67 FR 6226, February 11, 2002 A-549-817 AR: 07/19/99 - 06/30/00 Public Document Enforcement III/8:PH,MF,DA MEMORANDUM TO: Faryar Shirzad Assistant Secretary for Import Administration FROM: Joseph A. Spetrini Deputy Assistant Secretary for Import Administration Group III SUBJECT: Issues and Decision Memorandum for the Final Results of the Antidumping Duty Administrative Review of the Suspension Agreement on Certain Hot-Rolled Flat-Rolled Carbon Quality Steel Products from Brazil and Termination of the Suspension Agreement Summary We have analyzed the comments and rebuttals of interested parties in the antidumping duty administrative review of the suspension agreement on certain hot-rolled flat-rolled carbon-quality steel products (HR) from Brazil (A-351-828). As a result of our analysis, we have made changes in the margin calculations. We recommend that you approve the positions we have developed in the Discussion of the Issues section of this memorandum. Below is the complete list of the issues in this investigation for which we received comments and rebuttals by parties: 1. Sales Involving Trading Companies / Agency Sale Approach 2. Adjustment to U.S. Price for Comparison to Reference Price - Commissions 3. Adjustment to U.S. Price for Comparison to Reference Price - Ocean Freight 4. Adjustment to U.S. Price for Comparison to Reference Price - U.S. Inland Freight 5. Adjustment to U.S. Price for Comparison to Reference Price - Credit Insurance 6. Violation of Suspension Agreement - Alleged Inadvertent Nature 7. Margin Calculation - Entry Basis versus Sales Item Basis 8. U.S. Commission Offset - Margin Calculation 9. U.S. Warranty - Direct versus Indirect Expense 10. U.S. Credit Expense - Credit Days 11. U.S. Credit Expense - Interest Rate 12. Freight Costs - Estimated versus Actual 13. PIS /COFINS Taxes Background We published in the Federal Register the preliminary results of review of suspension agreement on August 8, 2001. See Certain Hot-Rolled Flat-Rolled Carbon Quality Steel Products from Brazil: Preliminary Results of Antidumping Duty Administrative Review of Suspension Agreement, 66 FR 41500 (August 8, 2001) (Preliminary Results). The period of review (POR) is July 19, 1999 through June 30, 2000. We invited parties to comment on our preliminary results of review. We received a case brief from respondents, Companhia Siderurgica Nacional (CSN), Usinas Siderurgicas de Minas Gerais (USIMINAS), and Companhia Siderurgica Paulista (COSIPA), on September 7, 2001. We received a rebuttal brief from petitioners, Bethlehem Steel Corporation, LTV Steel Company Inc., National Steel Corporation, and United States Steel LLC, on September 14, 2001. A hearing was not requested. Scope of Investigation The products covered are certain hot-rolled flat-rolled carbon-quality steel products of a rectangular shape, of a width of 0.5 inch or greater, neither clad, plated, nor coated with metal and whether or not painted, varnished, or coated with plastics or other non-metallic substances, in coils (whether or not in successively superimposed layers) regardless of thickness, and in straight lengths, of a thickness less than 4.75 mm and of a width measuring at least 10 times the thickness. Universal mill plate (i.e., flat-rolled products rolled on four faces or in a closed box pass, of a width exceeding 150 mm, but not exceeding 1250 mm and of a thickness of not less than 4 mm, not in coils and without patterns in relief) of a thickness not less than 4.0 mm is not included within the scope of this agreement. Specifically included in this scope are vacuum degassed, fully stabilized (commonly referred to as interstitial-free (IF)) steels, high strength low alloy (HSLA) steels, and the substrate for motor lamination steels. IF steels are recognized as low carbon steels with micro-alloying levels of elements such as titanium and/or niobium added to stabilize carbon and nitrogen elements. HSLA steels are recognized as steels with micro- alloying levels of elements such as chromium, copper, niobium, titanium, vanadium, and molybdenum. The substrate for motor lamination steels contains micro-alloying levels of elements such as silicon and aluminum. Steel products to be included in the scope of this agreement, regardless of HTSUS definitions, are products in which: (1) iron predominates, by weight, over each of the other contained elements; (2) the carbon content is 2 percent or less, by weight; and (3) none of the elements listed below exceeds the quantity, by weight, respectively indicated: 1.80 percent of manganese, or 1.50 percent of silicon, or 1.00 percent of copper, or 0.50 percent of aluminum, or 1.25 percent of chromium, or 0.30 percent of cobalt, or 0.40 percent of lead, or 1.25 percent of nickel, or 0.30 percent of tungsten, or 0.012 percent of boron, or 0.10 percent of molybdenum, or 0.10 percent of niobium, or 0.41 percent of titanium, or 0.15 percent of vanadium, or 0.15 percent of zirconium. All products that meet the physical and chemical description provided above are within the scope of this agreement unless otherwise excluded. The following products, by way of example, are outside and/or specifically excluded from the scope of this agreement: • Alloy hot-rolled steel products in which at least one of the chemical elements exceeds those listed above (including e.g., ASTM specifications A543, A387, A514, A517, and A506). • SAE/AISI grades of series 2300 and higher. • Ball bearing steels, as defined in the HTSUS. • Tool steels, as defined in the HTSUS. • Silico-manganese (as defined in the HTSUS) or silicon electrical steel with a silicon level exceeding 1.50 percent. • ASTM specifications A710 and A736. • USS Abrasion-resistant steels (USS AR 400, USS AR 500). • Hot-rolled steel coil which meets the following chemical, physical and mechanical specifications: C Mn P S Si Cr Cu Ni ----------------------------------------------------------------------------------------------------- 0.10-0.14% 0.90% Max 0.025% Max 0.005% Max 0.30-0.50% 0.30-0.50% 0.20-0.40% 0.20% Max ----------------------------------------------------------------------------------------------------- Width = 44.80 inches maximum; Thickness = 0.063- 0.198 inches; Yield Strength = 50,000 ksi minimum; Tensile Strength = 70,000 - 88,000 psi. • Hot-rolled steel coil which meets the following chemical, physical and mechanical specifications: C Mn P S Si Cr Cu Ni Mo ------------------------------------------------------------------------------------------------------------- 0.10-0.16% 0.70-0.90% 0.025% Max 0.006% Max 0.30-0.50% 0.30-0.50% 0.25% Max 0.20% Max 0.21% Max ------------------------------------------------------------------------------------------------------------- Width = 44.80 inches maximum; Thickness = 0.350 inches maximum; Yield Strength = 80,000 ksi minimum; Tensile Strength = 105,000 psi Aim. • Hot-rolled steel coil which meets the following chemical, physical and mechanical specifications: C Mn P S Si Cr Cu Ni V(wt.) Cb --------------------------------------------------------------------------------------------------------------------------- 0.10-0.14% 1.30-1.80% 0.025% Max 0.005% Max 0.30-0.50% 0.50-0.70% 0.20-0.40% 0.20% Max 0.10% Max 0.08% Max --------------------------------------------------------------------------------------------------------------------------- Width = 44.80 inches maximum; Thickness = 0.350 inches maximum; Yield Strength = 80,000 ksi minimum; Tensile Strength = 105,000 psi Aim. • Hot-rolled steel coil which meets the following chemical, physical and mechanical specifications: C Mn P S Si Cr Cu Ni Nb Ca Al ------------------------------------------------------------------------------------------------------------------------------------------- 0.15% Max 1.40% Max 0.025% Max 0.010% Max 0.50% Max 1.00% Max 0.50% Max 0.20% Max 0.005% Min Treated 0.01 - 0.07% ------------------------------------------------------------------------------------------------------------------------------------------- Width = 39.37 inches; Thickness = 0.181 inches maximum; Yield Strength = 70,000 psi minimum for thicknesses 0.148 inches and 65,000 psi minimum for thicknesses > 0.148 inches; Tensile Strength = 80,000 psi minimum. • Hot-rolled dual phase steel, phase-hardened, primarily with a ferritic- martensitic microstructure, contains 0.9 percent up to and including 1.5 percent silicon by weight, further characterized by either (i) tensile strength between 540 N/mm2 and 640 N/mm2 and an elongation percentage 26 percent for thicknesses of 2 mm and above, or (ii) a tensile strength between 590 N/mm2 and 690 N/mm2 and an elongation percentage 25 percent for thicknesses of 2 mm and above. • Hot-rolled bearing quality steel, SAE grade 1050, in coils, with an inclusion rating of 1.0 maximum per ASTM E 45, Method A, with excellent surface quality and chemistry restrictions as follows: 0.012 percent maximum phosphorus, 0.015 percent maximum sulfur, and 0.20 percent maximum residuals including 0.15 percent maximum chromium. • Grade ASTM A570-50 hot-rolled steel sheet in coils or cut lengths, width of 74 inches (nominal, within ASTM tolerances), thickness of 11 gauge (0.119 inch nominal), mill edge and skin passed, with a minimum copper content of 0.20%. The merchandise subject to this agreement is classified in the Harmonized Tariff Schedule of the United States (HTSUS) at subheadings: 7208.10.15.00, 7208.10.30.00, 7208.10.60.00, 7208.25.30.00, 7208.25.60.00, 7208.26.00.30, 7208.26.00.60, 7208.27.00.30, 7208.27.00.60, 7208.36.00.30, 7208.36.00.60, 7208.37.00.30, 7208.37.00.60, 7208.38.00.15, 7208.38.00.30, 7208.38.00.90, 7208.39.00.15, 7208.39.00.30, 7208.39.00.90, 7208.40.60.30, 7208.40.60.60, 7208.53.00.00, 7208.54.00.00, 7208.90.00.00, 7210.70.30.00, 7210.90.90.00, 7211.14.00.30, 7211.14.00.90, 7211.19.15.00, 7211.19.20.00, 7211.19.30.00, 7211.19.45.00, 7211.19.60.00, 7211.19.75.30, 7211.19.75.60, 7211.19.75.90, 7212.40.10.00, 7212.40.50.00, 7212.50.00.00. Certain hot- rolled flat-rolled carbon-quality steel covered by this investigation, including: vacuum degassed, fully stabilized; high strength low alloy; and the substrate for motor lamination steel may also enter under the following tariff numbers: 7225.11.00.00, 7225.19.00.00, 7225.30.30.50, 7225.30.70.00, 7225.40.70.00, 7225.99.00.90, 7226.11.10.00, 7226.11.90.30, 7226.11.90.60, 7226.19.10.00, 7226.19.90.00, 7226.91.50.00, 7226.91.70.00, 7226.91.80.00, and 7226.99.00.00. Although the HTSUS subheadings are provided for convenience and Customs purposes, the written description of the merchandise under this agreement is dispositive. Changes since the Preliminary Determination For business proprietary details of our analysis of the above mentioned changes to our final margin and reference price comparison calculations, see USIMINAS/COSIPA and CSN Final Analysis Memoranda, February 4, 2002. Discussion of the Issues Comment 1: Sales Involving Trading Companies - Agency Sale Approach Respondents argue that the Department's reference price analysis was flawed. Respondents argue that the Department's decision to reject the treatment of trading company customers as U.S. agents on certain sales made by CSN and USIMINAS is inconsistent with the intention of the Suspension Agreement, as well as common agency principles. According to respondents, the principle behind the Agreement was to ensure that the price paid for Brazilian hot-rolled steel was at or above a minimum reference price representing a point of direct competition with U.S.-made steel. Respondents assert that the Department, during the suspension agreement negotiations, refused the Brazilian mills suggestion to base the U.S. price on a minimum FOB-Brazil port basis, the sale term that respondents state was historically used for their sales to the U.S. market. Respondents state that the Agreement refers to the reference price as including "all transportation charges to the U.S. port of entry, together with port fees, duties, offloading, wharfage and other charges incurred in bringing the steel to the first customs port of discharge in the U.S. market" (citing Suspension Agreement). According to respondents, the Suspension Agreement's reference price is based on the prices in Purchasing Magazine because these prices are the most equivalent point at which imports and domestic products compete head-to-head in the market. Respondents explain that the Purchasing Magazine price terms are supposed to be FOB-U.S. mill, which they state is the equivalent of FOB-U.S. port (the closest International Commercial Term (Incoterm) for which is delivered-duty-paid or landed-duty-paid (DDP)). Respondents state that complying with the Suspension Agreement was difficult for the Brazilian mills, since none of the Brazilian mills had affiliated importers in the United States. Therefore, respondents claim that using an affiliated party as a method of charging U.S. customers a landed-duty-paid price was not possible, and this left the Brazilian mills with three options: 1) sell to the trading companies at a FOB or cost-and- freight price (C&F) below the reference price, hoping that the additional expenses incurred would put the total price at or above the reference price, and hoping that the unaffiliated importers would allow the Department to verify these importers' expenses; 2) sell to the trading companies at a DDP price, at or above the reference price, which would mean that someone in the United States would have to be reimbursed for post-shipment expenses associated with "bringing the steel to the first customs port of discharge in the U.S. market" because DDP terms require the shipper to pay these expenses; or 3) have the U.S. trading companies act as agents, whereby the Brazilian mills receive a "cash advance" from the agents at a FOB or C&F price (below the reference price), and the trading companies cover the remaining expenses, ensure that the U.S. customer is charged a price at or above the reference price, and permit the Department to verify these expenses. Respondents note that while option two is used, and has worked, for COSIPA, CSN determined that this option was cumbersome and unworkable, and involved the risk of violating the Suspension Agreement because of the various expenses that could end up being overpaid by CSN. Respondents state that CSN and USIMINAS determined that the Suspension Agreement language allowed them the flexibility to sell under option three, which is the option these two companies have used for most of their sales to the United States under the Suspension Agreement. According to respondents, option three - having U.S. trading companies act as agents - allows the Department to have the rare opportunity to know the actual price at which the product is sold to the final U.S. customer. Respondents claim the sale to the final U.S. customer is the point at which the product directly competes with sales of U.S.