67 FR 35482, May 20, 2002 A-469-811 Investigation Public Document IA/Group I/Office 2: ST MEMORANDUM TO: Faryar Shirzad Assistant Secretary for Import Administration FROM: Richard W. Moreland Deputy Assistant Secretary, Group I Office of AD/CVD Enforcement DATE: May 13, 2002 SUBJECT: Issues and Decision Memorandum for the Antidumping Duty Investigation of Structural Steel Beams from Spain Summary We have analyzed the case and rebuttal briefs of interested parties in the investigation of sales at less than fair value of structural steel beams from Spain. As a result of our analysis, we have made changes in the margin calculations for the final determination. 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 from interested parties: Comment 1: Level of Trade in the Home Market Comment 2: Level of Trade for U.S. Sales/CEP Offset Comment 3: Arm's Length Test Comment 4: Strength Codes Comment 5: Billing Adjustments Comment 6: Home Market Rebates Comment 7: Home Market and U.S. Freight Expenses Comment 8: Inland Freight Expenses of the Affiliated Resellers Comment 9: Home Market Credit Expenses Comment 10: U.S. Rebates Comment 11: U.S. Brokerage and Handling Expenses Comment 12: U.S. Indirect Selling Expenses of Arbed Americas Comment 13: Interest Expenses Included in U.S. Indirect Selling Expenses Comment 14: Clerical Errors in the Preliminary Determination Comment 15: Calculation of the Overall Dumping Margin Comment 16: Calculation of Raw Materials Costs Comment 17: Exchange Gains and Losses Comment 18: Acceptance of Revised Sales Databases Background On December 28, 2001, the Department of Commerce (the Department) published the Notice of Preliminary Determination of Sales at Not Less Than Fair Value and Postponement of Final Determination: Structural Steel Beams From Spain, 66 FR 66207 (Preliminary Determination). We invited parties to comment on our preliminary determination. The petitioners and the respondent, Aceralia Corporacion Siderurgica (Aceralia), both requested a public hearing, which was held at the Department on April 16, 2002. The period of investigation (POI) is April 1, 2000, through March 31, 2001. Margin Calculations We calculated constructed export price (CEP), export price, and normal value using the same methodology stated in the preliminary determination, except as follows: 1. We revised our margin calculations for Aceralia to take into account our findings at verification. For further discussion, see the May 13, 2002, memorandum from Patrick Connolly to the file entitled "Calculations Performed for Aceralia Corporacion Siderurgica, S.A. (Aceralia) for the Final Determination in the Less Than Fair Value Investigation on Structural Steel Beams from Spain" and the May 13, 2002, memorandum from Gina Lee to Neal M. Halper entitled "Cost of Production and Constructed Value Adjustments for Final Determination." 2. We included the downstream sales of certain of Aceralia's affiliated customers (i.e., those in the Velasco Group) in our analysis. We found that the sales to these customers were made at a different level of trade than those made by Aceralia to its customers. Where appropriate, we made margin comparisons across levels of trade. See Comment 2. 3. We determined that sales made to Aceralia's affiliated party in the United States, TradeARBED, were made at the same level of trade as Aceralia's direct sales in the home market. Therefore, we did not grant Aceralia a CEP offset for purposes of the final determination. See Comment 3. 4. We revised the product characteristics code for strength for certain products sold in the home market. We also revised the control numbers associated with these products using the revised codes and recalculated the cost of production (COP) accordingly. See Comment 4. 5. We did not take into account the information reported in the field REBATE3H in Aceralia's revised database because: 1) this information relates to future possible refunds of actual rebates that Aceralia paid to the customer; 2) Aceralia had not contacted the customer to request that the refunds be made; and 3) thus, there was no guarantee that the customer would grant these refunds. See the memorandum from Shawn Thompson and Elizabeth Eastwood to Louis Apple entitled "Verification of the Sales Responses of Aceralia Corporacion Siderurgica, S.A. in the Antidumping Duty Investigation on Stainless Steel Beams from Spain" (the home market sales verification report) at page 1 and verification exhibit 1. Rather, we based home market rebates on the amounts actually granted to Aceralia's customers on its POI sales. 6. We made an adjustment for credit expenses in the home market because we verified that Aceralia's payment dates were reported accurately. See Comment 9. 7. We made no deduction for commissions on export price sales because we found at verification that the reported amounts merely represented an intra-company markup (i.e., the "commission" amounts were not recorded in Aceralia's accounting records, nor were they paid to the affiliated reseller). See the home market sales verification report at page 34. 8. We increased certain U.S. foreign inland freight expenses by the highest positive percentage difference observed at verification. See Comment 7. 9. We accepted Aceralia's rebates in the U.S. market, as reported, based on our finding at verification that TradeARBED did not grant rebates on sales in the fourth quarter of 2000. See Comment 10. 10. We based the amount of U.S. brokerage and handling expenses for shipments through certain ports in the United States on adverse facts available. As adverse facts available, we used the highest expense on the record of this investigation. See Comment 11. 11. We recalculated U.S. imputed credit expenses using the combined U.S. interest rate of the two affiliated parties in the United States which had short-term borrowings during the POI. See the U.S. sales verification report at page 15. 12. We included a portion of the indirect selling expenses incurred by an affiliate of TradeARBED during the POI in the calculation of U.S. indirect selling expenses. See Comment 12. 13. We revised the offset to the interest expense component of U.S. indirect selling expenses to account for imputed credit expenses on a company- wide basis. See Comment 13. 14. We corrected a clerical error in the preliminary determination involving the conversion of certain expenses from a per-metric-ton to a per- hundredweight basis. See Comment 14. Discussion of the Issues Comment 1: Level of Trade in the Home Market During the POI, Aceralia sold structural steel beams in the home market through two channels: 1) directly to its own customers; and 2) through affiliated distributors. Based on our request, Aceralia reported downstream data for a number of its affiliated parties (collectively known as "Velasco"). In our preliminary determination, we rejected Velasco's reported downstream information because we found that it was so incomplete that it could not be used. Nonetheless, we gave Aceralia a chance to correct its data problems, and we verified this information in Spain. Both Aceralia and the petitioners agree that the Velasco data "passed" verification and thus should be used in the final determination. However, Aceralia and the petitioners disagree as to the appropriate use of this data. Aceralia contends that Velasco's downstream sales are at a different level of trade (LOT) than Aceralia's mill-direct sales. Specifically, Aceralia contends that Velasco's sales to its unaffiliated customers (which Aceralia characterizes as the "reseller LOT") are at a more advanced marketing stage and are supported by a substantially greater level of selling activities as compared to Aceralia's direct sales (which Aceralia terms the "Aceralia LOT"). Aceralia bases this conclusion on the following facts: 1) sales at the Aceralia LOT are primarily to large distributors/resellers, whereas sales at the reseller LOT are to smaller end-user customers and small local warehouses; 2) Aceralia's direct customers purchase in large volumes for inventory and resale, while the resellers' customers purchase in small volumes for consumption; 3) Aceralia sells only one type of product, while the resellers offer a range; 4) Aceralia's direct customers need minimum attention in terms of their purchases of beams, while the end-users at the reseller LOT have numerous and specific requests for special payment terms, specific freight and delivery arrangements, value-added services, technical advice and assistance, and marketing support; and 5) Aceralia provides limited payment terms and it generally establishes sales terms that govern over a period of time, while Velasco handles sales on a case-by-case basis. Aceralia contends that many of these conclusions were echoed by the Department in the verification report issued for Velasco, because this report describes the extensive services provided by the resellers to their unaffiliated customers. Aceralia notes that, in setting out the Department's approach to LOT, the preamble to the Department's regulations states that "{e}ach more remote level must be characterized by an additional layer of selling activities, amounting in the aggregate to a substantially different selling function. See the Preamble to the Department's Final Antidumping Duty Regulations, 62 FR 27296, 27371 (May 19, 1997) (Preamble). Aceralia maintains that under the Department's working methodology for analyzing whether differences in LOT exist within a market, the Department distinguishes between LOTs on the basis of qualitative and quantitative differences in selling activities, as well as differences in class of customer. For example, Aceralia maintains that it is the Department's practice to find two different LOTs where downstream sales benefit from services that other sales do not. As support for this assertion, Aceralia cites Notice of Preliminary Determination of Sales at Less Than Fair Value: Certain Cold- Rolled Flat-Rolled Carbon-Quality Steel Products from Brazil, 64 FR 61249, 61256 (Nov. 10, 1999) (Cold-Rolled Steel from Brazil) (where the Department found that downstream sales were at a different LOT than mill- direct sales because the resellers performed a higher degree of freight and delivery arrangements and provided inventory maintenance); and Circular Welded Non-Alloy Steel Pipe From Mexico: Preliminary Results and Partial Recission of Antidumping Duty Administrative Review, 65 FR 77560, 77563 (Dec. 12, 2000) (Steel Pipe from Mexico) (where the Department found that downstream sales were at a different LOT because the respondent performed additional selling functions to sell to its affiliated resellers which allowed the resellers, in turn, to perform selling functions for their unaffiliated customers). Consequently, Aceralia asserts that Aceralia's selling activities are quantitatively less and qualitatively different than Velasco's, which place: 1) Aceralia at a less advanced stage of marketing than Velasco; 2) sales by Velasco at a more remote stage of the chain of distribution. Thus, Aceralia contends that the Department should follow its practice, as outlined in Cold-Rolled Steel from Brazil and Steel Pipe from Mexico, and find two levels of trade in the home market. The petitioners agree that, in assessing level of trade, the Department will compare the cumulative selling functions along the chain of distribution of the producer and its affiliated reseller in the home market. The petitioners note that, if the cumulative selling functions are essentially the same, the Department will determine that the sales were made at the same level of trade. According to the petitioners, because the selling functions performed by both Aceralia and the resellers are essentially the same, the Department should find only one level of trade in the home market. Specifically, the petitioners note that Aceralia provided two charts in its response, one showing the selling functions performed by Aceralia and another showing the selling functions provided by Velasco. The petitioners assert that a simple comparison of these charts shows that both companies not only performed essentially the same selling activities, but that