentally different from that of the FINAME program which, unlike FINEM, is used primarily to finance the purchase or leasing of machinery and equipment. In addition, petitioners note that these two programs were not linked at their inception, but were created at different times. FINAME did not come into existence until 1966, while petitioners believe that information on the record indicates that FINEM was established when BNDES was created in 1952.

Department’s Position: We find that the record evidence does not demonstrate that the FINEM and FINAME programs are integrally linked pursuant to section 351.502(c) of the Department’s regulations. Each program has a specific purpose distinct from the other. The FINEM program’s purpose is to support fixed asset investments and working capital financing for companies that intend to increase their inventory or develop new projects, whereas FINAME is strictly limited to the financing of machinery and equipment acquisitions without any provisions related to expansions or new projects. See GOB Verification Report at 21 and 23.

Although both programs provide similar benefits in the form of loans, the level of benefits and the manner of financing used in each program are generally different. The vast majority of FINEM financing is directly financed through BNDES in excess of seven million reais for project funding, including construction, engineering, and working capital projects, as well as studies that support special projects. Id. at 21. In contrast, much of FINAME financing is provided through agent banks, for the purchase of Brazilian-made machinery and equipment. Often the amount of such financing is less than 7 million reais. Id. at 24. We therefore, conclude that these two programs are not set up to confer similar levels of benefits on similarly situated firms.

Finally, there is no evidence on the record of this investigation that indicates that these two programs were linked at their inception. Based on our analysis, we do not find the FINEM and FINAME programs to be integrally linked and have therefore, continued to examine their specificity on an individual program basis.

Comment 9: Financial Contribution and Benefit of BNDES Loan Programs

Petitioners’ Argument: Petitioners state that respondents’ data contained in exhibits 3, 4, and 5 of its case brief show that there were numerous instances where the interest rates charged on BNDES FINAME and ExIm loans were lower than respondents’ comparable commercial loans. Petitioners argue that FINAME loans provided a financial contribution through the direct transfer of funds from the GOB. In Brazil Wire Rod, petitioners note, the Department found that the funds being loaned through the FINAME program were GOB funds; that the largest portion of the financing charges paid by the borrower are paid to the GOB; and that, while these loans are processed through agent banks, they must be approved by BNDES.

Petitioners add that even if the Department does not consider FINAME loans to be a direct transfer of funds, these loans would still provide a subsidy under article 1677(5)(B)(iii) of the Act because the GOB used the agent banks as a funding mechanism to provide a financial contribution. According to petitioners, the agent bank function is one that would otherwise be vested in the government and is similar in nature to other GOB programs such as FINEM which provides funding directly.

Respondents’ Argument: Under the statute, respondents state that a program must provide a financial contribution through a “direct transfer of funds” from the government in order to be considered countervailable. Respondents state that no financial contribution is conferred due to the intermediate actions of agent banks which have a great degree of latitude in deciding the terms of FINAME loans. Therefore, FINAME loans cannot be considered a direct transfer of funds since their administration, approval, and servicing is largely handled by independent, nongovernmental financial institutions. Specifically, respondents note that these agent banks are responsible for analyzing the borrower’s operation and assessing the risk level, disbursing the money, and carrying the loan risk. In addition, there is no government limitation on or participation in setting the agent’s spread on the interest rates charged on these loans since the agent spread is exclusively negotiated between the agent lending institution and the client.

Respondents contend that a legal and procedural separation exists between FINAME and the companies requesting the loans. There have been several Brazilian Superior Court decisions, according to respondents, that have prohibited companies from including FINAME in their lawsuits made against the agent financial institutions, indicating that FINAME is not held legally liable to these companies requesting the loans. Respondents add that FINAME deals only with the financial institutions who play a large role in the development and control of the loans. Respondents contend that the BNDES-ExIm loans are similar to commercial export financing and confer no special benefit to respondents. Respondents state that BNDES-ExIm financing is made through a network of accredited financial agents that comprise nearly all the banks in Brazil and operate in conjunction with over 75 foreign banks. Companies apply to the agent banks, and BNDES and the agent bank review the application to determine if they meet the eligibility criteria. The funding for the loans is provided by BNDES, but respondents state that the agent bank is responsible for the loan and evaluating the credit risk which is covered by the agent spread. According to respondents, only COSIPA and USIMINAS obtained BNDES-ExIm loans for pre-shipment financing used for the production of goods for export that are linked to specific shipments.