-made steel. CSN and USIMINAS further claim that the agency option follows the spirit and letter of the Suspension Agreement (better than other options), and has additional benefits: the Brazilian mills' avoid the difficulties of administratively handling DDP sales, and the Department can monitor the price at which the imported product entered the U.S. marketplace. Respondents state that under non-agency options, CSN's and USIMINAS' trading companies could potentially sell to the final U.S. customer at a price below the reference price to cover any of their potential losses, by higher prices on other products resulting in the circumvention of the Suspension Agreement. Respondents maintain that with the agency option, the prices charged by CSN and USIMINAS comply with the reference price as it is defined in the Suspension Agreement, and the Department can effectively monitor the sales at the point of direct competition with the mills. Respondents compare the steps involved with agency sales and direct sales in order to demonstrate that the agency sales option is preferable. According to respondents, the C&F price to the U.S. trading company may be lower than the reference price, but the price to the final U.S. customer includes expenses that result in a total price that is at or above the reference price. On the other hand, according to respondents, a direct sale from the Brazilian mill to the U.S. trading company could be made at a DDP price that is at or above the reference price, but pursuant to DDP sales terms, the Brazilian mill would be required to reimburse the trading company for various import expenses, resulting in a net payment to the Brazilian mill that is below the reference price. Additionally, respondents state that the U.S. trading companies could then resell the subject merchandise at a price that is unknown to the Brazilian mill or the Department. In fact, according to respondents, under a non-agency option the U.S. trading companies could resell the product below the reference price because they are not obligated to operate under the terms of the Suspension Agreement. Respondents claim that the evidence on the record confirms that sales made by CSN and USIMINAS were legitimate agency sales. Respondents affirm that all of the agency sales made by CSN and USIMINAS were between the mill and the end customer, as demonstrated by the following: USIMINAS, for example, was aware of the customer and the price terms that were negotiated with the customer; and the customer knew which mill produced the steel and what type of product to expect. Respondents state that the U.S. trading companies represented themselves merely as agents of the mills, and the final U.S. customer was aware that the Brazilian mill controlled the terms of the sale. Additionally, according to respondents, the export license that is issued states that the trading company is "operating as agent for" the mill. Respondents state that the trading company's invoice also indicates that the trading company is "acting as a sales agent in the USA market" for the mills. Therefore, according to respondents, all of the parties had an understanding that the trading company was acting on behalf of the seller (the Brazilian mill). Respondents assert that the transfer of title during the sales process demonstrates that the end-user and not the trading company was the customer for agency sales. Citing Uniform Commercial Code section 2-401, n. 14 ("Sellers' delivery of goods to a third person at the direction of the buyer is effective to transfer title to the buyer"), respondents state that the trading companies were agents acting on behalf of the seller (the principal) and did not receive title to the merchandise because the sales contract was between the Brazilian mills and the final U.S. customer, not between the mills and the trading companies. Respondents state that the trading companies negotiated the final terms of the sales on behalf of the mills, but the mills remained the seller for the purpose of the sales contract. Respondents claim that, based on CSN's and USIMINAS' terms of sale, the title passed directly from the mill to the customer at the point the merchandise entered the United States because the invoice to the end- customer was on a landed-duty-paid basis, while the advance payment that the trading company made to the mill was on a C&F basis. In any case, respondents state that issuance of an invoice in and of itself is not the conclusive evidence of transfer of title. According to respondents, the customer's awareness of the agency relationship, which is documented in the trading company's invoice, is more relevant than the terms on the invoice. Respondents also argue that the time lag between issuance of an invoice and transfer of title is inconsequential. Respondents maintain that the more appropriate benchmark for transfer of title is when risk of loss passes to the buyer, which is determined by the sale term. Respondents give an example of CSN's direct sales during the POR, which involved the Cayman subsidiary issuing a commercial invoice to the customer at a DDP price, and argue that when delivery is required, title does not transfer until the seller's performance is completed (after the product enters U.S. Customs, in the case of DDP sales). Since the invoice may have been issued days or weeks before (at the time of shipment from the Brazilian port), respondents conclude that it is common for passage of title and issuance of the invoice to occur on different dates. Additionally, respondents state that the fact that the agent carried the risk of non-payment and attendant credit cost does not nullify the validity of the trading companies' agency relationships with the mills. Instead, respondents argue that these responsibilities are among the many responsibilities the agent agreed to undertake on behalf of the mills, and these responsibilities tend to prove, not disprove, that the trading companies acted on behalf of the mills. Respondents point out the Department's statement in the Preliminary Determination that the trading companies have the power to negotiate and set the price and terms of sales, except that the price must be above the reference price (and in the case of USIMINAS, the trading companies do not have unlimited freedom during price negotiations). Respondents argue that this statement reveals that the agreement between the mills and the agents ensured the mills' compliance with the Suspension Agreement. Respondents assert that the trading companies did not have control over the final price, which indicates that there was no sale between the mill and the trading company; rather, the trading company was operating under the restricted authority that is given under a legitimate agency agreement. Respondents maintain that the mills insisted on selling at or above the reference price, and because of the agency agreement, the trading companies complied with this requirement. Respondents assert that the trading companies acknowledged that absent the agency agreement, they would have sold inventoried product at the prevailing market price. Petitioners counter that the Department properly determined that respondents violated the reference price provisions of the Suspension Agreement. According to petitioners, respondents have no basis for asserting that "the point of direct competition" was the Suspension Agreement's intended point of comparison. Petitioners emphasize that respondent's assertion is inconsistent with the Department's analysis, which, petitioners state, focuses on the first sale to an unaffiliated purchaser in the U.S. market. Petitioners disagree with respondents' statement that, in a "real-world sense," the prices charged by CSN and USIMINAS fully comply with the Suspension Agreement's definition of the reference price. Petitioners state that the Suspension Agreement does not deal with price visibility (or transparency) issues or the type of customer to whom a sale should be measured. Petitioners stress that the Suspension Agreement instead includes the standard definition of a U.S. sale, which, to whatever type of customer and with the inclusion of certain expenses, must be above the stated reference price. Petitioners maintain that respondents sold below the stated reference price that is plainly defined in the Suspension Agreement. Petitioners argue that respondents' sales to the U.S. trading companies were not agency sales but were sales to the trading companies themselves. Petitioners state that substantial evidence, such as the contractual language between the mills and the trading companies, shows that sales were between the mills and the trading companies. Petitioners point out that USIMINAS' and CSN's invoicing patterns (e.g., USIMINAS invoices its trading companies, which then independently invoice their customers) also demonstrate that the sales were to the trading companies. Additionally, according to petitioners, the price negotiation and payment patterns show that the trading companies are the customers because the mills negotiate a price with the trading companies, and then the trading companies negotiate separately with the U.S. customer. Petitioners note that the trading company's sole economic interest derives from the separate negotiations it conducts with CSN or USIMINAS on one hand, and with its customer on the other hand. Petitioners state that the trading companies' "commissions," the difference between the price charged to the final U.S. customer and the price paid to CSN or USIMINAS (plus expenses), is unlike a commission in that there is no set formula or method for determining the difference. Petitioners assert that the trading companies actually receive a profit (or loss) from the sales. Petitioners state that the less CSN or USIMINAS benefits from the transaction, the more profit the trading companies make. Petitioners state that this is inconsistent with an agency relationship, where the agent acts on behalf of the principal. Petitioners point out that the fact that the trading companies can lose money is also inconsistent with an agency relationship because an agent is supposed to earn a flat fee or a percentage of the price, and only an entity that buys and then resells products will be willing to run the risk of a loss. Petitioners call attention to the fact that payment from the trading company to the Brazilian mill is due "upon receipt of paperwork" while the U.S. customer pays the trading company when it receives the merchandise, which results in a lengthy gap between the trading company's payment to the mills and its demand for payment from its customer. Petitioners state that the trading company bears the risk of loss in the event that the U.S. customer cancels a sale after the trading company has already paid the Brazilian mill, since the sale between the mill and the trading company is final. Petitioners challenge respondents' statement that bearing the risk of loss is one of the responsibilities that the trading companies assume on behalf of the mills. Petitioners maintain that the fact that the trading companies bear the risk of loss after paying the mills for the merchandise before reselling is an indication of ownership. According to petitioners, the trading companies are not acting "on behalf" of the mills in bearing the risk of loss, rather they bear the risk because they can profit from transactions that are not cancelled. Petitioners assert that the fact that the trading companies bear the risk of loss after paying the mills for the merchandise but before reselling it is a plain indication of ownership. Petitioners point out that there is no evidence on the record that indicates that the Brazilian mills have agency agreements with the trading companies. Petitioners argue that the mills' invoices to the trading companies show that the trading companies are customers, not agents. According to petitioners, the fact that there are no agency agreements, along with the facts stated in the previous paragraphs, refutes respondents' claim that the first sale to an unaffiliated customer is between the mill and the U.S. end-user. Because, according to petitioners, the trading companies are customers, title passes to the trading companies. Regarding respondents' citation of the Uniform Commercial Code for the proposition that the seller's delivery of goods to a third person at the direction of the buyer is effective to transfer title to the buyer, petitioners argue that there is no evidence indicating that the mills deliver the merchandise to the trading companies at the direction of the end-users. Rather, petitioners state that respondents pointed out that delivery to the trading companies is at respondents' discretion. Petitioners dispute respondents' claim that the invoice terms from the mills show that the title passes to the end-user, by-passing the trading companies. Petitioners state that the invoice from the mills to the trading companies specify shipping obligations that the mills owe to the trading companies, and the mills' performance is completed upon shipment to the trading companies. Thus, citing respondents' statement that title does not transfer until the seller's performance is completed, petitioners state that the title transfers to the trading company when the mill completed its performance (which occurred prior to payment of duties, brokerage, etc.). Petitioners add that respondents rely on the invoicing terms standard but then negate it by stating that regardless of the invoice terms, the end-user is aware of the agency relationship. Petitioners explain that when sales from the mill to the trading company are made at C&F terms, the mill is responsible for freight to the U.S. port, but the trading company covers duties and brokerage. Regarding the terms of sale between the trading company and the end-user, petitioners state that the trading company is responsible for expenses incurred after shipment, including delivery to the end-user. Therefore, according to petitioners, title does not pass to the end-user until it receives the merchandise in the United States. Petitioners maintain that title passes from the mill to the trading company, and then from the trading company to the end-user. Regarding respondents' comment that the more appropriate benchmark for transfer of title is the passing of risk of loss, petitioners state that the trading company bears the risk of loss, especially if an end-user cancels a sale after the trading company has paid the mill for the merchandise. Petitioners conclude that this risk of loss is consistent with transfer of title from the mill to the trading company rather than an agency relationship between the mill and the trading company. Petitioners argue that restrictions placed on the trading companies in negotiating price with the end-user is not indicative of agency relationships between the mills and the trading companies. Petitioners state that while respondents use the existence of restrictions as a way of substantiating that there is an agency agreement, the fact remains that there is no evidence of an agency agreement. Petitioners conclude that respondents' claim that an agreement restricted the trading company's authority is unsubstantiated. Petitioners restate that the record evidence demonstrates that the first sale to an unaffiliated party in the United States was the sale from the Brazilian mill to the trading company. Petitioners state that the Department appropriately treated these sales as export