there was no qualitative difference between them. The petitioners argue that it is not relevant that Velasco serves as large number of small customers who often purchase in small quantities, because the Department has held in other cases that transaction size or unit value of sales are not selling functions and thus they are not considered in deciding LOT issues. See Notice of Final Results of Antidumping Duty Administrative Review, Partial Rescission of Antidumping Duty Administrative Review and Revocation of Antidumping Duty Order in Part: Certain Pasta From Italy, 67 FR 300 (Jan. 3, 2001) and accompanying decision memorandum at Comment 6. Therefore, the petitioners contend that only one level of trade exists in the home market. Department's Position: The framework for determining the number of levels of trade in a given market is set out in 19 CFR 351.412(c)(2). According to this provision, the Department will determine that sales are made at different levels of trade if they are made at different marketing stages (or their equivalent). Substantial differences in selling activities are a necessary, but not sufficient, condition for determining that there is a difference in the stage of marketing. Some overlap in selling activities will not preclude a determination that two sales are at different stages of marketing. In determining whether substantial differences in selling activities exist, the Department examines all selling activities which occur along the chain of distribution. Contrary to the petitioners' assertion, this does not simply involve comparing Aceralia's selling functions to the Velasco Group's selling functions. Rather, when a producer sells through an affiliated reseller, we consider the relevant functions to be the selling functions of both the producer and the reseller (i.e., the cumulative selling functions along the chain of distribution) for purposes of comparing the selling activities related to the affiliate's sale with those related to the producer's sale to its customers. (1) If the reseller performs selling functions which add substantial selling activity to make the sale, we may find that sales by the reseller are made at a different LOT than the LOT of the producer. In this case, we have examined the selling functions performed by Aceralia in selling to its direct customers (including the Velasco Group companies), as well as the additional selling functions performed by the Velasco Group companies to sell to their unaffiliated customers. We find that the additional selling functions performed by the Velasco Group are substantial. For example, we note that Velasco companies perform numerous selling functions, which include, among other things: 1) the maintenance of sales forces which solicit and process orders for their sizeable customer base (many of whom are end users); 2) the making of sales calls and demonstrations to customers; 3) the maintenance of a fleet of trucks for deliveries; 4) the maintenance of pre-sale inventories of products; and 5) the handling of certain additional services for their clients (e.g., cutting). See the February 20, 2002, memorandum from Michael Strollo and Irina Itkin to Louis Apple entitled "Verification of the Sales Questionnaire Responses of Laminados Velasco, S.L. in the Antidumping Duty Investigation of Structural Steel Beams from Spain" (the Velasco sales verification report) at page 3. Moreover, Velasco performed certain selling functions to a much greater degree or frequency than Aceralia. (2) When taken as a whole, these additional selling functions lead to the conclusion that the sales by the Velasco Group companies are at a more remote level in the chain of distribution and are made at a different LOT than Aceralia's direct sales. Therefore, we have treated sales by the Velasco Group as made at a separate LOT for purposes of the final determination. Comment 2: Level of Trade for U.S. Sales/CEP Offset In the preliminary determination, we found that Aceralia's home market sales were made at a single LOT. We also found that this level was at a different, and more advanced, LOT than the LOT of Aceralia's CEP sales. Therefore, we made a CEP offset adjustment in our preliminary calculations. Both Aceralia and the petitioners agree that there is no need to make a CEP offset for the final determination because both contend that CEP comparisons can be made at the same LOT. (3) However, Aceralia and the petitioners disagree on the reasoning behind this action. Aceralia contends that the Aceralia LOT is equivalent to the LOT of all U.S. sales, including both the export price LOT and CEP LOT after all deductions are made pursuant to section 772(d) of the Act. According to Aceralia, both its home market and U.S. sales involve substantially similar selling activities and expenses. Indeed, Aceralia maintains that the only differences in the selling functions performed for each market are limited to the provision of certain engineering services and inventory maintenance for home market direct sales. Aceralia contends that this analysis was confirmed at verification. According to Aceralia, the standard used to determine whether U.S. and home market sales are at the same LOT is that these sales must be made at "comparable marketing stages" with "similar" selling functions performed by the sellers. See Micron Technology Inc. v. United States, 243 F.3d 1301, 1305 (Fed. Cir. 2001) (Micron) and the Statement of Administrative Action (SAA) at 829. Thus, Aceralia concludes that minor differences between two LOTs do not warrant a finding that there are two different LOTs. Rather, Aceralia notes that the preamble to the Departments regulations explains that the Department to base normal value on the sales of foreign like product at the LOT "corresponding to" the U.S. or equivalent LOT. See Preamble, 62 FR at 27372. Finally, Aceralia contends that the law requires that the Department exhaust all possible matches of U.S. sales to home market sales at the same LOT before resorting to matching sales at a different LOT. However, Aceralia contends that, where this is not possible, the Department must make a LOT adjustment to normal value because there is record evidence supporting a pattern of price differences between the two LOTs in Spain that affects price comparability between U.S. price and normal value. In contrast, the petitioners argue that Aceralia's CEP sales are at the same LOT as home market sales made by Velasco. Again, the petitioners contend that a side-by-side comparison of the selling function charts submitted by Aceralia shows that there is considerable overlap in the activities performed by each of these parties. The petitioners contend that the appropriate starting point for this analysis is the CEP sale prior to the deductions under section 772(d) of the Act. (4) Nonetheless, the petitioners contend that, if the Department finds that two levels of trade exist in the home market, and a third level of trade exists in the United States, it would still compare CEP sales to sales made by Velasco. However, the petitioners concede that, under this scenario, the Department should make a CEP offset in accordance with its practice. According to the petitioners, this action would be consistent with the action taken by the Department in Notice of Final Determination of Sales at Less Than Fair Value: Stainless Steel Bar from Italy, 67 FR 3155 (Jan. 23, 2002), where the Department compared a respondent's CEP sales to its home market sales by an affiliated reseller based on a finding that the starting price sales for those were similar. Department's Position: The Department has consistently stated that the Act and the SAA support analyzing the level of trade of CEP sales at the constructed level, after expenses associated with economic activities occurring in the United States have been deducted pursuant to section 772(d) of the Act. In the preamble to our proposed regulations, we stated: With respect to the identification of levels of trade, some commentators argued that, consistent with past practice, the Department should base level of trade on the starting price for both export price EP and CEP sales... The Department believes that this proposal is not supported by the SAA. If the starting price is used for all U.S. sales, the Department's ability to make meaningful comparisons at the same level of trade (or appropriate adjustments for differences in levels of trade) would be severely undermined in cases involving CEP sales. As noted by other commentators, using the starting price to determine the level of trade of both types of U.S. sales would result in a finding of different levels of trade for an EP sale and a CEP sale adjusted to a price that reflected the same selling functions. Accordingly, the regulations specify that the level of trade analyzed for EP sales is that of the starting price, and for CEP sales it is the constructed level of trade of the price after the deduction of U.S. selling expenses and profit. See Antidumping Duties; Countervailing Duties; Notice of Proposed Rule Making and Request for Public Comments, 61 FR 7308, 7347 (Feb. 27, 1996). Consistent with the above position, in those cases where a level of trade comparison is warranted and possible, the Department evaluates the level of trade for CEP sales based on the price after adjustments are made under section 772(d) of the Act. See, e.g., Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, Germany, Italy, Japan, Sweden, and the United Kingdom; Final Results of Antidumping Duty Administrative Reviews and Revocation of Orders in Part, 66 FR 36551 (July 12, 2001) and accompanying decision memorandum at Comment 22; Large Newspaper Printing Presses and Components Thereof, Whether Assembled of Unassembled, From Japan: Notice of Final Determination of Sales at Less Than Fair Value, 61 FR 38139, 38143 (July 23, 1996). We note that, in every case decided under the revised antidumping statute, we have consistently adhered to this interpretation of the SAA and of the Act. See, e.g., Static Random Access Memory Semiconductors From Taiwan; Final Results of Antidumping Duty New Shipper Review, 65 FR 12214 (Mar. 8, 2000) and accompanying decision memorandum at Comment 3; Extruded Rubber Thread From Malaysia; Final Results of Antidumping Duty Administrative Review, 65 FR 6140, 6141 (Feb. 8, 2000); Notice of Final Determination of Sales at Less Than Fair Value: Dynamic Random Access Memory Semiconductors of One Megabit and Above ("DRAMs") from Taiwan, 64 FR 56308, 56313 (Oct. 19, 1999). Moreover, the CAFC has held that the statute unambiguously requires the Department to deduct the selling expenses set forth in section 772(d) of the Act from the CEP starting price prior to performing its LOT analysis. See Micron, at 1314-1315. Consequently, we have continued to adjust the CEP, pursuant to section 772(d) of the Act, prior to performing the LOT analysis, specified in 19 CFR 351.412. In this case, we have reconsidered the position, stated in the Preliminary Determination, that Aceralia's home market direct sales were at a different, and more advanced, LOT than its CEP sales after the deductions have been made pursuant to section 772(d) of the Act. We verified that Aceralia performs essentially the same selling functions to sell to its home market direct customers as it does to its U.S. affiliate, TradeARBED, in the U.S. market, after all selling expenses incurred in the United States were deducted from the CEP. For example, we note that Aceralia makes delivery arrangements for both home market and U.S. sales; it makes sales calls on both home market and U.S. customers; it accepts and processes orders; and it provides technical advice to all parties. See the home market sales verification report at page 5, as well as Aceralia's January 9, 2002, submission on level of trade. At verification, we noted that the degree or frequency of the selling functions associated with the activities differed; however, we find that these differences are not so significant as to rise to the level of a distinct LOT. Consequently, based on the foregoing analysis, we find that Aceralia's sales to TradeARBED are at the same LOT as the company's