Respondents state that export financing is readily available from commercial lenders because these loans are collateralized by the exported good. Thus, these loans have seniority in relation to other loans in the event that Brazil experiences problems servicing its external debt. In addition, respondents add that both Brazilian and foreign banks commonly offer export financing to large exporters like the steel industry.

According to respondents, record evidence indicates that the associated cost of BNDES financing obtained by each respondent company exceeded the cost of their comparable commercial loans. Therefore, respondents received no countervailable benefit from these loans in accordance with section 351.505(a)(1) of the Department’s regulations. Respondents argue that the companies’ commercial loans are comparable because they satisfy the first two criteria of section 351.505(a)(2) in that they match BNDES financing by its structure and maturity of the loans (variable interest rate and long-term).

Respondents note that some of the companies’ BNDES financing was denominated in Brazilian currency or BNDES’s basket currency (UMBNDES), whereas the commercial loans were in U.S. dollars. Respondents state that the Department has found BNDES financing to be the only source of long-term lending in reais used in Brazil, and as such, the Department has used the World Bank’s average private lending rate in U.S. dollars in Brazil as a benchmark. See Preliminary Negative Countervailing Duty Determination: Carbon and Certain Alloy Steel Wire Rod from Brazil (Brazil Wire Rod Preliminary Determination), 67 FR 5967, 5970 (February 8, 2002). By using this rate, the Department would ignore company-specific loans simply because they are denominated in a different currency, and would be using a broad industry average that is not representative of the company’s actual private financing experience, according to respondents. Respondents also state that a country-wide average would include loans from all types of Brazilian companies which would unfavorably skew the comparison since the Department found at verification that large Brazilian companies obtain financing at better terms than smaller firms.

Respondents state that although BNDES-ExIm loans are export-specific, no benefit was provided by these loans since their financing costs exceeded those found in commercial loans. For example, according to respondents, a commercial export loan to USIMINAS examined at verification has an interest rate slightly lower than the effective interest rate on USIMINAS’s two BNDES-ExIm loans received during the POI. A similar examination undertaken by respondents of COSIPA’s ExIm loans also shows that the interest rate exceeds the interest rate on COSIPA’s comparable commercial loans.

Respondents argue that the financial cost of respondent companies’ FINAME and FINEM loans exceeded the cost of their comparable commercial loans. Respondents calculated a weightedaverage interest rate for each year that a respondent company obtained FINAME or FINEM financing using the commercial loans reported by each respondent company or the information from its financial statements, to demonstrate that the cost of its private financing is less than the cost of its FINAME or FINEM loans. Respondents add that even if the Department decides to use the World Bank Debt Tables in lieu of each company’s cost of private financing, those rates are also generally lower than either the FINAME or FINEM rates.

Department’s Position: As noted in the discussion under Programs Found Not Countervailable above, we find that FINEM loans are not specific and therefore, not countervailable as a domestic subsidy. Accordingly, we have not addressed respondents’ arguments regarding this program. However, we do find FINAME and BNDES-ExIm loans to be specific as an import substitution subsidy and export subsidy, respectively, as discussed in the Programs Determined to Confer Subsidies and therefore, we have examined respondents’ arguments concerning the existence of a financial contribution and benefit for these two programs, below.

We disagree with respondents that FINAME and ExIm loans do not provide a financial contribution because there was no “direct transfer of funds.” Consistent with the Department’s determination in Brazil Wire Rod, the funds being loaned through the FINAME and ExIm program are GOB funds; the largest portion of the financing charges paid by the borrower are paid to the GOB; the GOB established the “financial cost” and the “basic spread” which comprise most of the charges on these loans as compared to the portion represented by the agent bank’s spread; and, these banks are approved by BNDES and act as agents on behalf of the GOB. See Brazil Wire Rod Decision Memo in the “Analysis of Programs” section. We find that the FINAME and ExIm loans provide a financial contribution in the form of a direct transfer of government-provided funds in accordance with section 771(5)(D)(i) of the Act.