price (EP) sales in the Preliminary Results and should continue to do so in the Final Results. Petitioners add, however, that, if the Department determines that the trading companies acted as agents, and the end-user was the first unaffiliated customer, then the statute requires the Department to treat these sales as constructed export price (CEP) sales. Furthermore, according to petitioners, because neither CSN nor USIMINAS reported indirect selling expenses in their databases, the Department is required to use facts available for those expenses. Department's Position: The statute directs the Department to base U.S. price upon the first sale to an unaffiliated party in or to the United States. See sections 772(a) and 772(b) of the Act; see also section I. of the Suspension Agreement (defining U.S. price). The issue in this case is whether respondents CSN and USIMINAS are affiliated with the trading companies in question for purposes of determining U.S. price. Section 771(33) of the Act defines the term "affiliated persons." The relevant subsection, 771(33)(G), states that an affiliated person is "{a}ny person who controls any other person and such other person." Section 771(33) also states that "a person shall be considered to control another person if the person is legally or operationally in a position to exercise restraint or direction over the other person." Accordingly, to determine the appropriate price, we must establish whether in fact USIMINAS, and separately CSN, was in a position to exercise restraint or direction over the trading company in question for the sales at issue. In this case, we find that the respondents CSN and USIMINAS are not affiliated with the trading companies for the sales at issue, as discussed below. To determine whether a seller is acting as an agent under the control of the foreign manufacturer, and therefore, is affiliated with the manufacturer, the Department has previously focused on a range of criteria including, (1) the foreign producer's role in negotiating price and other terms of sale; (2) the extent of the foreign producer's interaction with the U.S. customer; (3) whether the agent/reseller maintains inventory; (4) whether the agent/reseller takes title to the merchandise and bears the risk of loss; and (5) whether the agent/reseller further processes or otherwise adds value to the merchandise. See, e.g., Final Determination of Sales at Less Than Fair Value; Stainless Steel Sheet and Strip in Coils from Germany, 62 FR 24394, 24403 (May 5, 1997) (citing Furfuryl Alcohol from South Africa, 60 FR 22550 (May 8, 1995) (Furfuryl Alcohol); and Electrolytic Manganese Dioxide from Japan; Final Results of Antidumping Duty Administrative Review, 58 FR 285510 (May 14, 1995) (EMD from Japan). Although the test applied in Furfuryl Alcohol focused upon the role of the foreign producer in the sales transactions, the characteristics of the transactions themselves and other factors have also formed part of our analysis. In EMD from Japan, for example, the petitioner argued that the reseller, Mitsubishi, was acting as the agent of the manufacturer, Tosoh Corporation. We disagreed, finding instead that Mitsubishi was an independent reseller. We based our decision upon the finding that Mitsubishi took title to the merchandise, bore the risk of loss, and was fully responsible for completion of the sale through its U.S. subsidiary. In addition to the factors considered in the above cases, other factors typically examined to distinguish an independent seller from a manufacturer's agent include whether (1) the reseller can fix the price at which it sells without accounting to the manufacturer for the difference between that price and the price paid to the manufacturer; (2) the reseller deals, or has the right to deal, in the goods of other suppliers; and (3) the reseller deals in its own name and does not disclose the supplier. The analysis of whether a relationship constitutes one of principal-agent is a fact-driven, case-specific determination; there is no bright line test. For example, although agency relationships are frequently established by a written contract, this is not essential for an agency relationship to exist. Under general principles of agency, the focus of the analysis is whether it is agreed that the agent is to act for the benefit of the principal, not for itself. (See Restatement on Agency, sections 1 cmt.b. and 25 cmt.a. See also section 14J and 14K.). In addition, control of the principal over its agent is the hallmark of an agency relationship. (See Restatement, section 14). More specifically, section 771(33) of the Act focuses the Department's inquiry upon the party's ability to exercise control, rather than the actuality of control over specific decisions. See Antidumping Duties; Countervailing Duties; Final Rule, 62 FR 27295, 27298 (May 27, 1997) (Final Rule); see also Antidumping Duties; Countervailing Duties; Proposed Rule, 61 FR 7307, 7311 (Feb. 27, 1996). Finally, to make an appropriate determination, we must examine the totality of the circumstances to understand the relationship of the parties, and to determine whether the foreign producer controls the trading company with respect to the sales at issue. In this case, respondents CSN and USIMINAS claim that the trading companies are their agents because the trading companies hold themselves out as agents, as demonstrated by the export license and invoice documentation, because the end-user is aware of the mill producing the merchandise and that the mill is aware of the identity of the end-user, and because the trading companies are making the sales to end-users on behalf of the respondents. We disagree. The fact that both the export license and invoice issued to the end-user state that the trading company is the agent of respondent does not by itself create an agency relationship. The overall arrangement, and the language of the contracts between the parties, indicate otherwise. The sales contracts indicate that the sales occurred between the trading companies and the end-users, and not between the respondents and end- users. The Brazilian mills invoiced the trading companies and received payments from the trading companies (upon receipt of paperwork) long before the trading companies sent invoices to and received payment from their U.S. customers. Second, in this case the trading companies take title to the merchandise, bear the risk of loss, and are fully responsible for completion of the sale. In addition to negotiating prices with their U.S. customers, the trading companies handled delivery to the customers, provided follow-up services, and handled complaints or returns, if there were problems with the merchandise. See USIMINAS/COSIPA analysis memo dated February 4, 2002. In this case, the trading companies can lose money if their customers cancel their orders. For example, in this case we found one instance in which an end-user canceled a sale after shipment from the respondent in Brazil. The trading company held the merchandise in inventory until another customer could be found (see trading company verification report dated May 11, 2001, at 6). In this case, we find that the trading companies bore the risk of loss and were responsible for completion of the sale. The trading companies interacted directly with the U.S. customers and the U.S. customers looked to the trading companies for quality, delivery, and price issues. The trading companies were responsible for marketing the product to the U.S. customers, and invoiced and received payments from the customers, based on prices that the trading companies negotiated separately with the Brazilian mills. The trading companies were not bound by any agreements and operated in order to make a profit. We agree with petitioner that under respondents' own standard for the passage of title, the trading companies take title to the subject merchandise. The invoices state the shipping obligations that the respondent owe to the trading companies. We agree the mills' performance is complete upon shipment of the subject merchandise to the trading companies. Thus, the terms of the invoice indicate that title does not bypass the trading companies as respondents suggest. Furthermore, it is logical to assume, as respondents suggest, that title transfers when a party assumes the risk of loss. As discussed above, the risk of loss is borne by the trading companies. Moreover, we agree with petitioners that the way in which these arrangements are structured is contrary to an agency relationship. The trading companies make money based on the difference between the price charged to the final U.S. customer and the price paid to the respondent. The structure of the arrangement is unlike a commission in that there is no formula or method for determining the amount for commission. As petitioners aptly point out, the less the respondents benefit from the transaction, the more profit can be made by the trading companies. Such an arrangement is contrary to the agency relationship where the agent acts in the interest of, and to benefit of, the principal. Moreover, the fact that the trading companies can lose money in buying and selling the merchandise further indicates that the trading companies in this case are not acting as agents, who would normally earn a fee or a percentage of the final selling price. An entity that buys and sells merchandise and incurs the risk of loss is acting more in the capacity of an independent seller, rather than as an agent for respondent. Third, the trading companies at issue deal, or have a right to deal, in goods of other suppliers. The trading companies routinely perform the role of trading companies for other suppliers. In fact, during the LTFV investigation in this case, respondents claimed that these same trading companies were unaffiliated customers, and therefore the respondents reported their sales to such trading companies as the relevant sales - i.e., the sales to the first unaffiliated parties - in that segment of the proceeding. See Notice of Final Determination of Sales at Less Than Fair Value; Hot-Rolled Flat-Rolled Carbon-Quality Steel Products From Brazil, 64 FR 38756, 38759 (July 19, 1999); see also Notice of Preliminary Determination of Sales at Less Than Fair Value: Hot-Rolled Flat-Rolled Carbon-Quality Steel Products from Brazil, 64 FR 8299, 8303 (Feb. 19, 1999). During the instant review, however, the respondents assert that the trading companies are their agents for the sales subject to review. In our final regulations, we specifically noted the relevance of temporal differences in these relationships, stating that "the Department will consider the temporal aspect of a relationship as one factor to consider in determining whether control exists. In this regard, we do not intend to ignore a control relationship that happens to terminate at the beginning (or comes into existence at the end) of a period of investigation or review." Final Rule, 62 FR at 27298. In this case, the trading companies at issue normally perform the role of independent sellers, and not that of agents for respondents. Here, the respondents have attempted to establish an agency relationship solely for the purpose of obtaining a more favorable position with respect to the operation of the Suspension Agreement. Finally, as noted above, control of the principal over its agent is the hallmark of an agency relationship. More specifically, section 771(33) of the Act focuses the Department's inquiry upon a party's ability to exercise control, rather than the actuality of control over specific decisions. In this case, the respondents have not demonstrated any ability to control the trading companies in the sales to end-users. Instead of the agents acting for the benefit of the principal, the trading companies are entrusted to sell above the reference price. Respondents claim that the agency agreement requires them to do so. Although not by itself dispositive of the issue of control, we find that in the case of USIMINAS and CSN, there is no written agency agreement with the trading companies that would provide the basis for determining the relationship between the parties, nor is there any document that would indicate the measure of control the manufacturer holds with respect to the trading companies for the sales in question. Significantly, the respondents acknowledge that the trading companies are not obligated to operate under the terms of the Suspension Agreement. Rather, USIMINAS and CSN entrusted the trading companies to sell above the reference price (see USIMINAS verification report dated July 31, 2001 page 9 and trading company verification report dated May 11, 2001 page 5). But the Department found no evidence that the trading companies are bound by an agency agreement or contract with USIMINAS or CSN to sell at or above the reference price. In our view, USIMINAS and CSN do not have control over the price charged by the trading companies to the end-users. While the trading companies may have sold the merchandise at or above the reference price during the period of this review, that action does not demonstrate control by the respondents. To the contrary, selling above the reference price is consistent with the trading companies' interests in selling at the highest possible price for purposes of maximizing profit. We find nothing on the record to indicate that respondents have the ability to prevent the trading companies from selling subject merchandise at prices below the reference price. Absent an ability to control the actions of the trading companies with respect to the price term, we find that the respondents have not created an agency relationship with the trading companies. Based upon the above analysis, the Department determines that the trading companies acted as independent sellers outside the control of respondents with respect to the sales at issue, and therefore constitute unaffiliated U.S. customers for purposes of determining U.S. price. Comment 2: Adjustments to U.S. Price for Comparison to Reference Price - Commissions Respondents argue that U.S. commissions should not have been excluded from the price that was compared to the reference price. Respondents note that the Suspension Agreement intended to include "other charges incurred in bringing the steel to the first customs port of discharge in the U.S. market" in the reference price. Respondents argue that U.S. commissions earned on agency sales were a necessary and direct expense related to bringing the steel to the first customs port of discharge. Respondents assert that in applying the Department's methodology in past cases, including the original investigation of the instant case, the compensation the trading companies receive is by nature a commission. Respondents note that in the investigation, the Department found that "generally speaking a commission is a payment to a sales representative for engaging in sales activity." Respondents state that the Department also recognized in the investigation that commissions paid to trading companies have certain characteristics: (1) they are agreed upon in writing; (2) they are earned directly on sales made, based on flat rates or percentage