direct sales in the home market. As stated in the SAA, a constructed export price offset adjustment will be made only where: 1) the data available do not form an appropriate basis for determining an LOT adjustment under section 773(a)(7)(A)(ii) of the Act; and 2) normal value is established at a level of trade more remote from the factory than the level of trade of the CEP. Because we find that Aceralia's U.S. CEP and home market direct sales were made at the same LOT during the POI, we have not made a CEP offset adjustment for purposes of the final determination. Rather, in instances where we made comparisons across levels of trade, we made an LOT adjustment based on our finding that a pattern of consistent price differences existed between the home market levels of trade, in accordance with section 773(a)(7)(A) of the Act. Regarding Aceralia's EP sales, however, we note that the LOT differs from the LOT for CEP sales. Specifically, the Department is directed under 19 CFR 351.412(c) to base the LOT analysis for EP transactions on the starting price to the first unaffiliated customer, before any adjustments are made under section 772(c) of the Act. See Preamble, 62 FR at 27370. As a consequence, we attempted to analyze the selling functions performed by Aceralia's affiliated exporter, Aristrain Hispano Trade GmbH (AHT), in making the sale to the U.S. customer. (5) At verification, Aceralia stated that it performed similar selling functions to sell to AHT as it did to sell to TradeARBED. However, Aceralia did not provide the selling functions performed by AHT to sell to its unaffiliated customers, nor did it respond to questions on this topic posed to it in a supplemental questionnaire. (6) Given that we asked Aceralia to describe any differences between AHT's sales process and that used by TradeARBED and Aceralia did not identify any, we have made the reasonable assumption that the selling functions performed by AHT are the same, or substantially similar to, the selling functions performed by TradeARBED in selling to unaffiliated parties in the United States. As noted above, when a producer sells through an affiliated reseller, we consider the relevant functions to be the cumulative selling functions along the chain of distribution. In this case, we find that AHT performed many of the same selling functions performed by the Velasco Group companies in the home market. Because we do not have sufficient information on the record to determine whether AHT's selling functions in toto are quantitatively or qualitatively different than Velasco's, as facts available, we find that they are not different. Accordingly, we find that Aceralia's EP sales through AHT are at the same LOT as the Velasco Group's sales in the home market. As with CEP transactions, in instances where we made comparisons across levels of trade, we also made an LOT adjustment for purposes of the final determination. Finally, we disagree with Aceralia's contention that the Act requires the Department to exhaust all possible matches of U.S. sales to home market sales at the same LOT before resorting to matching sales at a different LOT. Rather, section 773(a)(1)(B) of the Act requires that, to the extent practicable, the Department base normal value on sales of the foreign like product at the same LOT as EP or CEP. Under section 771(16) of the Act, the term "foreign like product" is defined first as merchandise which is identical to the subject merchandise, and then as merchandise which is similar. We interpret this section of the Act to mean that we are required to compare products based on physical similarity first, before considering whether they are at an equivalent level of trade. To do otherwise would lead to the illogical result of disregarding identical comparisons where they exist in favor of similar comparisons, in direct contravention of the requirement of section 771(16) of the Act. Therefore, we have compared U.S. sales to sales of foreign like product based on the physical similarity of the merchandise, in accordance with our practice. In this case, we found that all U.S. products had an identical or most similar match at the same LOT in the home market. However, if we had been faced with the choice of making an identical comparison at the different LOT in the home market or a similar comparison at the same LOT, we would have made the identical product comparison at a different LOT. Comment 3: Arm's Length Test In its initial questionnaire response, Aceralia treated all customers who were in a particular corporate grouping as a single customer (i.e., it reported a single customer code for the group). In response to a supplemental questionnaire, Aceralia also provided separate customer codes for each individual company within each corporate group code. We treated these individual customers as separate entities in our margin calculations performed for the preliminary determination, and thus we performed the arm's length test for each individual affiliated customer. In accordance with our practice, we disregarded sales to affiliated customers only where we found that these sales were not made at arm's length. The petitioners claim that the Department misunderstood Aceralia's response on this topic. Specifically, the petitioners assert that the individual customers within each grouping are not separate entities, but rather are merely locations or divisions of the larger customer. According to the petitioners, the Department's policy in such situation is to apply the arm's length test to the corporate group as a single entity. For example, the petitioners contend that the Department confronted an identical fact pattern in an administrative review involving carbon steel flat products, where the Department collapsed all divisions or delivery points of particular customers into a single entity before performing the arm's length test. See Certain Corrosion-Resistant Carbon Steel Flat Products From Japan: Final Results of Antidumping Duty Administrative Review, 65 FR 8935, 8945 (Feb. 23, 2000) (Carbon Steel Flat Products from Japan). Consequently, the petitioners assert that the Department should treat all separate customers within the customer code grouping for Aceralia's affiliated home market distributor, Velasco, as a single entity for purposes of the arm's length test. The petitioners further argue that, if Velasco's sales fail the arm's length test, the Department should base normal value on Velasco's downstream sales, rather than on Aceralia's sales to Velasco. Aceralia argues that the Department should continue to perform the arm's length test on each of the individual customer codes reported in its home market sales listing. According to Aceralia, each of the companies is a separate legal entity that prepares its own financial statements. Aceralia notes that it offers different terms of sale to these companies, as well as different rebate programs. Finally, Aceralia states that the group customer codes are maintained only for internal management purposes, and thus they do not support a finding that individual Velasco codes represent different customers. Aceralia asserts that cumulating the sales of the Velasco group when the arm's length test is performed would only serve to mask any differences in price caused by the rebates granted to some customers and not others. Aceralia contends that the factual situation here is not analogous to the situation in Carbon Steel Flat Products from Japan, because in that case (unlike here) the separate customer codes merely represented different divisions of the same company. Rather, Aceralia maintains that this situation is more similar to that in Grey Portland Cement and Clinker From Mexico; Final Results of Antidumping Duty Administrative Review, 67 FR 12518 (Mar. 19, 2002) and accompanying decision memorandum at Comment 18 (Mexican Cement), where the Department did not use codes that potentially grouped customers but instead used codes that identified each customer separately. Department's Position: The information on the record shows that the companies in the Velasco group are separate legal entities, rather than mere divisions or "delivery points" of a single company. Specifically, Aceralia has placed on the record of this investigation the consolidated financial statements of the Velasco Group, which show that there are a number of individual affiliates. In addition, Aceralia provided excerpts from the financial statements of several of these companies at the verification conducted at Laminados Velasco, S.L. (which is itself a separate and distinct legal entity). See the Velasco sales verification report at verification exhibit 2. In cases where customers are separate legal entities, we find that it is appropriate to treat them as separate affiliates rather than as units of a single company. We note that the petitioners argued at the public hearing held in this case that Aceralia's use of a single name for these companies (i.e., Velasco) and the preparation of a single rebate worksheet for them proves that Aceralia itself believes that these companies constitute a single entity. However, we find that the former merely represents a matter of convenience, (7) while the latter was done in an apparent attempt by Aceralia to mitigate its reporting burden. As noted in the home market sales verification report, we found at verification that Aceralia correctly matched all program-specific rebates to its home market sales. (8) Moreover, even though Aceralia granted one type of rebate based on the collective experience of the Velasco companies, (9) this action does not outweigh the fact that these companies are separate legal entities, with their own sales forces and their own customer bases. We disagree with the petitioners that the facts in this case are analogous to those in Carbon Steel Flat Products from Japan. In that case, the entities in question were not separate companies, but rather were "divisions or delivery points within a single customer." See Carbon Steel Flat Products from Japan, 65 FR at 8945. Thus, these entities simply represented certain functional or administrative units of particular companies, but were not companies in and of themselves. We note that 19 CFR 351.403 directs the Department to perform the arm's length test for each affiliated party, rather than on the basis of a broader group. Therefore, in accordance with our practice and our regulations, we have continued to perform the arm's length test on home market sales made by each separate affiliated reseller during the POI. Comment 4: Strength Codes According to Department's questionnaire, yield strength is one of the criteria used to determine product similarity between beams sold in the home and U.S. markets. The questionnaire instructed respondents to report the strength of their subject beams within certain ranges using specific codes. For example, products with yield strengths between 50,000 pounds per square inch (psi) and 65,000 psi with atmospheric corrosion resistance were assigned a code of "41," products in the same range which were heat treated, not for corrosion resistance, were assigned a code of "42," etc. Where specific codes were not identified, the questionnaire instructed Aceralia to report a strength code of "43" (i.e., other) within the 50,000 to 65,000 psi range and "52" for strengths in excess of 65,000 psi. In its initial questionnaire response, Aceralia reported a strength code of "43" for two different grades of beams, S 355 and A-572-50. Because Aceralia indicated that these grades had different yield strengths, we requested that Aceralia revise its reported codes to assign discrete strength codes to products having unique yield strengths and to reassign its control numbers accordingly. Aceralia did so, and we used the revised strength codes and control numbers in our analysis for purposes of the preliminary determination. The petitioners contend that the Department's instructions in this case were inconsistent with those in the seven other concurrent investigations on structural steel beams. Specifically, the petitioners assert that the Department did not direct the respondents in any of the other seven cases to