According to section 351.505(a)(5) of the Department’s regulations, in order to determine whether a long-term variable rate loan confers a benefit, we must first compare the rate on the government-provided loan for the year in which the government loan terms were established, to a variable long-term interest rate on a comparable commercial loan taken out in the same year as the government-provided loan, if such commercial loans were taken out. If the comparison shows that the effective interest rate on the government-provided loan is equal to or higher than the effective interest rate on the comparable commercial loan, the Department will determine that the government-provided loan did not confer a benefit. However, if the interest rate on the government-provided loan is lower, we will examine the loan to measure the benefit in the POI in accordance with section 351.505(c)(4) of the Department’s regulations. Finally, if the Department is unable to make a comparison using the comparable benchmark, section 351.505(a)(5)(ii) allows us to modify our method in order to find a bases to measure the benefit. As discussed above in the Benchmark for BNDES Loans section, we examined all three respondent companies’ FINAME and BNDES-ExIm loans in this manner to determine if a benefit was received from these loans.

We found that in years in which no comparable commercial loans were available, the most appropriate benchmarks to use were the respondent companies’ long-term, variable rate commercial loans that were approved in 2000 and verified by the Department. The long-term financing information reported in respondents’ annual reports was too general to use as a benchmark because there were many years for which we could not determine the composition of the funding for these long-term loans and/or notes, or for which we were unable to segregate the BNDES financing from the total amount. In addition, the loan information in these annual reports was not specific enough for us to ascertain the dates on which these loans were approved. We also did not find the loan information contained in the World Debt Tables to be a better alternative since this information uses predominantly fixed-rate loans that are not as appropriate as the long-term variable rate commercial loans taken out during the POI which are more similar in loan structure to the variable-rate BNDES financing.

Comment 10: BNDES-ExIm Financing and Specificity

Petitioners’ Argument: Petitioners note that USIMINAS and COSIPA both received export financing from the GOB through BNDES. By definition, petitioners argue, these loans meet the specificity requirement under the statute as an export subsidy, and the Department should find them countervailable in the final determination.

Department’s Position: Since receipt of BNDES-ExIm financing is contingent upon export performance, we find this program to be an export subsidy pursuant to section 771(5A)(B) of the Act. See discussion in Programs Determined to Confer Subsidies, section, supra.

Comment 11: BNDESPAR Program

Respondents’ Argument: Respondents state that the loan originating from BNDESPAR, a subsidiary of BNDES, is actually a term sale of assets made through an agent bank for the purpose of financing the purchase of privatization currency from BNDES as part of Brazil’s National Privatization Program (NPP). The original BNDESPAR loan in question was made to Banco Bozano Simonsen (“Bozano”), the agent bank, to purchase shares in COSIPA which were then resold to Brastubo, a Brazilian pipe and tube manufacturer. In September 1993, Brastubo sold some of the COSIPA shares, representing 49.8 percent of COSIPA’s voting capital, to Anquilla, its wholly-owned subsidiary. Respondents state that USIMINAS bought Anquilla and assumed its indebtedness to Bozano in connection with the privatization currencies used to obtain shares in COSIPA. USIMINAS later dissolved Anquilla and became a direct shareholder of COSIPA. According to respondents, the same terms in the initial contract between BNDES and Bozano are present in the subsequent contracts between USIMINAS and Bozano, and USIMINAS and BNDES.

Respondents argue that there was no financial contribution associated with this loan due to the intervening involvement of a private financial institution. According to respondents, all financial institutions were eligible to buy these privatization currencies for resale to third party investors without any restrictions. Respondents contend that no financial contribution was provided since this program did not involve a direct transfer of funds to USIMINAS from the GOB. Moreover, BNDES was not involved in USIMINAS’s subsequent assumption of Bozano’s debt when it obtained Anquilla’s shares in COSIPA. The transaction between USIMINAS and Bozano was distinct from the direct transfer of funds to Bozano from BNDESPAR.

Finally, respondents contend that BNDES financing of privatization assets through BNDESPAR did not confer a countervailable benefit because it was part of a privatization process that involved broad participation by financial agents and resulted in significant premiums being paid for shares in privatized companies at the highest market price. According to respondents, the distribution of this financing was widespread, as shown by record evidence indicating that no one financial agent or privatized company predominated in this process. As such, the record demonstrates that this financing was not specific to the steel industry as defined in the statute.