rates applied to the value of individual orders; (3) they take into consideration the expenses which a trading company incurs; and (4) they take into consideration the sales and marketing services performed by a trading company in lieu of an exporter/manufacturer establishing its own larger sales force. See Notice of Final Determination of Sales at Less than Fair Value: Certain Hot-Rolled Flat Rolled Carbon Steel Products from Brazil, 64 FR 38756, 38770 (July 19, 1999) (Final Determination). Respondents assert that the mills' arrangement with the trading companies, including agent's compensation, clearly meets the first, third, and fourth criteria. Regarding the first criterion, respondents argue that the agency relationship was in writing, citing to export licenses, sales contracts, and invoices on the record of this review and which state that the trading companies acted on behalf of the Brazilian mills (i.e., CSN or USIMINAS). Respondents assert that it was established that the trading companies' obligations as agents included making sure the mills' U.S. sales were in accordance with reference price and undertaking sales activities to bring the product to the market (citing to page 15 of the May 9, 2001, CSN Verification Report and page 9 of the July 31, 2001, USIMINAS Verification Report). Respondents further note that the trading companies confirmed this understanding of the agency relationship and the trading companies' obligations (citing to page 5 of the May 11, 2001, Trading Company Verification Report and page 3 of the May 9, 2001, Trading Company Verification Report). Regarding the second criterion, respondents argue that the mills and the trading companies agreed that the trading companies would be compensated in the form of a commission for undertaking these responsibilities (citing to May 11, 2001, Trading Company Verification Report). According to respondents, the commission was determined as the difference between the trading companies' price to the U.S. end-user and the expenses incurred in bringing the merchandise to the U.S., including the price paid to the mill. The respondents note that although the commission is not based on a flat rate or percentage, the companies agreed to the basis for the commission and the formula was applied consistently throughout the period of review. To satisfy the third and fourth criteria, the respondents point out that the trading companies were compensated for typical services that trading companies provide to import products - arranging for brokerage, discharging, and inland freight to the port of entry. The respondents argue that the mills would have incurred these expenses regardless of the trading companies' involvement in these U.S. sales. The mills conclude that they chose to use the trading companies in lieu of incurring these expenses directly. Petitioners argue that there are no trading company commissions to include in U.S. price. Petitioners assert that respondents concede that a crucial characteristic of a commission is that it is based on flat rates or percentage rates applied to the value of individual orders. Petitioners contend that the respondents do not argue that the trading companies earned commissions based on flat rates or percentage rates, but rather, the respondents conceded by silence that the trading companies earned profits rather than commissions. Petitioners note that the trading company earns the difference between the price it charges its customer and the sum of what it pays CSN or USIMINAS (plus expenses), but as the Department found in its Preliminary Analysis Memoranda for CSN and USIMINAS, unlike a commission, there is no set formula or method for determining this difference. Citing respondents' argument that the arrangement in which profit earned equals revenue less cost of goods plus expenses, is a formula applied consistently to determine the trading company commission amounts, petitioners argue that the mills cannot be found to have "paid" this amount (commission) to the trading companies, since a commission was never due or payable to the mill in the first place. Petitioners contend that this "commission" is not an amount in which the mill has any right, and therefore, the trading company cannot receive this "commission" from the mill. Petitioners go on to suggest that under respondents' theory, all entities that sell goods they do not produce earn commissions rather than profits. Department's Position: We agree with petitioners that there were no commissions paid by USIMINAS and CSN to the trading companies to include in the U.S. price for comparison to the reference price. The "commission" described by respondents is the trading companies' profit and therefore does not meet the four characteristics for a commission the Department outlined in the investigation (see Final Determination). At the Department's verifications of the mills in Brazil and of the trading companies in the United States, neither the mills nor the U.S. trading companies provided documents demonstrating that the "commission" was agreed upon in writing, in order to meet the first criterion. While respondents cite various documents involved in the sales process between the mills and the trading companies as evidence of an agency relationship between the mills and the U.S. trading companies, these documents do not reference any commission on U.S. sales. Whereas respondents assert that "the formula was applied consistently throughout the period of review," the Department notes that the alleged commission was not a flat rate or percentage applied to the value of individual orders. Therefore, the Department classifies the "commission" as profit earned by the trading companies for sales from USIMINAS and CSN. The Department confirmed at the trading companies' verifications that they arrange for brokerage, discharging, and inland freight to the port of entry. Additionally, we found that the trading companies provide sales and marketing services to U.S. end-users for merchandise the trading companies purchase from USIMINAS and CSN. Respondents view these sales and marketing services as evidence of an agency relationship and reason for claiming a "commission" on U.S. sales, thus fulfilling the third and fourth criteria defining a commission, outlined in the investigation. The Department, on the other hand, finds that the trading companies perform these services regardless of an agency agreement between the mills and the U.S. trading companies. The trading companies must carry out sales and marketing services in order to re-sell merchandise in the United States purchased from the Brazilian mills. (See Comment 1.) The "commissions" as defined by respondents do not meet the commission criteria set forth by the Department in the original investigation (see Final Determination). The Department finds that the "commissions" were not agreed upon in writing, nor are they a flat rate or percentage. Instead, the "commissions" more closely resemble a profit earned by the trading company through re-sale of merchandise purchased from the Brazilian mills. Consequently, we conclude that the trading companies' sales and marketing services are services essential to re-selling merchandise in the United States. Therefore, the Department is continuing not to make an adjustment for this amount in the calculation of the U.S. price for comparison to the reference price. Comment 3: Adjustments to U.S. Price for Comparison to Reference Price - Ocean Freight Respondents point out that the signatories to the Suspension Agreement agreed that subject merchandise will be sold at or above the established reference price. Respondents state that for agency sales, CSN Cayman and USIMINAS invoiced the trading company at FOB or C&F unit prices, which are sales terms that do not include all of the transportation and other expenses contemplated by the Agreement. Respondents explain that in order to determine whether U.S. prices for agency sales were in accordance with the applicable reference price, the Department started with CSN's Cayman or USIMINAS' price and added certain sales expenses to calculate a DDP- equivalent price. The respondents claim that the Department failed to add unit costs for pre-entry inland freight and credit insurance. Respondents state that for proper comparison to the applicable reference prices, these costs must be added to the U.S. price because they are either transportation charges to the U.S. port of entry or charges incurred in bringing the steel to the U.S. market. Respondents argue that these expenses are legitimately included in the price to the U.S. customer in accordance with the Agreement. Respondents offer that, as an alternative, the Department could compare the gross unit price reported in the GRSUPRU field, which includes these expenses, to the applicable reference price. Respondents state that the Department should add ocean freight as a transportation charge to the first port of entry to certain CSN sales. Respondents state that CSN's questionnaire responses indicate that CSN Cayman invoiced agency sales on either a FOB or C&F basis. Furthermore, respondents argue that in order to build-up to a DDP-equivalent price, ocean freight should be added to Cayman's FOB prices, but not Cayman's C&F prices, which already included ocean freight. Respondents stated that in the preliminary results memorandum for CSN, the Department incorrectly assumed that all of CSN's agency sales were invoiced by CSN Cayman on C&F terms. Respondents assert that the omission of ocean freight was probably inadvertent and that CSN had originally reported the affected sales as FOB. However, according to respondents, CSN revised the sales terms to DDP at the request of the Department to clearly indicate the terms of the invoice reported in the gross unit price (GRSUPRU) field. Respondents suggest programming language to correct the error. Petitioners did not comment on this issue. Department's Position: We agree with respondents that pre-entry ocean freight should be added to CSN Cayman's FOB prices. For certain sales, CSN originally reported the sales as FOB, and subsequently recoded the sales as DDP. For the final results, since these sales were originally on a FOB basis by CSN Cayman and not C&F like the rest of the sales, we have added ocean freight to the U.S. price for the affected sales for comparison to the reference price. However, because the sales are on a FOB basis, for purposes of the margin program, we have not subtracted ocean freight for these sales since it is not part of the gross unit price. Comment 4: Adjustments to U.S. Price for Comparison to Reference Price - U.S. Inland Freight The respondents state that the Department should adjust the price for agency sales to account for pre-entry U.S. inland freight, citing to Section H, (Art. I.H) of the Suspension Agreement. Respondents claim that the Suspension Agreement is unambiguous about the inclusion of such expenses in the DDP price or its equivalent. Respondents point out that the Agreement contemplates inclusion of pre-entry inland freight in the reference price, as a transportation charge to the U.S. port of entry. For some agency sales, respondents state that CSN and USIMINAS shipped the product to one U.S. port, but the product was not cleared through Customs until the U.S. agent arranged for transportation by barge to another port for entry. Respondents assert that the customer would not take delivery at the first port but wanted delivery closer to its operations. Also, respondents point out that given this delivery requirement, the importer delayed entry and therefore cash disbursement to Customs. By delaying entry, the respondents explain that this would allow the importer more time to gather necessary documentation to make the entry. The respondents claim that the Department determined that the Suspension Agreement (and therefore the reference prices established in the Agreement) (1) "eliminate completely the injurious effect of exports to the United States" and (2) "prevent the suppression or undercutting of price levels of domestic hot-rolled steel by imports." Respondents further state that in order to test whether CSN's and USIMINAS' sales to the United States meet these objectives, the Department must compare import prices at a level that competes with U.S. shipments. Respondents argue that imported material cannot compete with U.S. material if it is not formally entered for consumption in the United States because it is not legitimately in the marketplace. Specifically, respondents argue that because the costs of importation continue to accrue until the product is entered, which in turn increases the cost to the customer, those costs must be added to the FOB or C&F price in order to make a proper comparison. Petitioners agree that inland freight incurred before the point of entry into the U.S. Customs territory should be included in the calculation of the net U.S. price for comparison to the reference price. Petitioners contend that this pre-entry U.S. inland freight should also be included as a U.S. movement expense in the Department's calculation of net price in the Department's margin calculation program. Petitioners state that adjustments made to EP should be reduced by expenses "which are incident to bringing the subject merchandise from the original place of shipment in the exporting country to the place of delivery in the United States" (citing to section 772(c)(2)(A) of the Act). Therefore, petitioners claim that inland freight incurred prior to entry into the U.S. customs territory falls into this definition of movement expenses, and the Department should include inland freight in the U.S. movement expenses and add it to the reference price and deduct it in the dumping margin calculation. Department's Position: The Department agrees with both respondents and petitioners that the Suspension Agreement contemplates inclusion of pre-entry inland freight in the reference price, as a transportation charge to the U.S. port of entry. However, even if the inland freight charges were added for comparison to reference price, the companies would still not be in compliance with the reference price. These freight expenses are not deducted when calculating net price in the margin calculation because they are paid for by the trading companies and are not part of the transaction between the mills and the trading companies. Comment 5: Adjustments to U.S. Price for Comparison to Reference Price - Credit Insurance Respondents argue that credit insurance is another appropriate charge incurred in bringing the product to the United States. Respondents claim that as CSN's and USIMINAS' agents, the trading companies performed a variety of services to sell CSN's and USIMINAS' product in the United States. In addition, respondents point out that the trading companies incurred the risk of non-payment by the U.S. customer because the trading companies paid upon invoicing but did not receive payment from the U.S. customer until after delivery of the goods. Respondents state that the trading companies acquired credit insurance to cover their credit expense (i.e., insurance that covers the risk of