report separate strength codes for products that would otherwise fall within the "other" strength code category. Indeed, the petitioners assert that the Department expressly rejected the respondents' efforts to introduce new strength codes in the companion investigations on structural steel beams from Germany and Luxembourg. The petitioners contend that there is no justification for the Department to make a change to the model match criteria where the difference does not reflect any meaningful difference between two products in the market. The petitioners further contend that such a change is especially unjustified here, given that Aceralia admits that the two grades are equivalent. Thus, the petitioners contend that the Department should reassign both the strength codes and control numbers to treat these two grades as the same product. In addition, the petitioners contend that the Department should weight average the costs of production for these grades using the revised control numbers. Aceralia contends that the reporting of distinct grade codes for the two grades in question is appropriate. According to Aceralia, there is no requirement in the Act or the Department's regulations that binds the Department to use the same product definitions among different antidumping proceedings. Indeed, Aceralia notes that the Department considers each investigation or administrative review to be a separate proceeding, with the facts guiding each proceeding separately. As support for this assertion, Aceralia cites: 1) Heavy Forged Hand Tools From the People's Republic of China; Final Results and Partial Rescission of Antidumping Duty Administrative Review and Determination Not To Revoke in Part, 66 FR 48026 (Sept. 26, 2001) and accompanying decision memorandum at Comment 20 (Hand Tools from the PRC), where the Department calculated ocean freight rates in a manner claimed by the petitioners to be inconsistent with other recent investigations; and 2) Notice of Final Determination of Sales at Not Less Than Fair Value: Pure Magnesium from the Russian Federation, 66 FR 49347 (Sept. 27, 2001) and accompanying decision memorandum at Comment 4, where the Department treated by-product costs in a manner claimed by the respondent to be inconsistent with a previous investigation on the same product. Thus, Aceralia contends that the Department has wide latitude to diverge from the product characteristics used in the companion cases in order to define the product characteristics with more accuracy here. Aceralia asserts that the petitioners acknowledge the differences in the physical characteristics of the products in question, but minimizes the importance of the distinction. Aceralia contends that it is the Department's preference to compare U.S. sales to identical merchandise in the home market, and that acceptance of the petitioners' argument will not result in an identical comparison. Thus, Aceralia asserts that the petitioners' argument does not provide any valid basis for the Department to reverse its instructions to assign more accurate strength codes and, therefore, the petitioners' argument should be rejected. Department's Position: During the POI, Aceralia sold beams made of S 355 grade steel in the home market. In addition, it sold beams made of A-572-50 grade steel in the United States. Because both of these grades have a minimum yield strength between 50,000 psi and 65,000 psi and were neither heat-treated nor corrosion-resistant, Aceralia assigned both grades a strength code of "43" (i.e., other). In its supplemental questionnaire response, Aceralia stated the following: These grades are equivalent to each other, but there is a slight difference in their yield strengths. The minimum yield strength of the S 355 is 355 newtons/millimeter while the minimum yield strength of the A-572-50 is 345 newtons/millimeter. See Aceralia's November 8, 2001, supplemental response, at pages 3 and 4. Although we accepted a distinction in the two grades for purposes of the preliminary determination, we have reconsidered our position on this matter. By Aceralia's own admission, these grades are equivalent, having only "slight" differences in their yield strengths. Thus, we find that this difference is not meaningful in the context of the Department's model matching framework. In crafting the criteria used both to define specific control numbers and to determine appropriate model matches, the Department does not attempt to identify every physical characteristic which may differentiate products. Rather, it is to establish certain categories which capture the essential nature of various product groups. For example, in this case we defined the shape of structural steel beams using a combination of beam type (e.g., wide flange, standard, etc.), section depth ranges (e.g., up to12 inches, between 12 and 18 inches, etc.) and size ranges (e.g., greater than 65 lbs. per foot, etc.). (10) Thus, it is not the goal of the Department to define products so narrowly that "identical" products are identical in every particular, but instead to treat models that are essentially the same as a single product. The purpose behind this methodology is to capture the various price and cost break points in order to make meaningful comparisons. In this case, we have determined that the Department's questionnaire defines the "other" categories for yield strength in a manner which is sufficiently narrow to permit appropriate comparisons in this particular instance. Specifically, we note that this category includes only yield strengths in a particular strength range with properties not explicitly identified in other categories. (11) Therefore, we have assigned the same yield strength to both grades S 355 and A-572-50 for purposes of the final determination, and we have redefined Aceralia's control numbers accordingly. Finally, because it is the Department's practice to calculate costs by control number, we have recomputed the cost for all products made of these grades of steel as the weighted-average of the reported costs. Comment 5: Billing Adjustments According to the petitioners, the Department discovered at verification that Aceralia had double counted the majority of the billing adjustments reported in its home market sales listing - once in the calculation of gross unit price and again under the billing adjustment field. The petitioners assert that the Department should, therefore, set billing adjustments equal to zero in all instances where they were double counted. Aceralia notes that it netted all billing adjustments out of gross unit price in its revised database submitted after verification, using the data contained in the relevant home market sales verification exhibit. Therefore, Aceralia agrees that no further adjustment to gross unit price related to billing adjustments is necessary. Department's Position: We have confirmed that Aceralia deducted all billing adjustments from the gross unit price in its post-verification home market sales listing. Therefore, we have made no further adjustment to the home market price for billing adjustments. Comment 6: Home Market Rebates According to the petitioners, the Department was unable to verify the accuracy of certain rebates reported in the home market sales listing. The petitioners assert that Aceralia was unable to tie the rebates in question to its accounting system, nor was it able to demonstrate that Aceralia paid the customer for one specific type of rebate. Therefore, the petitioners contend that the Department should disallow the rebates in question for purposes of the final determination. Aceralia disagrees that the Department could not verify the rebates at issue. Rather, Aceralia asserts the Department verified the existence of Aceralia's rebate programs, as well as the per-unit amounts by program for specific invoices without discrepancy. Aceralia concedes that its accounting system does not contain a specific account for rebates and thus it was unable to verify its total rebates as a separate item. However, Aceralia notes that it was able to demonstrate that the net sales figure (after deduction of discounts, rebates, and commissions) reported in the home market sales listing reconciled to the total net sales amount recorded in its accounting system to within a negligible amount. Furthermore, Aceralia notes that the other "discrepancy" identified by the petitioners relates to one adjustment to a single rebate program for only one of the company's four mills. Aceralia asserts that, because: 1) this rebate program did not vary by customer or mill; and 2) it was able to verify the same rebate paid to another customer for another mill, the Department was able to verify the accuracy of this rebate program. Thus, Aceralia contends that the Department should accept its rebates as reported for purposes of the final determination. Department's Position: We disagree that Aceralia was unable to support its rebate calculations. At verification, we examined Aceralia's's rebate programs in depth. We found that rebates were not recorded in a separate account in Aceralia's accounting system, but rather were treated as offsets to sales revenue. Moreover, these rebates were not segregated in the company's accounting records from other types of credit notes issued by Aceralia. Given the limitations of Aceralia's record keeping, we were unable to tie the total rebate expenses for specific programs directly to the general ledger. Nonetheless, we are confident that Aceralia did not overstate the amount of the rebates in question based on our findings at verification. Specifically, we note that Aceralia demonstrated that the rebates reported in the home market sales listing tied in toto to the amounts reported in the accounting system. (12) In fact, based on the testing procedures performed at verification, we found that the expenses reported in the home market sales listing were lower than those recorded in the accounting system. Thus, we find that Aceralia's methodology for reporting rebates was conservative. Regarding the petitioners' specific concern that Aceralia was unable to demonstrate that it granted one specific type of rebate to one customer, we note that this issue arose in the context of examination of pre-POI rebates, and, consequently, the amount in question represented a deduction to the total rebates reported for this customer. Thus, we find that allowing the deduction in question is appropriate. (13) Comment 7: Home Market and U.S. Freight Expenses The petitioners contend that at verification the Department discovered wide-ranging discrepancies in Aceralia's reported freight expenses. Specifically, the petitioners assert that the Department found that Aceralia either under- or over-reported the freight expenses for six of the 11 home market sales, and for all of the 12 U.S. sales, examined during verification in Spain. According to the petitioners, where the Department cannot verify the accuracy of the information submitted with respect to a given adjustment, it must apply partial facts available. As facts available, the petitioners contend that the Department should use the lowest inland freight expense reported on home market sales and the highest foreign inland freight expense reported on U.S. sales. Aceralia argues that the Department should accept the freight expenses in question as reported. According to Aceralia, these expenses were based on the contract rates in effect with the company's freight suppliers. Aceralia asserts that use of the contract rates was the most accurate methodology available to it, given that the reporting of transaction- specific data would have been virtually impossible because: 1) this data was not maintained on the company's computer system; and 2) the company had tens of thousands of sales of beams during the POI. In any event, Aceralia notes that the freight expenses were identical to the contract rates for half of the home market sales examined at verification and approximately the same for the remaining home market transactions. Aceralia further notes that, where the home market expenses differed, Aceralia under-reported the expense to a greater degree than it over- reported the