[Federal Register Volume 63, Number 10 (Thursday, January 15, 1998)] [Notices] [Pages 2558-2585] From the Federal Register Online via the Government Publishing Office [www.gpo.gov] [FR Doc No: 98-944] [[Page 2557]] _______________________________________________________________________ Part III Department of Commerce _______________________________________________________________________ International Trade Administration _______________________________________________________________________ Tapered Roller Bearings and Parts Thereof, Finished and Unfinished From Japan, etc.; Notice Federal Register / Vol. 63, No. 10 / Thursday, January 15, 1998 / Notices [[Page 2558]] DEPARTMENT OF COMMERCE International Trade Administration [A-588-054, A-588-604] Tapered Roller Bearings and Parts Thereof, Finished and Unfinished, From Japan, and Tapered Roller Bearings, Four Inches or Less in Outside Diameter, and Components Thereof, From Japan; Final Results of Antidumping Duty Administrative Reviews AGENCY: Import Administration, International Trade Administration, Department of Commerce. ACTION: Notice of Final Results of Administrative Reviews. ----------------------------------------------------------------------- SUMMARY: On September 9, 1997, the Department of Commerce (the Department) published the preliminary results of the 1995-96 administrative reviews of the antidumping duty order on tapered roller bearings (TRBs) and parts thereof, finished and unfinished, from Japan (A-588-604), and the antidumping finding on TRBs, four inches or less in outside diameter, and components thereof, from Japan (A-588-054) (see Tapered Roller Bearings and Parts Thereof, Finished and Unfinished, from Japan, and Tapered Roller Bearings, Four Inches or Less in Outside Diameter, and Components Thereof, from Japan; Preliminary Results of Antidumping Duty Administrative Reviews, 62 FR 47452 (September 9, 1997) (TRB Prelim)). The review of the A-588-054 finding covers two manufacturers/exporters and two resellers/exporters of the subject merchandise to the United States during the period October 1, 1995, through September 30, 1996. The review of the A-588- 604 order covers three manufacturers/exporters, two resellers/ exporters, and the period October 1, 1995, through September 30, 1996. We gave interested parties an opportunity to comment on our preliminary results. Based upon our analysis of the comments received we have changed the results from those presented in our preliminary results of review. EFFECTIVE DATE: January 15, 1998. FOR FURTHER INFORMATION CONTACT: Charles Ranado, Stephanie Arthur, or Valerie Owenby, Office of AD/CVD Enforcement III, Office 8, Import Administration, International Trade Administration, U.S. Department of Commerce, 14th Street and Constitution Avenue, N.W., Washington, D.C. 20230, telephone: (202) 482-3518, 6312, or 0172, respectively. SUPPLEMENTARY INFORMATION: Applicable Statute and Regulations Unless otherwise indicated, all citations to the statute are in reference to the provisions effective January 1, 1995, the effective date of the amendments made to the Tariff Act of 1930 (the Act) by the Uruguay Round Agreements Act (URAA). In addition, unless otherwise indicated, all citations are to the Department's regulations, 19 CFR part 353 (1997). Background On August 18, 1976, the Treasury Department published in the Federal Register (41 FR 34974) the antidumping finding on TRBs from Japan, and on October 6, 1987, the Department published the antidumping duty order on TRBs from Japan (52 FR 37352). On October 1, 1996 (61 FR 51529), the Department published the notice of ``Opportunity to Request an Administrative Review'' for both TRB cases. The petitioner, the Timken Company (Timken), and one respondent requested administrative reviews. We initiated the A-588-054 and A-588-604 administrative reviews for the period October 1, 1995, through September 30, 1996, on November 15, 1996 (61 FR 58513). On September 9, 1997, we published in the Federal Register the preliminary results of the 1995-96 administrative reviews of the antidumping duty order and finding on TRBs from Japan (see TRB Prelim at 47542). We held a hearing for the 1995-96 administrative reviews of both the A-588-054 and A-588-604 TRBs cases on October 30, 1997. The Department has now completed these reviews in accordance with section 751 of the Act, as amended. Scope of the Review Imports covered by the A-588-054 finding are sales or entries of TRBs, four inches or less in outside diameter when assembled, including inner race or cone assemblies and outer races or cups, sold either as a unit or separately. This merchandise is classified under the Harmonized Tariff Schedule (HTS) item numbers 8482.20.00 and 8482.99.30. Imports covered by the A-588-604 order include TRBs and parts thereof, finished and unfinished, which are flange, take-up cartridge, and hanger units incorporating TRBs, and tapered roller housings (except pillow blocks) incorporating tapered rollers, with or without spindles, whether or not for automotive use. Products subject to the A-588-054 finding are not included within the scope of this order, except for those manufactured by NTN Corporation (NTN). This merchandise is currently classifiable under HTS item numbers 8482.99.30, 8483.20.40, 8482.20.20, 8483.20.80, 8482.91.00, 8484.30.80, 8483.90.20, 8483.90.30, and 8483.90.60. These HTS item numbers and those for the A-588-054 finding are provided for convenience and Customs purposes. The written description remains dispositive. The A-588-054 reviews cover TRB sales by two TRB manufacturers/ exporters (Koyo Seiko Ltd. (Koyo) and NSK Ltd. (NSK)), and two resellers/exporters (Fuji Heavy Industries (Fuji) and MC International (MC)). The reviews of the A-588-604 case cover TRB sales by three manufacturers/exporters (Koyo, NSK and NTN Corporation (NTN)), and two resellers/exporters (Fuji and MC). Because Fuji and MC had no shipments in the A-588-604 review, for the reasons explained in our notice of preliminary results, we have not assigned a rate to these firms for these final results (see TRB Prelim at 47453). The period of review (POR) for both cases is October 1, 1995, through September 30, 1996. Duty Absorption On December 11, 1996, Timken requested that the Department