non-payment) between the date it pays the mill and the date it receives payment from the customer. The respondents claim that in order to determine whether prices for agency sales were in accordance with the applicable reference prices, the Department should add credit insurance (and all other reported sales expenses that the trading companies incurred on the mills' behalf) to the price the trading companies paid. Petitioners did not comment on this issue. Department's Position: As discussed in comment 1 above, the Department has determined that the sales to the first unaffiliated party in the United States are the sales to the trading companies. Credit insurance is paid for by the trading companies, and it is related to the risk of non-payment by their customers. Therefore, it is not related to the sales between the mills and the trading companies. Consequently, the Department has not adjusted the U.S. price to include credit insurance. Comment 6: Violation of Suspension Agreement - Alleged Inadvertent Nature Respondents assert that if the Department ultimately determines that USIMINAS' and CSN's sales were not agency sales, and that these sales were made at prices below the reference price, the Department must then conclude that any "violation" that may have been committed was inadvertent, and that there are no sufficient grounds for terminating the Suspension Agreement. According to respondents, any violations that arise will be the result of the Department's interpretation of what is a true "sale" (i.e., if the Department determines that the true sale was from the mill to the trading company). Respondents allege that the Suspension Agreement does not clearly define what is a sale, and the Department did not state its interpretation of a "sale" until after sales had been made. Respondents state that the Department knew of the Brazilian mills' intention to use U.S. trading companies as agents, since, according to respondents, the mills discussed this procedure with the Department before any sales were made by any mill. Respondents argue that the Department never issued any oral or written acceptance or disapproval of agency sales under the Suspension Agreement. Therefore, respondents claim, that without any guidance, the mills had to act according to their best interpretation of the Suspension Agreement. Respondents declare that the mills made their best efforts to comply with the Suspension Agreement, never hid anything from the Department pertaining to the agency sales, and believe that their sales processes comply with the terms of the Suspension Agreement. Respondents maintain that the mills acted in good faith and could not have known that the Department would interpret the term "sale" in a different manner. Therefore, according to respondents, the Department cannot conclude that any of the mills intentionally violated the Suspension Agreement. Respondents note that Article XI.B. of the Suspension Agreement permits the Department to take whatever action it deems appropriate, presumably including termination of the Suspension Agreement, if it determines there has been a violation. Respondents point to the definition of a "violation" in section I.K. of the Suspension Agreement, which states that violation means noncompliance, except for noncompliance that is inconsequential, inadvertent, or does not substantially frustrate the purposes of the Suspension Agreement. Respondents maintain that the use of the word "or" in the definition means that if the noncompliance meets any of the criteria (e.g., inconsequential) then it is not a violation. Respondents reiterate that USIMINAS and CSN believed that they were complying with both the letter and the spirit of the Suspension Agreement, and the Department indicated that their interpretation was incorrect only in the Preliminary Determination, almost one full year after the sales under review were made. Respondents assert that the prices to the "true" customers were above the reference price. Regarding the not "inconsequential" issue, respondents claim that a large number of USIMINAS' and CSN's sales were found to be under the reference price because these companies made mostly agency sales. Respondents argue that, under these circumstances, any noncompliance with the Suspension Agreement was inadvertent and did not substantially frustrate the purposes of the Suspension Agreement, and, therefore, there is no basis for terminating the Suspension Agreement. Petitioners counter that respondents made willful violations, and that the Suspension Agreement should be terminated and an antidumping order issued, in accordance with Article XI.B. of the Suspension Agreement. Petitioners state that nothing on the record supports respondents' claim that its actions were inadvertent. Petitioners argue that while "inadvertent" is not defined in the Suspension Agreement, common usage defines it as not duly attentive or accidental. According to petitioners, the respondents' "acts" were the antithesis of inadvertent. Petitioners further assert that respondents' acts were part of a pattern of calculated and systematic conduct that rises beyond the Suspension Agreement's threshold for finding that the Agreement has been violated. In response to respondents' argument regarding the use of the word "or" in the provision listing the three exceptions to the violation provision of the Suspension Agreement as meaning that if noncompliance meets any of the criteria then there is not a violation, petitioners state that during the debate on the Trade Agreements Act of 1979, Congress interpreted the disjunctive "or" in a different manner that contradicts respondents' interpretation (citing to Cong. Rec. § 10311 (July 23, 1979)). Petitioners state that if one of these criteria is not met (i.e., if the violation is not inconsequential, immaterial, or unimportant), then material injury has resulted. Petitioners assert that respondents deliberately fashioned the transactions which the Department found violated the reference price provision of the Suspension Agreement. Additionally, petitioners assert that by violating the reference price provision, the cornerstone of the Suspension Agreement, respondents substantially frustrated the purposes of the Suspension Agreement. Petitioners declare that respondents violated the Suspension Agreement because its actions were neither inadvertent nor inconsequential, and they substantially frustrated the purposes of the Suspension Agreement. Therefore, according to petitioners, the Suspension Agreement should be terminated and an antidumping order should be issued. Department's Position: Section I.K. of the Suspension Agreement defines a violation as "noncompliance with the terms of this Agreement, whether through an act or omission, except for noncompliance that is inconsequential, inadvertent, or does not substantially frustrate the purposes of this Agreement." The Department found that a substantial amount of sales were made below the reference price, resulting in noncompliance with the terms of the Suspension Agreement. As stated in section I.H. of the Suspension Agreement, ". . . the Signatories will ensure that the actual terms are equivalent to a price that is not lower than the Reference Price." Respondents state that the mills acted according to their interpretation of the Suspension Agreement and that any violations that have occurred were inadvertent. Inadvertent is defined as accidental or unintentional, and as petitioners point out, CSN and USIMINAS chose to alter their selling methods from the direct sales approach that was used at the outset of the period of review. Therefore, their actions were not inadvertent. Respondents notified the Department of their intentions to sell to a customer via a U.S. company, but the Department had no way of knowing whether the Suspension Agreement was being violated until all of the information was collected and verified. The large number of violations of the reference price, in addition to the mills' violation of the 15% limit of the weighted average amount by which the estimated normal value may exceed the export price for all LTFV entries, cannot be classified as "inconsequential" and they have frustrated the purposes of the Suspension Agreement. The Suspension Agreement was established to ensure that subject merchandise was sold at or above the established reference price. Respondents' acts have resulted in noncompliance that is not inadvertent, is not inconsequential, and does frustrate the purpose of the Suspension Agreement. Comment 7: Margin Calculation - Entry Basis versus Sales Item Basis The respondents state that the Department's analysis of sales is contrary to the requirements of the statute and of the Agreement. The respondents argue that the statute (citing to section 734 (c)(1)(b) of the Act) and the Suspension Agreement require the Department to conduct the dumping calculation on an entry basis and not on a sales item or observation basis. Respondents cite the U.S. Customs' definition of entry from the Customs Service's regulations, section 141.0a(a). Respondents state that an entry comprises all of the merchandise listed in the documents of entry for a particular day. The respondents state that the entry is given a particular number, and may include more than one product, sale, and even customer. The respondents argue that the Department ignored U.S. Customs language and calculated margins on a sales-item by sales-item basis. The respondents indicate that the Suspension Agreement mirrors the statutory provision with the Agreement stating that each Signatory agreed to the dumping margin limitation on an entry-specific basis. The respondents state that the Department should revise its calculation in the final results to comply with the statute and the Agreement. Respondents state that the Department can determine which observations represent an entry from the data reported by the mills. Respondents state that for CSN several observations with the same entry date and entered value constitute one entry. The respondents state that for USIMINAS, the Department reviewed documentation at U.S. verifications demonstrating that all merchandise arriving at a port (DESTU) and on a given vessel (VESSELU) was entered together on a single U.S. Customs' entry form. Respondents additionally state that the identification of an entry must also take into account the bill of lading date (SHIPDATU) for the vessels that sailed more than once during the period of review. Respondents point out that the Department must consider the importer (IMPORTER) because each importer entered the merchandise separately. Respondents state that the Department reviewed U.S. Customs entry summaries at the March 2001 COSIPA verification. The respondents state that, for COSIPA, the entry summaries reviewed by the Department indicated that all merchandise with the same invoice prefix destined for a particular port was entered together as a single entry. Respondents argue that the verification documentation demonstrates that COSIPA's entries are INVOICU and DESTU-specific. The respondents proposed specific SAS programming to group each respondents' sales observations into specific entries. Petitioners argue that the Department properly calculated dumping margins on an entry basis. Petitioners state that respondents' argument appears to be attacking the Department's methodology for calculating dumping margins in all antidumping administrative reviews. Petitioners point out that section 751 (a)(2)(A) of the Act, regarding administrative reviews, holds that the administering authority shall determine 1) the normal value and export price (or constructed export price) of each entry of the subject merchandise, and 2) the dumping margin for each such entry. Petitioners note that the statute uses identical terminology (i.e., "entry") in various sections as in the Suspension Agreement. Petitioners state that they are unaware of an administrative review where multiple products, sales, and invoices were combined together prior to product comparison for purposes of calculating a dumping margin. Petitioners also state that respondents provide no case precedent for the alternative methodology they are proposing. Petitioners argue that there are discrepancies between the respondents' narrative description of the proposed change to the calculation of the margins on an "entry" basis and the actual programming. Petitioners assert that calculation on an entry-by-entry basis would not allow for model- specific matching in order to make an appropriate comparison of the U.S. and foreign market price. Petitioners conclude that the Department should continue with the calculation methodology used in the Preliminary Results. Department's Position: We agree with petitioners. Consistent with our practice, we have calculated the dumping margins in this case on an entry basis. For purposes of section 734(c)(1)(B) of the Act, we have interpreted and applied the term "entry," as defined in section 771(23), consistent with our interpretation and application under section 751(a)(2)(A) for calculating dumping margins in administrative reviews. Further, the respondents do not cite, nor are we aware of, any case in which the Department has interpreted and applied the term as respondents suggest. To do so would require that we combine multiple products for product comparisons, and thereby defeat the Department's comparison to the foreign like product established by the Department's model-match, pursuant to section 771(16 ) of the Act. Such an approach would also distort the Department's dumping margin calculations by inflating or deflating margins, depending upon the mix of products that were combined. Finally, we note that this issue is not relevant to our decision in this case because, regardless of whether we calculate dumping margins based upon sales specific observations, or upon what the respondents have termed "entries" in this case, we find that the Suspension Agreement has been violated. Accordingly, we have continued to calculate the dumping margins in this case as in our Preliminary Results. Comment 8: U.S. Commission Offset - Margin Calculation Respondents argue that the statute and the Department's regulations require an offset for U.S. commissions in the dumping analysis. Respondents point out that section 773(a)(6)(c)(iii) of the Act states that in order to achieve a fair comparison with the export price, normal value shall be adjusted to account for differences in circumstances of sales between the two markets. Respondents also cite to section 351.410(e) of the Department's regulations. Respondents state that the Department should allow a commission offset for all U.S. commissions on CSN's and USIMINAS' agency sales (if the Department were to accept the agency relationship between the Brazilian mills and the U.S. trading companies) because no commissions were paid in the home market. The respondents state that CSN paid commissions to its agent on certain direct sales and all of the sales that CSN classifies as agency sales. Respondents state that USIMINAS paid commissions on all of its U.S. sales, which USIMINAS classifies as agency sales. The respondents