expense. Finally, Aceralia asserts that the data reported is far more accurate that the petitioners' proposed expense because it takes into account the difference in expenses arising from shipping different lengths of beams to different regions and/or from different ports in Spain. Department's Position: At verification, we examined the freight expenses for nine shipments to home market customers and 12 shipments to the United States. (14) We found that the expenses for all but two of the home market shipments were either accurately reported or understated. Regarding the remaining transactions, we found that the amounts reported were overstated to such a slight degree that this error would have a negligible impact on the margin. See the home market sales verification report at pages 23 and 24. Based on these facts, we find that Aceralia's methodology for reporting home market freight expenses was conservative. Consequently, we have continued to accept these expenses for purposes of our final margin calculations. Regarding U.S. sales, however, we agree with the petitioners that none of the U.S. foreign inland freight expenses examined at verification could be tied accurately to source documentation. Specifically, we found that Aceralia understated the expenses on approximately half of the transactions, and it overstated them on the remainder (sometimes by a significant amount). See the home market sales verification report at page 32. Consequently, because Aceralia's methodology for reporting these expenses yielded results which were inaccurate in every case, we have based the amount of foreign inland freight for all U.S. transactions not examined at verification on facts available, in accordance with section 776(a)(2)(D) of the Act. Moreover, because Aceralia had the ability to provide accurate information to the Department, we find that Aceralia failed to cooperate by not acting to the best of its ability, in accordance with section 776(b) of the Act, and thus we find that an adverse inference is appropriate. As adverse facts available, we have increased the reported expenses by the highest percentage difference observed a verification. We disagree with the petitioners' suggested methodology of using the highest reported foreign inland freight expense as facts available. We find that this methodology would be distortive, given: 1) the large variation in the verified per-unit expenses; and 2) there was no consistent pattern of under-reporting associated with these expenses. Comment 8: Inland Freight Expenses of the Affiliated Resellers The petitioners note that Aceralia was unable to report transaction- specific freight information for its affiliated resellers, and thus it reported an average freight expense for these companies. The petitioners contend that the Department should reject this average expense because: 1) it could not be verified; and 2) the resellers were unable to support the alternative calculation provided at verification. Therefore, the petitioners argue that the Department should base the amount of freight expenses for these companies on the lowest per-ton amount reported by Aceralia in its home market sales listing. Aceralia maintains that the Department fully verified the revised per- unit freight expense provided at verification. Therefore, Aceralia contends that this freight expense should be used for purposes of the final determination, in the event that the Department matches U.S. sales across levels of trade. Department's Position: We agree with the petitioners that the freight expenses reported by the affiliated resellers could not be tied to source documentation. However, we disagree that the revised calculation provided at verification could not be verified. Rather, the verification report clearly indicates that all components of the revised calculation tied to audited financial statements without discrepancy. See the Velasco sales verification report at page 12. Therefore, because accurate and verified information exists on the record of this proceeding, we have accepted this information for purposes of the final determination. Comment 9: Home Market Credit Expenses In its questionnaire response, Aceralia indicated that it was unable to report transaction-specific dates of payment. As an alternative, it reported estimated dates of payment based on the payment terms stated on the invoice. It also reported average credit periods by customer and mill. For purposes of the preliminary determination, we disallowed an adjustment for credit expenses because the information reported by Aceralia on its customer-specific credit periods appeared to be contradictory and/or inaccurate. In its case brief, Aceralia argues that the Department should use the reported home market credit expenses in its final calculations. Specifically, Aceralia contends that the Department verified that the reported payment dates were either the same, or slightly earlier, than the actual payment dates for 10 out of 11 sales examined at verification and thus were either accurate or conservative. Finally, Aceralia contends that the Department should not recalculate credit expenses using the reported customer-specific average payment periods because: 1) the periods reported for one of the four mills were distorted by different payment terms offered on non-subject merchandise; and 2) the Department verified that the periods reported for the other mills contained a methodological error in the calculation formula. The petitioners did not comment on this issue. Department's Position: At verification, we confirmed that Aceralia received payment for the vast majority of the home market sales examined on or after the due date set forth in the customer's payment terms. Therefore, we have made an adjustment for home market credit expenses in our final margin calculations. Specifically, where verified actual payment dates exist on the record of this proceeding, we calculated credit expenses using these dates. Where actual payment dates were not available, we accepted Aceralia's credit expenses based on the reported payment terms because we find these expenses to be a conservative estimate of the actual credit amount. Comment 10: U.S. Rebates In its questionnaire response, Aceralia stated that its U.S. affiliate, TradeARBED, granted rebates to certain U.S. customers in 2000. However, Aceralia did not report any rebates on sales to these customers in the fourth quarter of the year. Consequently, in our preliminary determination, we based the amount of the rebates for this quarter on facts available. Aceralia contends that the Department verified that its affiliate did not grant any rebates on sales made in the fourth quarter of 2000. Therefore, Aceralia argues that the Department should use the rebates reported, without resorting to facts available, for purposes of the final determination. The petitioners did not comment on this issue. Department's Position: At verification, we confirmed that TradeARBED did not grant rebates on the sales in question during the fourth quarter of 2000. See the March 25, 2002, memorandum from Irina Itkin to Louis Apple entitled "Verification of the Sales Questionnaire Responses of TradeARBED Inc. in the Antidumping Duty Investigation of Structural Steel Beams from Spain" (the U.S. sales verification report) at page 11. Therefore, we made no adjustment for rebates for these sales for purposes of the final determination. Comment 11: U.S. Brokerage and Handling Expenses The petitioners contend that the Department should base Aceralia's U.S. brokerage and handling expenses on facts available. The petitioners maintain that Aceralia failed to report transaction-specific expenses, even though: 1) it admitted that it was able to do so; and 2) it received several extensions of the time limit for submitting its U.S. sales response (and thus it had adequate time to do so). Moreover, the petitioners assert that the Department was unable to verify the accuracy of port-specific data reported in lieu of transaction-specific expenses. Specifically, the petitioners claim that the expenses reported for both of the ports examined at verification contained errors. More significantly, the petitioners note that the Department was unable to examine additional calculations at verification because the documentation was not available, despite the fact that TradeARBED had months to prepare. According to the petitioners, the Department should find that TradeARBED did not act to the best of its ability in reporting brokerage and handling expenses and consequently it should apply adverse facts available to these expenses for the final determination. As adverse facts available, the petitioners argue that the Department should use the highest expense calculated for brokerage and handling. Aceralia argues that the Department should accept the brokerage and handling expenses presented at verification. Specifically, Aceralia contends that, of the two port-specific calculations examined at verification, the Department found that one of the figures was correct and the errors noted for the other were minor. Aceralia contends that, given that the data at TradeARBED on the whole verified, the Department should not assume that all of TradeARBED's brokerage and handling expenses are unreliable. Regarding the petitioners' claim that TradeARBED did not act to the best of its ability in reporting brokerage and handling expenses, Aceralia notes that this topic was one of the last examined at verification. Aceralia claims that it is unreasonable to expect respondents to prepare all details of every item of every expense in advance, in case the Department decides to randomly sample additional items. Finally, regarding the petitioners' allegation that TradeARBED should have reported transaction-specific expenses, Aceralia contends that it was unnecessary to do so, given that the charges incurred at each port were standard. According to Aceralia, it would be inappropriate to apply as facts available only the highest per-unit charge incurred because to do otherwise would distort the actual, verified information on the record. Department's Position We disagree with Aceralia, in part. On the first day of the U.S. sales verification, TradeARBED presented the Department with a list of corrections to the U.S. sales database. As part of these corrections, TradeARBED provided revised U.S. brokerage and handling expenses for the majority of the ports through which it imported subject merchandise during the POI. See the U.S. sales verification report at pages 1 and 2. In order to demonstrate the accuracy of the revised expenses, TradeARBED provided the worksheets used to calculate the amounts for two ports, San Francisco and Mobile. We tied the expenses reported for San Francisco to the company's accounting system without discrepancy; however, when we attempted to perform the same procedures for Mobile, we found that TradeARBED had misallocated the expenses in question and, as a result, had understated the revised amount. At this point, we requested that TradeARBED provide brokerage invoices for certain other ports, in accordance with normal verification procedures. TradeARBED informed us that it was not prepared to provide these invoices, claiming that they were "not available." See the U.S. sales verification report at page 13. We disagree with Aceralia that it was unreasonable to expect TradeARBED to provide documentation for additional ports at verification. The information requested was not a random sample (as Aceralia implies), but rather related to transactions explicitly identified in the verification outline issued to TradeARBED more than two weeks prior to the Department's requests at verification. (15) Indeed, we note that TradeARBED must have been in possession of these invoices prior to the start of verification, in light of the fact that it revised the brokerage and handling expenses for the majority of its ports during its preparation for verification. Thus, we find that it is disingenuous of Aceralia to argue that it could not anticipate our request to review further documentation, and consequently that it had inadequate time to gather the data. Section 776(a) of the Act provides