Department’s Position: As discussed in more detail in the Programs Determined to Confer Subsidies section, supra, we find this program to be de facto specific under section 771(5A)(D) of the Act. In 1993, the year that the original BNDESPAR loan was provided to another company to purchase shares in COSIPA, BNDESPAR funds went to a limited number of companies in a few industries for investments in a small number of state-owned enterprises. See BNDES verification exhibit 8 submitted July 3, 2002; also Calculations Memorandum for the Final Determination: Certain Cold-Rolled Flat-Rolled Carbon Quality Steel Products from Brazil which contains detailed information on the actual number of industries and companies involved. Information on loan approvals in 1995, the year in which USIMINAS refinanced this loan directly with BNDES when it purchased shares in COSIPA shows that in 1995, although 43 industries were identified by BNDES in exhibit 4 of the GOB’s April 9, 2002 submission, BNDESPAR loans were approved in only 13 industries for investments only in state-owned enterprises. Furthermore, the top five recipients of these loan approvals, including the steel industry, represent over 80 percent of the BNDESPAR loans approved in 1995. Accordingly, we find this program to be de facto specific under section 771(5A)(D) because it is limited to a group of enterprises or industries.

We disagree with respondents’ argument that there was no financial contribution associated with this BNDESPAR loan due to the intervening involvement of private agent banks. As the record evidence indicates, USIMINAS’s BNDESPAR loan has subsequently been refinanced directly through BNDES, and therefore in repaying this loan prior to and during the POI, USIMINAS makes payments to BNDESPAR and not to Bozano. Therefore, we find that this program did provide a financial contribution in the form of a direct transfer of funds in the form of loans from the GOB pursuant to section 771(5)(D)(i) of the Act.

Comment 12: Pro-Industria- Specificity

Respondent’s Argument: Respondents argue that the Pro-Industria program is not countervailable because it is not de facto specific to the steel industry within the state of Minas Gerais. Respondents note that only USIMINAS is eligible to use the Pro-Industria program because it is the only company that is located within the state of Minas Gerais.

Respondents further argue that in determining whether this program is specific, the Department must look at a breakdown of Pro-Industria usage by industry sector. Respondents argue that the steel industry, which is included in the sector identified as “basic metallurgy,” received approval for an amount that was comparably less than other industries that used this program. Respondents further argue that because the steel industry was such a small participant among many different industries, the steel industry is not the “predominant” user of the program. Additionally, respondents note that with such a broad distribution of program use among industrial sectors using the program, the recipients are not limited in number. As a result, respondents argue that the Department must find that Pro-Industria is not countervailable because it is not de facto specific to the steel industry.

Department’s Position: We disagree with respondents. From 1997 through 2000, the Pro- Industria program was used predominantly by a very small group of industries, the automobile and basic metallurgy industries. The total usage of these industries combined for each year from 1997 to 2000 add up to more than sixty percent of total program usage. Because this program was used primarily by a small group of industries, we determine that the Pro-Industria program is specific to a group of enterprises or industries within the meaning of section 751(5A)(D)(iii)(II) of the Act.

Comment 13: Non-Use of PROEX

Respondents’ Argument: Respondents note that the Department has since verified that neither the three steel producers under investigation nor the trading companies involved in the export of subject merchandise to the United States used this program.

Department’s Position: We agree with respondents. The Department verified that this program was not used. See GOB Verification Report at 14-15.

Total Ad Valorem Rate

We have revised the net subsidy rate that was calculated in the Preliminary Determination. We determine the total estimated net subsidy rate for each company to be the following:
Producer/Exporter Net Subsidy Rate
USIMINAS/COSIPA 13.99%
CSN 7.90%
All Others 13.07%

Recommendation

Based on verification and our analysis of the comments received, we recommend adopting all of the above positions. If these recommendations are accepted, we will publish the results of the final determination in the Federal Register.

__________ __________
Agree Disagree
______________________
Faryar Shirzad
Assistant Secretary
for Import Administration
______________________
Date