conclude that the mill compensated the agents for services they performed to bring the product to the U.S. market, while no such commissions were paid for sales in the home market. Therefore, respondents argue the Department should follow the statute and Regulations to account for the difference in the circumstance of sales and calculate a U.S. commission offset to normal value. Petitioners argue that the Department did not allow an offset for "commissions" in the dumping analysis because CSN and USIMINAS did not pay commissions to the trading companies, and the trading companies did not earn commissions. Petitioners classify what respondents call commissions as profit, and the petitioners therefore conclude that the Department should continue to deny the commission offset to CSN and USIMINAS in the Final Results. Department's Position: We agree with petitioners. The Department maintains its position that USIMINAS' and CSN's sales to the first unaffiliated party in the United States were to the trading companies. We find that what respondents term commissions to the trading companies is actually profit. A commission is a flat rate or percentage applied to the value of individual orders. The trading companies did not apply this formula in determining the return on these sales. Instead, the trading companies' applied a formula to determine profit or loss: The difference between the price the trading company bought the hot-rolled steel from the mills in Brazil and the price the U.S. trading company charged the U.S. customer. Respondents state that it is possible for the trading companies to incur a "negative commission." Negative commissions are not possible by applying a flat rate or percentage to the trading companies' U.S. sales. In such an instance, the trading company would incur a loss. See Comment 2. Thus, USIMINAS and CSN did not pay any commissions to the U.S. trading companies for all of their U.S. sales. Therefore, we find there is no basis for allowing a commission offset for these amounts. In the Preliminary Determination, we allowed for a commission offset for certain CSN sales that were initially sold to non- U.S. trading companies before being sold to the U.S. trading companies. This commission was based on a percentage of the value of individual orders (see CSN Verification Report dated July 31, 2001). We continue to allow for the commission offset for the sales between CSN and the non-U.S. trading company in our final results. Comment 9: U.S. Warranty Expenses - Direct Versus Indirect Expense Respondents contend that the Department improperly treated claims for defects on merchandise sold by USIMINAS and COSIPA to the United States as direct selling expenses. According to respondents, these claims are actually indirect warranty costs. Respondents state that the Brazilian mills did not intend to sell defective merchandise, and only a small portion ended up being defective. Respondents assert that while both companies determined to pay the claims, as of the time of their brief, the mills had not paid many of the claims. Respondents argue that claims for these defects were not part of the price that the mills agreed upon with their customers, and that claims against products already sold constitute warranty costs of the product. Respondents state that warranty costs form part of the price of the product only in the most general sense, in that a seller prices its product in order to cover warranty costs over the long run for all sales of the product. Respondents stress that a seller does not and cannot increase the price of individual sales to meet the cost of claims on those individual sales because it cannot know whether any individual product will incur any warranty costs. Therefore, respondents state the cost of warranty obligations must be spread over all sales rather than individual sales. Citing Notice of Preliminary Determination of Sales at Less Than Fair Value: Honey From Argentina, 66 FR 24108 (May 2001), and Circular Welded Non-Alloy Steel Pipe and Tube From Mexico: Final Results of Antidumping Duty Administrative Review, 65 FR 37518 (June 2000), respondents claim that the Department has recognized that warranty expenses should be spread over all sales. Additionally, respondents state that, in a 1999 decision, the Court of Appeals for the Federal Circuit approved the allocation of warranty expenses over total sales in each market (citing to NSK v. United States, 190 F 3d 1321 (Fed. Cir. 1999)). According to respondents, the Federal Circuit pointed out that the Department's customary practice is not only to spread warranty costs over all sales in a market, but also to take costs incurred during the period of investigation and apply them to all sales, regardless of whether they were incurred on the merchandise actually sold during the period. Respondents assert that this methodology takes into account the nature of warranty costs, which is that at the time of sale the possibility of claims cannot be known, and, therefore, the relationship between warranty costs and individual sales is indirect at best. Respondents argue that the Department should not deduct warranty costs for those sales with claims that have not yet been paid, since according to respondents, this is the Department's long-standing policy. Furthermore, respondents argue that even though USIMINAS and COSIPA can identify which sales had the defects that resulted in claims, it does not mean that the warranty costs should be applied only to those sales. Respondents state that most companies can isolate which sale generated the warranty claim because such information has to be available for the company to agree to pay the claim, but it does not mean that the warranty costs affected the price of that particular sale. Respondents argue that if the Department applies the warranty costs to the individual sales with the defects, then it will have to change its warranty cost methodology in all future cases. According to respondents, this will end up reducing the amount of warranty deductions in both the U.S. and home markets because the Department will have to identify which sales during the period actually had claims rather than taking claims paid during the period and applying them to sales during the period. However, if the Department determines that warranty cost adjustments are warranted, then respondents stress that the Department should adjust normal value by applying those warranty costs to all U.S. sales. Petitioners state that warranty expenses are direct selling expenses because the Department's regulations define direct selling expenses as "expenses such as commissions, credit expenses, guarantees, and warranties, that result from, and bear a direct relationship to, the particular sale in question" (citing to section 351.410(c) of the Department's regulations). Petitioners also disagree with respondents' argument that the Department should not make adjustments because the claims have accrued but have not been fully paid. Petitioners state that this does not make a difference because respondents account for expenses on an accrual basis as opposed to when the expenses have actually been paid. Petitioners argue that respondents' warranty claims are not indirect, since they do not relate to non-variable costs and can be associated with particular sales. Citing Certain Cut-to-Length Carbon Quality Steel Plat Products from Italy; Final Determination of Sales at Less Than Fair Value, 64 FR 73,234, 73,242-2 (December 29, 1999), petitioners state that the Department has treated customer claims concerning defective or non- conforming merchandise as direct warranty expenses. Petitioners assert that the Department's preferred practice is to apply warranty expenses to individual sales, and that the Department will typically permit allocation over all sales only if sale-specific reporting is not feasible. Petitioners cite to Brass Sheet and Strip from Canada, Final Results of Antidumping Duty Administrative Review, 64 FR 46,344, 46,347 (August 25, 1999), to support their assertion that, because certain warranty expenses are direct selling expenses, the Department typically will ask respondents to associate these expenses with the individual sale that caused the expenses. Petitioners point out that, in this case, respondents were able to validate the warranty claims on a sales-specific basis. Therefore, petitioners maintain that the Department properly adjusted only the specific sales to which the claims applied. Department's Position: The Department recognizes that warranty expenses can be direct or indirect expenses. For instance, in the September 2001 Questionnaire that was sent to respondents, the Department states, at page C-25 for field number 42.0 (WARRU or warranty expense), that "{w}arranty expense should include only the direct expense less any reimbursement received from the customer or unaffiliated parts supplier." Additionally, the Department states that the respondents should, "{r}eport indirect warranty expenses as part of indirect selling expenses." Accordingly, warranty costs may be a direct expense, and, in their Supplemental Questionnaire response, respondents provided a list of sales that had warranty claims and the Department was able to verify this information. According to section 351.410(c) of the Department's regulations, direct selling expenses are expenses that result from, and bear a direct relationship to, the particular sale in question. USIMINAS' and COSIPA's warranty expenses are expenses that are a direct and unavoidable consequence of the sale (i.e., in the absence of sales to the customers, the corresponding warranty claims would not have been incurred). Although respondents claim that most of the expenses have not been paid, as of the date of Respondents' Brief, documentation reviewed during verification indicates that the mills and the customers had agreed on the claim amounts. See Sales Verification of Section A - C Questionnaire Responses Submitted by Companhia Siderurgica Paulista ("COSIPA") in the Administrative Review of the Suspension Agreement on Hot-Rolled Flat- Rolled Carbon Quality Steel from Brazil, dated July 31, 2001, on page 41, and pages 1 - 3 of Verification Exhibit 17. Therefore, we have continued to classify USIMINAS' and COSIPA's warranty expenses as direct selling expenses. Comment 10: U.S. Credit Expense - Credit Days Respondents state that the Department improperly added additional days to USIMINAS' credit costs calculation for U.S. sales. Respondents state that USIMINAS receives payment from the trading companies via letters of credit, and funds are deposited into a crediting bank when shipping documents are presented to the bank. As a result, according to respondents, USIMINAS has full right to the funds, and, in fact, is supposed to withdraw the funds to its own Brazilian bank within 20 days of notification of receipt of the funds in the crediting bank. Respondents claim that USIMINAS does not withdraw funds from the crediting bank immediately because the funds are in U.S. dollars, which must be converted to reais for deposit to the Brazilian bank, and USIMINAS prefers to maximize the value of the sale in reais by converting the money when there is a favorable exchange rate. However, according to respondents, there is no practical or legal impediment to USIMINAS' withdrawing the funds from the crediting bank as soon as they are deposited. Respondents state that it is, in fact, USIMINAS' normal accounting procedure to credit the funds as received in U.S. dollars, and to record the sale as being liquidated (if an export/exchange contract advance is involved), as soon as the funds are placed in the U.S. crediting bank. Respondents maintain that USIMINAS' actual credit costs on U.S. sales do not extend beyond the time that the funds are placed in the crediting bank in the United States. Additionally, respondents maintain that USIMINAS incurs no actual credit costs between the time that the funds are deposited in the U.S. bank in dollars, and the time USIMINAS deposits the funds in the Brazilian bank in reais. Therefore, according to respondents, the Department's use of additional days in calculating credit costs is without any support in the administrative record. Petitioners argue that respondents have not provided evidence to support their position. Petitioners state that respondents are relying on verification documents that the Department used as a basis for adjusting the payment dates in the first place. Petitioners recommend that the Department disregard USIMINAS' claim that the alteration to the payment days was without support, and continue to adjust the payment days as in the Preliminary Results. Department's Position: Contrary to respondents' statement that USIMINAS records the sale as soon as the funds are placed in the U.S. crediting bank, during verification at USIMINAS, the Department found that USIMINAS records payments in its accounting records when the funds are deposited in the Brazilian bank, and not when they are deposited in the intermediary bank. See Department's Analysis Memorandum dated February 4, 2002. Additionally, regardless of respondents' assertions in their brief, it is not evident from the record that USIMINAS had full rights to these funds prior to their withdrawal from the intermediary bank. We are continuing to adjust USIMINAS' reported credit costs calculation for U.S. sales by the appropriate number of days. Comment 11: U.S. Credit Expense - Interest Rate Respondents argue that the Department improperly measured U.S. credit costs based on USIMINAS' and COSIPA's local currency borrowings, rather than dollar borrowings. The respondents state that the Department's conclusion that USIMINAS' and COSIPA's loans under Resolution 63 were dollar borrowings is wrong and is unsupported by any evidence on the record. The respondents explain that under the Resolution 63 program, Brazilian banks borrow dollars from abroad based on their own creditworthiness in international markets at rates based on the London Interbank Offered Rate (LIBOR). The respondents state that Brazilian banks then lend these funds in reais to Brazilian companies as "repass" loans, marked up by the banks to account for the creditworthiness of the Brazilian companies. A "repass" loan is a loan where Brazilian banks borrow from foreign lenders and then loan out these funds to third parties throughout Brazil. Respondents state that the banks calculate the repayment amount using the appropriate dollar/reais exchange rate. Respondents state that the bank takes out the loans in dollars, not the Brazilian company. Respondents point out that the company receives reais and is required to pay back the loan in reais. Respondents argue that while the amount of the loan and interest may be measured in dollars, these factors do not make it a dollar loan. Respondents further assert that many domestic contracts between Brazilian companies, including loans, are expressed in dollars and corrected for exchange rate calculations. Respondents contend that this affects the nature of the amount owing, not the nature of the loan itself. Respondents state that neither COSIPA nor USIMINAS has