that the Department may use facts available in a determination if an interested party either withholds requested information or supplies information which cannot be verified. Section 776(b) of the Act permits the Department to make adverse inferences in cases where it finds that the interested party failed to cooperate by not acting to the best of its ability. At verification, we found that TradeARBED's brokerage and handling expenses related to certain U.S. ports could not be verified. Therefore, we find that the use of facts available is appropriate for these expenses, in accordance with section 776(a) of the Act. Moreover, TradeARBED refused to provide the documents in question at verification, although it was given adequate notice that these documents would be reviewed. Accordingly, we find that Aceralia did not act to the best of its ability in reporting these expenses, and as a consequence we have based the amount of brokerage and handling expenses for these ports on adverse facts available, in accordance with section 776(b) of the Act. As adverse facts available, we have used the highest per-port amount on the record of this proceeding. (16) Nonetheless, because TradeARBED was able to provide adequate documentation for two of the ports in question, we have accepted the expenses calculated for those ports for purposes of the final determination. We disagree with the petitioners that these expenses should be rejected because they were not reported on a transaction-specific basis. As we noted in our verification report at page 13, we were able to obtain accurate information for these two ports. Therefore, we have used these expenses in our final margin calculations. Comment 12: Indirect Selling Expenses of Arbed Americas Aceralia reported U.S. indirect selling expenses based on the expenses incurred solely by its U.S. sales agent, TradeARBED. At verification, we found that an affiliate of TradeARBED (i.e., Arbed Americas Atlantic (AAA)) also had some involvement in the sale of subject merchandise during the POI. Specifically, we found that AAA performed certain financing activities on behalf of the Arbed Americas group of companies. Because none of AAA's expenses had been included in the reported U.S. indirect selling expenses, we calculated a revised indirect selling expense factor using the consolidated financial statements of the Arbed Americas group for consideration in the final determination. This rate is set forth in the U.S. sales verification report. Aceralia argues that the Department should not base indirect selling expenses incurred in the United States on the consolidated experience of the Arbed Americas group because certain of the companies in the group have no involvement in the sale of subject merchandise in the United States. Indeed, Aceralia asserts that only AAA and TradeARBED are involved in U.S. sales of beams. Therefore, while Aceralia does not disagree that a portion of AAA's expenses should be allocated to subject merchandise, it contends that it is inappropriate to include the expenses of the other affiliates in the U.S. indirect selling expense ratio. Instead, Aceralia argues that the Department should allocate AAA's expenses over the revenues of the companies to which AAA provides financing, and then add the resulting percentage to the reported U.S. indirect selling expense factor. Aceralia provided this revised indirect selling expense rate and notes that the information necessary to perform this calculation is contained in the Arbed Americas consolidated financial statements, which are on the record of this investigation. The petitioners contend that the Department should use the indirect selling expense ratio calculated at verification. The petitioners maintain that it is typical for a U.S. selling company to be involved in selling many products other than subject merchandise. Moreover, the petitioners contend that Aceralia's methodology understates U.S. indirect selling expenses because the numerator and the denominator of Aceralia's calculation are not stated on the same basis. Department's Position: In calculating indirect selling expenses, the Department's practice is to base these expenses, as closely as possible, on the experience of the companies which actually made the sales of subject merchandise. For this reason, the Department generally bases indirect selling expenses in CEP situations on the experience of the U.S. affiliate which makes the sale, rather than on the expenses at a higher level of consolidation. In addition, the Department also includes an allocated portion of the expenses incurred by affiliates (e.g., parent companies) that provide support activities to the U.S. sales agent. In this case, we included the indirect selling expenses of TradeARBED and an additional amount for the expenses incurred by AAA, because AAA performed certain functions related to TradeARBED's sales. We are unable to allocate AAA's expenses using its sales revenue, however, because the revenues reflected on its financial statements represent income from the sales of services. As such they cannot be accurately applied to U.S. sales of subject merchandise, because these revenues do not include U.S. sales prices (i.e., the base to which the indirect selling expense ratio would be applied). Therefore, we have accepted Aceralia's methodology for allocating AAA's indirect selling expenses, as set forth in its April 3, 2002, case brief at Exhibit 4. However, we adjusted Aceralia's methodology to take into account the recalculation of interest expenses incurred by AAA, as described below at Comment 13. We find that this methodology is reasonable in that it appropriately distributes the selling expenses incurred by AAA to the entities which benefit from its provision of financial services. Moreover, this methodology has been used by the Department in past proceedings. See, e.g., Extruded Rubber Thread From Malaysia; Final Results of Antidumping Duty Administrative Review, 64 FR 12967, 12976 (Mar. 16, 1998) (where the Department allocated indirect selling expenses to all sales to which the expenses applied, rather than merely computing a ratio using the total sales value of the respondent). Comment 13: Interest Expenses Included in U.S. Indirect Selling Expenses In the preliminary determination, we included the interest expenses incurred by TradeARBED as part of U.S. indirect selling expenses. We then offset these expenses by the amount of imputed credit expenses incurred on subject beams sold by TradeARBED (including those in the companion investigations on Germany and Luxembourg). We also included certain of these interest expenses as part of the financing expense ratio computed for COP. (17) Aceralia argues that it is not appropriate to include these expenses as part of both indirect selling expenses and COP, because doing so results in double counting. According to Aceralia, the Department's long-standing practice has been to classify interest expenses as part of COP, and the Department should continue to treat these expenses as part of COP here. Nonetheless, in the event that the Department continues to treat these expenses as part of indirect selling, Aceralia argues that the Department should: 1) not include the accounts receivable "securitization" expenses reflected in the ARBED consolidated financial statements as part of COP because these expenses are related to sales, rather than cost; and 2) recalculate the imputed credit offset for purposes of the final determination using the Arbed America's company-wide experience. Regarding this latter point, Aceralia notes that U.S. sales of beams account for only around five percent of the company's sales revenues; thus, Aceralia argues that limiting the offset to sales of beams unjustly overstates indirect selling expenses. The petitioners agree that these expenses should not be double counted. Therefore, the petitioners contend that these expenses should be classified as indirect selling expenses, rather than as part of COP. Department's Position: We agree with Aceralia that the interest and offset components of U.S. indirect selling expenses were not calculated appropriately for purposes of the preliminary determination. Based on a recalculation of these components, for this final determination we have not included an amount of interest expenses in indirect selling because the recalculated offset exceeds the amount of the U.S. interest expenses. Therefore, we need not address the issue of double counting for purposes of the final determination. Regarding the calculation of the offset, we agree with Aceralia that the imputed credit calculation used to offset interest expenses included in indirect selling expenses is not correct. As noted in Stainless Steel Plate in Coils From the Republic of Korea; Final Results of Antidumping Duty Administrative Review, 66 FR 64107 (Dec. 11, 2001) and accompanying decision memorandum at Comment 14 (Stainless Steel Plate from Korea), under appropriate circumstances, we may exclude some portion or all of a U.S. sales affiliate's interest expenses in the calculation of its indirect selling expenses. Furthermore, we explained that we must determine the appropriate universe of CEP deductions on a case-by-case basis because the activities of U.S. sales affiliates differ considerably across cases. We agree with Aceralia that our calculation of indirect selling expenses, which was preliminarily done on a TradeARBED-wide basis, took into account all of the indirect selling expenses of the company and allocated them over all of TradeARBED's sales in the United States. As a result, the offset for imputed credit was not calculated on the same basis as all of the other expenses included in indirect selling expenses, since the offset was limited to sales of structural steel beams from Spain and its two companion cases from Germany and Luxembourg, whereas all of the other expenses are reported on a company-wide basis. Therefore, we have adjusted TradeARBED's reported interest expenses to determine a net interest expense ratio applicable to U.S. sales of Spanish beams alone, thereby eliminating the double-counted amount attributed to imputed credit, and appropriately attributing the reported expenses to sales of subject merchandise. This method, by allocating a portion of the total interest expense to subject merchandise and disregarding the remainder, directly addresses Aceralia's concern about including interest expenses associated with non-subject merchandise in the calculation of indirect selling expenses. We note that, because AAA also performed certain financing activities on behalf of TradeARBED during the POI, we have also included AAA's financing and indirect selling expenses in our re-computation. Therefore, consistent with the methodology set forth in Stainless Steel Plate from Korea, to calculate indirect selling expenses, we separately calculated an adjustment for net interest expense applicable to sales of subject merchandise. Specifically, we calculated the amount of interest expense applicable to sales of subject merchandise by multiplying the reported TradeARBED- and AAA-wide interest expenses by the ratio of total U.S. sales of subject merchandise to the amount of total U.S. sales of all merchandise. We then deducted the total value of imputed credit expense reported on the computer sales listing from the amount of interest expense applicable to sales of subject merchandise. Based on this calculation, we found that the reported credit expenses on CEP sales of Spanish beams exceeded the reallocated amount of interest expenses, and thus we did not include any interest expenses in our deductions from U.S. price. Finally, regarding "securitization expenses," we have continued to exclude these expenses from the calculation of financing expenses used in COP. These expenses represent bank charges paid by group companies on sales transactions. Therefore, we find that these expenses are more appropriately classified as selling, rather than financing, expenses. Comment 14: Clerical Errors in the Preliminary Determination Aceralia claims that the Department made two clerical errors in the preliminary determination. First, Aceralia claims that the Department failed to use the correct conversion factor to restate the COP in the calculation of CEP profit from kilograms to hundredweight (i.e., the unit of measure of the gross unit price). Second, Aceralia contends that the Department failed to convert U.S. credit expenses from metric tons to hundredweight before calculating the foreign unit price in dollars (which was stated in hundredweight). The petitioners did not comment on this issue. Department's Position: We have reviewed our calculations and agree that we failed to make conversions in question accurately. Therefore, we have revised our calculations for purposes of the final determination to apply the correct conversion factors. Comment 15: Calculation of the Overall Dumping Margin Aceralia argues that the Department should not set negative dumping margins to zero when calculating the overall dumping margin. According to Aceralia, this practice violates section 773(a) of the Act, which requires the Department to make a fair comparison between export price or CEP and normal value. Specifically, Aceralia contends that an aggregate dumping margin developed by a calculation that zeroes out the margins of specific product groups within the calculation does not reflect a true weighted-average dumping margin, as required by section 771(35)(B) of the Act. Aceralia maintains that it is mathematically impossible to calculate an average without considering the actual values of specific numbers in the numerator where their corresponding values are included in the denominator. Rather, Aceralia contends that this methodology results in an artificial increase in the overall dumping margin because it does not capture the real average price at which subject merchandise is sold in the United States. Finally, Aceralia argues that the Department's practice of zeroing negative margins violates Article 2.4.2 of the World Trade Organization (WTO) Agreement on Implementation of Article VI of the General Agreement on Tariffs and Trade (the Agreement), and thus it also violates the international obligations of the United States. As support for this assertion, Aceralia cites a decision by the WTO in which the WTO held that zeroing of negative margins does not constitute a fair comparison. See WTO Appellate Body in European Communities Anti- Dumping Duties on Imports of Cotton-Type Bed Linen from India, at page 15 (Bed Linen). According to Aceralia, the WTO decision in Bed Linen constitutes an interpretation of the obligations of every WTO member country under the Agreement, and because this interpretation does not conflict with the express language in U.S. antidumping law, it must be followed here. The petitioners contend that the Department's practice of zeroing negative margins is not only correct, but it is also mandated by U.S. antidumping law. Specifically, the petitioners note that the antidumping law requires the Department to determine an estimated weight-average dumping margin, which is defined under section 771(35)(B) of the Act as the percentage determined by dividing the aggregate dumping margins determined for a specific exporter or producer by the aggregate export prices and constructed export prices of such exporter or producer. The petitioners further note that the section 771(35)(A) of the Act defines the term "dumping margin" as the amount by which the normal value exceeds the export price or constructed export price of the subject merchandise. Thus, the petitioners conclude that the antidumping law expressly prohibits the Department from either calculating a negative dumping margin or including negative margins in the calculation of the weighted-average dumping margin. The petitioners point out that Aceralia's approach has been squarely rejected in several recent determinations. See, e.g., Notice of Final Determination of Sales at Less Than Fair Value: Certain Softwood Lumber Products from Canada, 67 FR 15539 (Apr. 2, 2002) and accompanying decision memorandum at Comment 12; Final Determination of Sales at Less Than Fair Value: Automotive Replacement Glass Windshields From The People's Republic of China, 67 FR 6482 (Feb. 12, 2002) and accompanying decision memorandum at Comment 34 (Windshields from the PRC); and Notice of Final Determination of Sales at Less Than Fair Value; Certain Hot-Rolled Carbon Steel Flat Products From The Netherlands, 66 FR 50408 (Oct. 4, 2001) and accompanying decision memorandum at Comment 1 (Hot-Rolled Steel from the Netherlands). Finally, the petitioners maintain that the WTO decision in Bed Linen has no impact on U.S. law or the Department's practice. The petitioners note that the SAA explicitly states that reports issued by the WTO appellate body have no binding effect under U.S. law, nor do they provide legal authority for federal agencies to change their regulations or procedures. See the SAA at 1032. Thus, the petitioners contend that the Bed Linen decision applies only to the European Community, and consequently it provides no basis for changing either U.S. law or practice. Department's Position: We disagree with Aceralia. As we have discussed in prior cases, our methodology is consistent with our statutory obligations under the Act. See, e.g., Hot-Rolled Steel from the Netherlands at Comment 1 and Windshields from the PRC at Comment 34. First, sales that did not fall below normal value are included in the weighted-average margin calculation as sales with no dumping margin. The value of such sales is included in the denominator of the weighted-average margin along with the value of dumped sales. We do not, however, allow sales that did not fall below normal value to cancel out dumping found on other sales. Second, the Act requires that the Department employ this methodology. Section 771(35)(A) of the Act defines "dumping margin" as "the amount by which the normal value exceeds the export price or constructed export price of the subject merchandise." Section 771(35)(B) of the Act defines "weighted-average dumping margin" as "the percentage determined by dividing the aggregate dumping margins determined for a specific exporter or producer by the aggregate export prices and constructed export prices of such exporter or producer." These sections, taken together, direct the Department to aggregate all individual dumping margins, each of which is determined by the amount by which normal value exceeds export price or constructed export price, and to divide this amount by the value of all sales. The directive to determine the "aggregate dumping margins" in section 771(35)(B) makes clear that the singular "dumping margin" in section 771(35)(A) applies on a comparison- specific level and does not itself apply on an aggregate basis. At no stage in this process is the amount by which export price or CEP exceeds normal value on sales that did not fall below normal value permitted to cancel out the dumping margins found on other sales. This does not mean, however, that sales that did not fall below normal value are ignored in calculating the weighted-average rate. It is important to note that the weighted-average margin will reflect any "non-dumped" merchandise examined during the investigation because the value of such sales is included in the denominator of the dumping rate, while no dumping amount for "non- dumped" merchandise is included in the numerator. Thus, a greater amount of "non-dumped" merchandise results in a lower weighted-average margin. This is, furthermore, a reasonable means of establishing duty deposits in investigations, and assessing duties in reviews. In an investigation such as the present case, the deposit rate calculated must reflect the fact that the Customs Service is not in a position to know which entries of merchandise entered after the imposition of a dumping order are dumped and which are not. By spreading the estimated liability for dumped sales across all investigated sales, the weighted-average dumping margin allows the Customs Service to apply this rate to all merchandise entered after an order goes into effect. Finally, with respect to the respondent's WTO-specific arguments, we believe U.S. law is consistent with our WTO obligations. Moreover, the WTO decision in Bed Linen concerned a dispute between the European Union and India. Consequently, we have no obligation under U.S. law to act on this decision. Comment 16: Calculation of Raw Materials Costs The petitioners claim that Aceralia did not properly account for its beam blank inventory in the reported costs. The petitioners state that Aceralia should not have used its POI consumption of beam blanks but instead should have used the cost of beam blanks produced during the POI. Aceralia argues that it properly reported the cost of beam blanks used to produce the subject merchandise during the POI. Department's Position: Beam blanks are an intermediate product (i.e., work-in-process) which is used as an input in producing beams, the finished product. The blanks consumed in production are valued at the inventory requisition value in Aceralia's normal books and records in determining the cost of manufacturing beams during the year. This treatment follows Aceralia's normal books and records, in accordance with its home country generally accepted accounting principles. We note that the Department uses current costing in cases where the country in which the respondent operates experienced high inflation during the period. This was not the case for Aceralia during the POI. Therefore, we have relied on Aceralia's reported costs for our final determination. Comment 17: Exchange Gains and Losses In the preliminary determination, we included exchange gains and losses in the G&A rate calculation for Aceralia's cost of production. Aceralia argues that the Department should not include exchange gains and losses in both the G&A and the financial rate calculations in order to avoid double counting these amounts. The petitioners did not comment on this issue. Department's Position: We agree with Aceralia that exchange gains and losses should not be double counted. We note that the amounts for exchange gains and losses that the Department has included in the G&A calculation for the final determination were generated by Aceralia's accounts payable only and thus were picked up at the company level. In addition, since we noted no exchange gains or losses associated with financing activities, we did not include any additional exchange gains or losses in the financial expense rate calculation at the consolidated level. Comment 18: Acceptance of Revised Sales Databases On April 1, 2002, Aceralia submitted revised sales databases at the request of the Department. These databases incorporated certain changes arising from verification, as well as various revisions to the databases provided in hard copy form prior to the deadline for submission of new factual information. See the April 9, 2002, memorandum from Shawn Thompson to the file entitled "Telephone Conversation with Counsel for the Petitioners in the Less Than Fair Value Investigation on Structural Steel Beams from Spain." On April 18, 2002, the petitioners filed a letter objecting to the acceptance of the revised databases. In this letter, the petitioners assert that the databases: 1) do not completely capture all changes arising from verification; and 2) contain new information related to home market rebates. Regarding this latter point, the petitioners concede that the source documentation for the majority of these changes was filed in a submission in January 2002, prior to verification. However, the petitioners assert that this information appears to be unverified because it is not mentioned in the home market sales verification report, and there is no evidence on the record that the Department requested it to be incorporated in the home market sales listing. Moreover, the petitioners note that the information in the January submission cannot be easily linked to the home market sales listing, and thus it is difficult to check that the revisions have been made accurately. Finally, the petitioners assert that the revised sales listing contains revisions to numerous transactions which do not appear to relate to the January