obtained any loans in dollars, and USIMINAS and COSIPA have not subjected themselves to credit ratings in the international market. The respondents point out that neither USIMINAS nor COSIPA classify Resolution 63 loans as dollar loans on their books. The respondents cite the June 2001 USIMINAS verification, where USIMINAS cited Note 7 of USIMINAS' consolidated financial statements (which include COSIPA) that identifies Resolution 63 loans as "local" loans measured in dollars. The respondents state that Note 7 explicitly distinguishes Resolution 63 loans from foreign loans, which are taken out in dollars. The respondents conclude that the Department's characterization of Resolution 63 loans as dollar loans is not supported by any evidence on the record. Respondents argue that contrasting USIMINAS' and COSIPA's Resolution 63 loans with their long-term dollar loans illustrates that the Resolution 63 loans are not dollar loans. Respondents point out that the long-term dollar loan interest rate is significantly different from Resolution 63 loans. Respondents argue that if the Department wishes to use dollar loans to determine USIMINAS' and COSIPA's credit costs, it can use USIMINAS' foreign borrowings in dollars, set forth in USIMINAS' financial statements. Petitioners argue that the Department correctly calculated USIMINAS' and COSIPA's U.S. credit costs. Petitioners state that respondents arguments are unfounded. Petitioners note that COSIPA correctly used an interest rate based on the Resolution 63 loans in calculating its reported U.S. credit expense for this review, whereas USIMINAS used the U.S. commercial short-term lending rate. Petitioners state that the respondents' Resolution 63 loans are dollar-denominated loans, provided to the company in reais. Petitioners point out that the Department's verification report clearly states that the loans are denominated in dollars, that the charge passed on to USIMINAS is based on dollar-denominated loans inclusive of a mark-up on the rate the Brazilian banks are charged, and that the loans take into account any exchange rate fluctuations. The Petitioners urge the Department to follow its policy established in Import Administration Policy Bulletin 98.2 (Feb. 23, 1998), and as followed in the Preliminary Results. Petitioners argue that in the Final Results, the Department should ignore the respondents' proposal to use the interest rate charged on the Brazilian mills' long-term U.S. dollar denominated loans in the U.S. credit calculation. Department's Position: We agree with petitioners that USIMINAS' and COSIPA's Resolution 63 loans are dollar denominated loans, provided to the companies in reais. Both companies had U.S. dollar borrowings under Resolution 63 and neither company had any other short-term U.S. dollar borrowings. We noted in loan documents examined at verification that the reais amount is calculated by applying an agreed upon exchange rate to the dollar amount. Thus the loans are denominated in dollars, which is the currency of U.S. sales. Policy Bulletin 98.2 (Feb. 23, 1998) states that the short-term borrowing rate realized by the respondent in the relevant currency is the best measure of the time value of money and the cost incurred by the respondent in extending credit to its customers. While the respondents did not directly subject themselves to credit rating in the international market and the amount received by the respondents was not in U.S. dollars, the respondents still accrued the cost of borrowing U.S. dollars. Given that both USIMINAS and COSIPA have short-term borrowings driven by the interest rate for U.S. dollars, these Resolution 63 loans most appropriately represent the cost incurred by USIMINAS and COSIPA in extending credit to its U.S. customers. The Resolution 63 loans do take into account the cost of borrowing in U.S. dollars. Additionally, we disagree with respondents' suggestion to use USIMINAS' foreign borrowings set forth in USIMINAS' financial statements because USIMINAS' financial statements do not differentiate between U.S. currency loans and other foreign currency loans. Comment 12: Freight Costs - Estimated versus Actual Respondents argue that the Department failed to adjust USIMINAS' and COSIPA's freight cost on individual sales in accordance with normal freight costs. Respondents note that the Department deducted actual freight costs of each shipment from the sales price of each individual sale item. Respondents point out that in a review of a suspension agreement under section 734 (c), the statute requires the Department to calculate "the amount by which the estimated normal value exceeds the export price (or constructed export price)." Respondents state that in a normal administrative review, the Department calculates "export price" and "normal value" in order to determine antidumping duties on each individual sale to the United States. Respondents assert that administrative reviews of suspension agreements are subject to different considerations from those that apply to normal administrative reviews. Respondents argue that the "estimated normal value" is the Department's signal to determine the extent to which the parties comply with their obligations under the Suspension Agreement, not to assess dumping duties as in a normal administrative review. Respondents assert that in complying with the Suspension Agreement, the parties must make their best efforts to avoid dumping the products. Respondents state that this obligation must be viewed in the context of the overall obligations that the companies and the Department have agreed to in the Suspension Agreement because a slavish conformity to the calculation of dumping margins could be counterproductive to the Suspension Agreement itself. Respondents note that Congress deliberately required the Department to determine estimated normal value, not normal value, within the larger context of negotiated agreements. Respondents explain that the Brazilian mills are obliged to price their products in their contracts on the basis of the landed price in the United States, regardless of what the ocean freight costs may actually work out to be when the product is shipped. Respondents state that the Brazilian mills estimate the cost of ocean freight. Respondents postulate that if actual ocean freight costs -incurred some months after the product has been sold - are higher than the amount initially estimated by the company when the company priced the product, then the sale can be below normal value. Respondents state that by using actual freight costs, the Brazilian mills cannot comply with the landed-duty-paid price terms of the Suspension Agreement without substantial risk of creating sales below normal value. Respondents assert that the deduction of actual freight costs creates an internal contradiction within the Suspension Agreement. Respondents state that in order to comply with the statute and the Suspension Agreement and resolve the internal contradiction, the Department should calculate "estimated normal value" based on the normalized costs of ocean freight instead of deducting actual ocean freight costs on each sale. Respondents clarify that the Department should adjust normal value by the amount that individual ocean freight costs on individual U.S. sales differ from the average ocean freight costs over the period. Respondents argue that normalizing freight values would ensure compliance with the landed-duty-paid price required by the Suspension Agreement and would comply with the requirement that the Department not create statutorily impermissible margins when compared with "estimated normal value." Respondents state that COSIPA and USIMINAS had sales that generated margins due to ocean freight being above the average ocean freight cost and higher than expected. Respondents contend that if the Department is to calculate entries below estimated normal value rather than sales below normal value as required by the statute, it would adjust normal value based on the difference between actual ocean freight and average normal freight. Respondents assert that using the average ocean freight costs over the period would eliminate COSIPA's and USIMINAS' impermissible margins generated by unusual ocean freight and therefore would no longer be below "estimated normal value." Petitioners argue that the Department properly calculated normal value by deducting actual ocean freight expenses. Petitioners note that respondents have not provided any statutory or regulatory authority, nor case precedent to support their claim that "normal value" and "estimated normal value" have different meanings or require different calculations by the Department. Additionally, petitioners point out that Article VIII of the Suspension Agreement states that the Department may conduct administrative reviews under section 751 of the Act. Section 751(a)(1)(C) of the Act provides that the Department shall review the current status of, and compliance with, any agreement by reason of which an investigation was suspended, and review the amount of any net countervailable subsidy or dumping margin involved in the agreement. Therefore, petitioners recommend that the Department disregard respondents' interpretation of estimated normal value and apply the same methodology from the Preliminary Results. Petitioners also point out that ocean freight costs are taken into consideration in the calculation of export price or constructed export price, and has no bearing on the calculation of normal value. Petitioners propose that if the Department wishes to use estimated ocean freight costs to compare to "estimated normal value," the Department should use average ocean freight costs to account for all fluctuations in freight (both increases and decreases in ocean freight costs). Department's Position: We agree with petitioners that normal value was properly calculated in the Preliminary Determination. The Department notes that section 751(a)(2)(A)(i) of the Act instructs the Department to calculate normal value and export price while section 734 of the Act uses the term "estimated normal value." However, neither the Act nor the Department's regulations instruct the Department to adjust normal value to account for aberrational cost adjustments that are actually deducted from the export price. The term "estimated normal value" is irrelevant to actual or average ocean freight expenses since ocean freight is not even deducted from normal value. Additionally, in prior final determinations, the Department did not allow for adjustments to normal value to account for aberrant actual expenses (high or low expenses that deviate from the norm) that are deducted from the export price. The Department's questionnaire and regulations instruct respondents to report averages only if expenses cannot be tied to a specific sale. The respondents reported actual ocean freight costs and as outlined by the Department's regulations (section 351.401), we are deducting these actual costs incurred by the mills from export price. For the above reasons, we will not adjust normal value to account for what respondents have characterized as aberrational ocean freight values. Comment 13: PIS and COFINS Taxes Respondents maintain that the Department should deduct PIS and COFINS taxes from home market sales prices. Respondents note that the tax adjustment provision of section 773(a)(6)(B)(iii) of the Act ensures that the Department makes a tax-neutral comparison when comparing normal value to export price by requiring the Department to adjust normal value by the amount of any indirect taxes imposed on home market sales, but not on export sales. Respondents state that until recently, the Department considered PIS and COFINS taxes to be indirect taxes that fall within the meaning of the tax adjustment provision. The Department's change in its treatment of these taxes, according to respondents, is based on a factually incorrect assumption that these taxes apply to total gross revenues and on a legally improper understanding of what indirect taxes are. Respondents point out that the statute and prior case law make clear that three circumstances must exist for the tax adjustment provision to apply to a particular tax. First, respondents state that the tax must be "directly" imposed on the home market product. Second, according to respondents, the tax must be rebated or not collected on export sales. Third, respondents note that the tax must be added to or included in the price of the home market sale. Respondents claim that the only requirements of significance in this review are the first and third requirements, as the second had never been an issue with respect to PIS and COFINS taxes. The respondents point out that in reaching its preliminary determination to deny an adjustment for PIS/COFINS taxes, the Department cited its "prior practice" as explained in the final determination from the original investigation. The respondents assert that the Department's current treatment of PIS/COFINS under section 773(a)(6)(B)(iii) of the Act mistakenly interprets PIS/COFINS as not being "imposed directly" on home market sales as required by the first prong of the statutory analysis. Respondents allege that the Department's current treatment of PIS/COFINS under section 773(a)(6)(B)(iii) centers on the first prong of the tax adjustment provision's requirements -- that the tax must be "directly" imposed on the home market product. The respondents maintain that in the final determination of this case, the Department did not adjust the mills' home market prices for Brazil's PIS/COFINS taxes because it incorrectly determined that these taxes are levied on gross revenue exclusive of export revenue. The respondents claim that the Department has not always taken this view of Brazil's PIS/COFINS taxes. The respondents contend that until the time of the final determination in the original investigations of this case, the Department consistently held that PIS/COFINS fall within the meaning of the tax adjustment provision. Respondents state that the Department only changed its established practice after it mistakenly determined that PIS/COFINS are taxes on total gross revenue excluding export revenues. Respondents state that the Department erroneously determined that PIS/COFINS are analogous to two Argentine taxes previously determined not to be indirect taxes within the meaning of the tax adjustment provision. Respondents state that in the Final Determination of the Less-Than-Fair Value; Silicon Metal from Argentina, 56 FR 37891 (August 9, 1991) (Silicon Metal from Argentina), the Department refused to make an upward adjustment to U.S. price for two Argentine taxes because these taxes were based on non-sales revenue as well as sales revenue. Respondents argue that the Department concluded that these taxes were not "directly" imposed on Argentine sales within the meaning of section 773(a)(6)(B)(iii) of the Act. According to respondents, the petitioners in Silicon Metal from Brazil; Final Results of Antidumping Duty Administrative Review and Determination Not to Revoke in Part, 62 Fed. Reg. 1970 (Jan. 14, 1997) (final admin. review)(Silicon Metal from Brazil) glossed over the