submission. For these reasons, the petitioners argue that the Department should reject the new databases. On April 26, 2001, Aceralia responded to the petitioners' submission. In its letter, Aceralia notes that the Department did not request that all changes outlined in the verification be made, nor did Aceralia ever claim that they were. Rather, Aceralia asserts that it only revised its sales listings to incorporate specific revisions determined through detailed discussions with the Department. Furthermore, Aceralia notes that the home market sales listing does not contain any new information. Specifically, Aceralia maintains that each of the revisions disputed by the petitioners can be linked to data on the record in either its January submission or the home market sales verification report. Aceralia points out that these revisions were also explicitly detailed in the narrative accompanying the revised databases. Therefore, Aceralia contends that the Department should reject the petitioners' arguments for purposes of the final determination. Department's Position: We have accepted these databases for use in the final determination. Regarding the petitioners' first point (i.e., that the databases should be rejected on the grounds that they do not completely incorporate all changes arising from verification), we find that Aceralia provided all changes requested by the Department (18), and it clearly indicated which revisions it made and (by omission from the narrative accompanying its submission) which revisions it did not. Because the remaining changes are easily made in our final margin calculations, we have incorporated them into our final analysis. In any event, we note that these databases were requested solely as a matter of convenience for the Department. Because they contain revisions that we otherwise would make to Aceralia's databases for the final determination, it is irrelevant to the final outcome of this case whether the revisions were made by Aceralia or by the Department itself. Moreover, we disagree with the petitioners that the revised databases contain new factual information in general. Instead, these databases are merely updated versions of Aceralia's most recent sales listings which reflect a number of minor corrections found during verification or contained in various pre-verification submissions. Each of the changes was clearly identified in the letter accompanying the databases and can be traced to information on the record on this proceeding as early as December of last year. In addition, each of these changes (including those related to rebates) was specifically requested by the Department. We also disagree that the home market sales listing specifically contains new rebate information. We have carefully examined the information submitted by Aceralia and find that it is entirely consistent with the information set out in the home market sales verification report and in the company's January 18, 2002, rebate submission. See the May 13, 2002, memorandum to the file from Patrick Connolly entitled "Calculations Performed for Aceralia Corporacion Siderurgica, S.A. (Aceralia) for the Final Determination in the Less Than Fair Value Investigation of Structural Steel Beams from Spain." We further disagree with the petitioners that this information cannot be easily linked to the data on the record, and we have included the programming language necessary to make a proper comparison of the old and new data in our memorandum to the file on this topic. We similarly disagree with the petitioners' contention that the rebate information in question is unverified. We note that Aceralia's January 18 submission is referenced on page 1 of the home market sales verification report. In addition, while this submission is not explicitly discussed in the section of the report related to rebates, the Department's silence on this matter cannot be interpreted to mean that the amounts were not verified. It is not the Department's practice to verify every component of every expense reported by a given respondent. Rather, the Department's standard practice is to conduct verification on a sample basis and to impute the results of verification findings to all items not specifically reviewed. (19) Because we find that Aceralia's data, as a whole, could be verified, we have accepted the January 18 rebate data for purposes of our final analysis. RECOMMENDATION Based on our analysis of the comments received, we recommend adopting all of the above positions and adjusting all related margin calculations accordingly. If these recommendations are accepted, we will publish the final determination in the Federal Register. Agree ______ Disagree ______ ______________________ Faryar Shirzad Assistant Secretary for Import Administration ______________________ (Date) _________________________________________________________________________ footnotes: 1. In other words, we compare the cumulative selling functions performed by both the producer and reseller in selling to the reseller's customers with the selling functions performed by the producer in selling to its customers. 2. Due to a claim for proprietary treatment, we are unable to discuss these selling functions here. However, they are set forth in footnotes 3 through 7 of Aceralia's January 9, 2002, LOT supplemental response. 3. Aceralia notes that it had previously argued that a CEP offset was appropriate. However, Aceralia maintains that additional evidence was placed on the record after the preliminary determination and yet more data was gathered during the company's preparations for verification on this topic. According to Aceralia, the Department now has a fully developed record which shows that U.S. and home market direct sales are at the same LOT; thus a CEP offset is no longer warranted. 4. In making this assertion, the petitioners misread a portion of the LOT supplemental questionnaire in this case, which states that the Department does not consider selling functions related to expenses, such as commissions, deducted under section 772(d) of the Act, when making LOT determinations. 5. AHT was a trading company located in Germany. This company was closed in June 2001, three months after the end of the POI. 6. For example, Aceralia was asked to supplement its selling functions information for each affiliated party for which it reported downstream sales, and to describe any differences in the sales process for sales made through AHT and TradeARBED. Aceralia provided information on its home market affiliated reseller in response to the former question and was silent as to any differences in response to the latter. See Aceralia's November 8, 2001, response at page 16. 7. Indeed, we have used the term "Velasco" to refer to the collective group of affiliated companies, when in fact we have never considered them to be a single entity. 8. See pages 19 through 21 of the home market sales verification report, in which we indicated that we found no discrepancies in tying specific rebates to specific invoices. 9. We are unable to discuss the specifics of this rebate here due to Aceralia's claim for business proprietary treatment. For further discussion, see the first rebate referenced on page 20 of the home market sales verification report. 10. As a result, a wide flange beam with an 8 inch section depth of 58 lbs. per foot would be treated as the same product as a wide flange beam with a 12 inch section depth of 53 lbs. per foot, all other things being equal. 11. The properties explicitly identified for yield strengths in the same range are: 1) for atmospheric corrosion resistance (i.e., code "41"); and 2) heat treated, not for atmospheric corrosion (i.e., code "42"). 12. At verification, Aceralia demonstrated that it completely accounted for all of the information recorded in its sales account by providing a worksheet which reconciled the sales and rebate information in the accounting system to the sales and rebate information reported in the home market sales listing. We found that Aceralia accurately reported each component of this calculation during other portions of the verification. For a full discussion of the procedures performed on this topic, see the home market sales verification report at pages 21 and 22. 13. In any event, Aceralia explained at verification that this particular rebate program was standard, and we found no evidence to contradict this assertion. See the home market sales verification report at pages 22 and 23. Moreover, because the rebate agreement was verbal, no documentation existed to prove the rebate schedule applicable to the customer group in question. Finally, we confirmed that Aceralia was unable to track the rebates paid under specific credit notes because of the limitations of its record keeping procedures (i.e., in various instances, the credit notes did not reference specific rebate programs or time periods). Nonetheless, as noted above, we established that the rebates of this type during the POI tied to Aceralia's accounting system in the aggregate, and we confirmed that they were allocated accurately to the customer group in question by tying the sales quantity for the group to Aceralia's accounting system. Thus, we have continued to allow these rebates for purposes of the final determination. 14. The number of home market shipments referenced here differs from that cited by the petitioners because: 1) Aceralia was unable to locate the documentation for one shipment; and 2) two transactions examined at verification were shipped under the same invoice. 15. Specifically, the verification outline requested that TradeARBED provide source documentation for its port expenses related to each of the sales which were pre-selected for verification. See the letter to Kermit W. Almstedt of O'Melveny & Myers from Shawn Thompson, dated February 27, 2002, at page 10. The sales selected for Spain were shipped through two U.S. ports, neither of which was Mobile or San Francisco. Similarly, the sales selected in the companion cases on structural steel beams from Luxembourg and Germany were shipped through additional (i.e., non- examined) U.S. ports. These sales were also identified in the Department's February 27 verification outline. Finally, the verification outline also clearly stated on page 1 that "the Department will perform procedures and request documents in addition to those in the outline." 16. We further note that, because TradeARBED reported the same per-port expenses in the companion investigations involving Germany and Luxembourg, and because the facts are identical with respect to those investigations as well (i.e., TradeARBED is the U.S. sales agent for all three proceedings and, consequently, the U.S. sales verifications for all three cases were combined), we have also based U.S. brokerage and handling expenses on adverse facts available for the investigations involving Germany and Luxembourg. 17. Specifically, we based the financing expenses used in the calculation of COP on the data shown on the financial statements of the Arbed Americas group, in accordance with our practice of calculating these expenses using audited financial statements at the respondent's highest level of consolidation. See, e.g., Frozen Concentrated Orange Juice from Brazil; Final Results and Partial Rescission of Antidumping Duty Administrative Review, 66 FR 51108 (Oct. 5, 2001) and accompanying decision memorandum at Comment 2. 18. In their April 18 letter, the petitioners assert that the Department "ordered" Aceralia to correct all errors identified at verification. This assertion is not correct. Rather, we requested that Aceralia make the changes listed in the narrative portion of its April 1 submission. 19. For example, we did not examine billing adjustments or early payment discounts during the verification conducted at Aceralia's affiliated reseller, Velasco. Even so, we have accepted this expense as reported based on our finding that the reseller's data, as a whole, was reported accurately. We note that the petitioners not only do not dispute this finding regarding Velasco's data, they urge the Department to accept it for purposes of the final determination. See Comment 1, above.