fact that Brazilian and Argentine taxes are, in fact, vastly different by asserting that PIS/COFINS are "almost identical" to the two Argentine taxes. Respondents assert that, unlike the Argentine taxes, PIS/COFINS are imposed only on a company's sales revenue. Respondents note that despite a handful of recent Department determinations asserting otherwise, Argentine taxes in fact are not like PIS/COFINS taxes. As explained in USIMINAS/COSIPA's Section B Response, the respondents state that PIS and COFINS are not imposed on profits, exports, or non- sales revenue but, rather, are imposed only on a company's total domestic sales. Respondents additionally point out that CSN reinforced this fact during the June 2001 verification by demonstrating that the basis for both PIS and COFINS is gross sales, minus credit billing adjustments, canceled sales, and IPI, plus "other" sales revenue such as sales of services and scrap. Respondents point out that CSN also provided a monthly income statement in Verification Exhibit 19 of the June 2001 verification, demonstrating that it calculates its PIS/COFINS tax liability on sales revenue alone. Respondents explain that the Brazilian mills calculate PIS/COFINS tax liability in this manner as required under Brazilian law. Respondents point out that Article 2 of Supplementary Law No. 70 of December 30, 1991 specifies that the COFINS tax "shall be two percent and charged against monthly billing, that is gross revenues derived from the sale of goods and services of any nature." Additionally, respondents explain that Article 7 of the same law makes clear that the export sales shall be exempt from the tax. Respondents state that Supplementary Law No. 85 of February 15, 1996, an amendment to the original COFINS law, further clarifies that income derived from the sale of goods or service abroad and sales specifically meant to be exported are exempt from COFINS. Finally, respondents state that the COFINS tax rate increased from 2 percent (the rate applicable during the original investigation) to 3 percent on November 27, 1998, but the method for determining companies' tax liability remained unchanged. With regard to the PIS tax, respondents explain that Supplementary Law No. 7 of September 7, 1970 created this tax. Respondents state that the sample Provisional Measure 1,546-22 of August 7, 1997, demonstrates that the PIS shall be calculated on 0.65 percent of invoicing -- "invoicing" being defined as the gross revenue originating from the sale of goods from own account, from the price of the services rendered and from the result obtained from alien's (i.e., consignee's) account. Respondents further explain that Article Four specifies that PIS does not apply to export sales. Respondents point out that as these laws require, the mills pay the 3.0 percent COFINS tax on gross sales revenues and the 0.65 percent PIS tax on gross sales revenue, equaling the combined 3.65 percent PIS/COFINS adjustment that the respondents seek under section 773(a)(6)(B)(iii) of the Act. Respondents believe that the discussion above illustrates that the PIS and COFINS taxes are indirect taxes imposed directly on sales. Respondents maintain that liability for these taxes increases or decreases based on whether a company makes a home market sale, because neither tax is based on non-sales generated revenue. Respondents conclude that PIS/COFINS are not "gross revenue taxes" and, therefore, not analogous to the Argentine taxes in Silicon Metal from Argentina. In addition, respondents claim that the Department's decision not to make an adjustment for PIS and COFINS is unsupported by any accounting or economic analysis. Respondents contend that PIS and COFINS are sales taxes calculated on an aggregate basis as opposed to on an invoice-specific basis. The respondents state that the Brazilian laws creating PIS and COFINS require the mechanical calculation of the amount owed to the government each month, and that this calculation is based on a two-step process. Respondents explain that first, a total value for all home market sales is identified through all invoices in the home market generated during the period. Second, respondents state that the tax rate is applied to the total amount of these invoices. Respondents point out that a tax applied on an invoice-specific basis is different only in the mechanics of calculating the amount owed to the government. With regard to a tax applied on an invoice-specific basis, respondents state that the tax is calculated separately with respect to every individual invoice. The respondents state that the sum of the invoice-specific taxes represents the total amount that the company must remit to the government so the ultimate tax owed to the government is the same in both manners of calculating tax liability. Respondents claim that the Department's characterization of a tax as a "gross revenue" tax simply because tax liability is calculated on an aggregate basis is misplaced when the tax liability calculated on such a basis is identical to the tax liability calculated on an invoice-specific basis. Respondents conclude that no basis exists to conclude that the manner of calculating a tax disqualifies a tax from an adjustment under section 773(a)(6)(B)(iii) of the Act when it has no effect on the result. Respondents claim that the Department is trying to establish a distinction where none exists. Respondents state that the Department has not, in any of its decisions relating to this issue, identified any support for its classification of a sales tax as a "gross revenue tax" simply because it is calculated on an aggregate basis. Respondents state that there is no accounting principal, tenet in public finance, or academic analysis that supports a conceptual distinction between taxes calculated on a sale-by-sale basis and taxes calculated on an aggregate basis. Respondents assert that absent any conceptual support for this distinction, the Department should find that the PIS and COFINS taxes are imposed directly on home market sales. Respondents conclude that there could not be a more direct relationship between the underlying sales and the ultimate tax liability. Respondents maintain that the third prong, inclusion of the taxes in the home market price, is satisfied in the instant case. Respondents assert that the Department has historically adjusted for PIS and COFINS in Brazilian dumping investigations. Respondents note that as such, all of these previous determinations stand for the proposition that PIS and COFINS are included in the home market price. Respondents state that since inclusion in the home market price has been a precondition for the tax adjustment in these previous investigations, the Department's adjustment for these taxes was based, necessarily, on a finding that these taxes were included in home market price. Respondents claim that although the Department decided not to adjust home market price for PIS and COFINS in a few scattered decisions before the final determination in this case, the Department has never based its denial of the PIS/COFINS adjustment on a specific or explained finding that the taxes were not included in the price and passed through to the home market customer. As a result, respondents conclude that none of the Department's previous decisions supports a finding in this proceeding that PIS and COFINS are not included in the home market price. Respondents claim that the Department has a long-standing practice of finding full pass-through of taxes imposed on home market sales. Respondents state that the Department has held that the Department may, under the dumping law, presume that a company includes the full amount of home market taxes in its home market price and passes the tax through to its home market customer. Respondents note that in the Final Administrative Review of Color Television Receivers from Korea, 49 FR 50420 (December 28, 1984), the Department made an adjustment for home market taxes based on the conclusion that the taxes were fully passed through to the home market customers. Respondents assert that the Department determined that it was authorized to make an adjustment under section 772(d)(1)(C) of the Act. Respondents point to a series of subsequent well-known court appeals in which the legality of the Department's finding of pass-through was challenged. Respondents contend that in Daewoo Electronics v. International Union, 6 F.3d 1511 (Fed. Cir. 1993), the Court of Appeals for the Federal Circuit resolved the issue in deciding that the antidumping law did not require the Department to conduct an econometric analysis to determine if a home market tax is included in home market price. Respondents note that the Court agreed that the Department was correct in finding that the home market taxes were included in home market price, and correct to make an adjustment for these taxes. Respondents remark that the Department's practice has been upheld in other recent appeals. Respondents note that the Department's finding of pass-through of certain home market taxes was challenged in the Silicon Metal From Argentina proceeding discussed above, the very case the Department relied on when it changed its treatment of Brazil's PIS/COFINS under the tax adjustment provision. Respondents believe that while Silicon Metal from Argentina serves as the mistaken basis for the Department's current treatment of PIS/COFINS, American Alloys, Inc. v. United States, 810 F.3d 1469, 1475 (Fed. Cir. 1994) (American Alloys) upholds the principle that company costs like PIS/COFINS are passed through to customers. Respondents explain that in American Alloys, the Court endorsed the Department's finding of tax pass-through based on the Department's analysis of the exporters' questionnaire responses, and verification of the base tax rate. Respondents note that the Court also based its finding on a presumption that a company passes on costs to its customers. Respondents state that both Courts acknowledge the propriety of a finding that is based on the economic reality that if a company incurs certain tax expenses with respect to certain sales (i.e., home market sales), it will recoup the expenses in pricing the sale. Respondents argue that from a theoretical standpoint, it would be absurd for the Department to assume that a company does not recoup the costs of home market taxes from home market customers. Respondents state that if one presupposes that the expense of this tax is not recovered from home market sales, then one would have to conclude that it is recovered in the pricing of export sales, its only other sales. Respondents conclude that this then leads to the conclusion that the company is using export sales to subsidize expenses related exclusively to home market sales. Respondents note however, this conclusion contravenes the very premise of the Department's dumping investigation -- that the exporter is using profits from home market sales to absorb losses in connection with U.S. sales. Respondents state that as a result, it would be ironic for the Department to determine that a company is dumping based on a finding that the company is absorbing home market sales taxes with export sales revenue. Respondents conclude that based on the Department's practice, and the relevant Court interpretations of the antidumping law, the Department should find that PIS and COFINS are included in the home market price, and passed through to home market customers. Respondents contend that any finding that PIS and COFINS were not passed through must be based on some evidence on the record. Respondents reiterate that the Department is not justified in finding no pass-through in this investigation. Respondents acknowledge that conceivably the Department might argue that PIS and COFINS were not included in the price because they were not itemized on the invoice (like the IPI and ICMS taxes). However, respondents note that this justification fails for two reasons. First, respondents state that PIS and COFINS were not itemized on the Brazilian invoices in all of the Department's previous investigations. Respondents pointed out, however, that the Department always found that these taxes were included in home market price, and qualified for an adjustment. Second, respondents claim, whether or not the tax is itemized on the invoice is irrelevant to a pass-through finding. Respondents state that any tax not itemized on the invoice is included in the gross unit price. Respondents claim that the fact that the tax is not identified in a separate box indicating "price before PIS/COFINS" and "price after PIS/COFINS" is not material. Respondents assert that itemization on the invoice only indicates how, as a practical matter, the tax is calculated and maintained in the accounting records of the company. Petitioners argue that the Department should continue to calculate normal value without deducting PIS and COFINS. Petitioners note that the statute clearly states that downward adjustments to normal value may only be made for taxes directly imposed upon sales of the foreign like product. Petitioners contend that the tax liability for PIS and COFINS taxes is imposed on all the company's domestic sales revenue (including service revenue) on an aggregate basis. Petitioners state that contrary to respondents' suggestions, the Department is intimately familiar with the way the PIS/COFINS taxes are imposed and collected, and the Department has painstakingly reviewed this issue in several recent cases, such as, Silicon Metal from Brazil, 64 Fed. Reg. 6305, 6319 (Feb. 9, 1999) (Final Results of Antidumping Duty Administrative Review) and Certain Cut-to-Length Carbon Steel Plate from Brazil, 63 Fed. Reg. 12,744, 12,745-46 (Mar. 16, 1998) (Final Results of Antidumping Duty Administrative Review) among others. Petitioners point out that after careful consideration of the issue in these cases, the Department's practice in this area is now settled. Petitioners state that the Department disallows claimed adjustments to normal value for these taxes. Indeed, petitioners point out that in Respondents' Case Brief at 54, respondents concede as much. Petitioners state that respondents simply seek to overturn the Department's current practice, based on no new facts or new arguments. Petitioners affirm that there is no reason for the Department to disturb its settled practice on this issue. Department's Position: We did not make any changes to the calculation of normal value (NV) for the Final Determination. This issue is not relevant to our decision in this case because, regardless of whether an adjustment is made for PIS/COFINS taxes, we find that substantial sales did not eliminate 85 percent of the dumping margin determined in the investigation and a substantial number of sales were below reference price. Recommendation Based on our analysis of the comments received, we recommend adopting all of the positions set forth above and adjusting all related margin and reference price comparison calculations accordingly. If these recommendations are accepted, we will publish the final results in the Federal Register. AGREE____ DISAGREE____ Faryar Shirzad Assistant Secretary for Import Administration Date