determine, with respect to all respondents, whether antidumping duties had been absorbed during the POR. Section 751(a)(4) of the Act provides for the Department, if requested, to determine during an administrative review initiated two or four years after the publication of the order, whether antidumping duties have been absorbed by a foreign producer or exporter if the subject merchandise is sold in the United States through an affiliated importer. The Department's interim regulations do not address this provision of the Act. For transition orders as defined in section 751(c)(6)(C) of the Act (i.e., orders in effect as of January 1, 1995), section 351.213(j)(2) of the Department's new antidumping regulations provide that the Department will make a duty-absorption determination, if requested, for any administrative review initiated in 1996 or 1998. See 62 FR 27394 (May 19, 1997). Because the finding and order on TRBs have been in effect since 1976 and 1987, respectively, they are transition orders in accordance with section 751(c)(6)(C) of the Act. The preamble to the new antidumping regulations explains that reviews initiated in 1996 will be considered initiated in the second year and reviews initiated in 1998 will be considered initiated in the fourth year (62 FR 27317, May 19, 1997). This approach [[Page 2559]] ensures that interested parties will have the opportunity to request a duty-absorption determination prior to the time for sunset review of the order under section 751(c) of the Act on entries for which the second and fourth years following an order have already passed. Since these reviews were initiated in 1996, and a request was made for a determination, we are making duty-absorption determinations as part of these administrative reviews. As indicated above, the statute provides for a determination on duty absorption if the subject merchandise is sold in the United States through an affiliated importer. In these cases, NTN, Koyo, NSK, and Fuji sold through importers that are affiliated within the meaning of section 751(a)(4) of the Act. We have determined that duty absorption has occurred with respect to the following firms and with respect to the following percentages of sales made through their U.S. affiliates: ------------------------------------------------------------------------ Percentage of U.S. affiliates' Manufacturer/exporter/reseller sales with dumping margins ------------------------------------------------------------------------ For the A-588-054 Case: Koyo Seiko............................................... 12.99 Fuji..................................................... 4.54 NSK...................................................... 13.30 For the A-588-604 Case: Koyo Seiko............................................... 98.10 Fuji \1\ NSK...................................................... 51.78 NTN...................................................... 66.36 ------------------------------------------------------------------------ \1\ No shipments or sales subject to this review. In the case of Koyo, the firm did not respond to our request for further-manufacturing information and we determined the dumping margins for these further-manufactured sales on the basis of adverse facts available. Lacking other information, we find duty absorption on all such sales of further-processed TRBs (see Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof from France, et. al.; Preliminary Results of Antidumping Administrative Review, 62 FR 31568 (June 10, 1997) (where we found duty absorption with respect to all sales for which respondent provided no data in response to the Department's questionnaire)). With respect to other respondents with affiliated importers (NSK, NTN, and Fuji), for which we did not apply adverse facts available, we must presume that the duties will be absorbed for those sales which were dumped. Where Koyo's margins were not determined on the basis of adverse facts available (i.e., for non-further-manufactured sales), we must presume that duties will be absorbed for those sales which were dumped. Our duty-absorption presumptions can be rebutted with evidence that the unaffiliated purchasers in the United States will pay the ultimately assessed duty. After publication of our preliminary results, we gave interested parties the opportunity to submit evidence that the unaffiliated purchasers in the United States will pay the ultimately assessed duties. However, we received no such evidence. Under these circumstances, we find that antidumping duties have been absorbed by Koyo, NTN, NSK, and Fuji on the percentages of U.S. sales indicated. Specific arguments relating to duty absorption are discussed in the ``Miscellaneous'' section below. Analysis of Comments Received We received case briefs from Koyo, NSK, NTN, and Timken on October 16, 1997. We received rebuttal briefs from the same four parties, as well as from Fuji, on October 23, 1997. The comments which were contained in all of the case and rebuttal briefs we received are addressed below in the following order: 1. Facts Available/Further Manufacturing 2. Adjustments to Normal Value 3. Adjustments to United States Price 4. Cost of Production and Constructed Value 5. Miscellaneous Comments Related to Duty Absorption, Level of Trade, the Arm's-Length Test, and Sample Sales 6. Clerical Errors 1. Facts Available/Further Manufacturing Comment 1: Koyo argues that the Department's application of adverse facts available with respect to its sales of A-588-604 further- manufactured TRBs was inappropriate and contrary to law for the following reasons. First, Koyo states, the Department apparently decided to require further-manufacturing data because there was an insufficient quantity of sales of imported finished A-588-604 TRBs to serve as a surrogate in accordance with section 773(e) (1) and (e)(2) of the Act, the special rule provision for further-processed merchandise. Koyo asserts that the Department presumably reached its conclusion by comparing the entered and sales values of imported finished over 4'' TRBs to the entered and sales values of over 4'' TRB components which were further manufactured. However, Koyo contends, the Department's comparison of over 4'' finished TRBs to over 4'' further- manufactured TRBs was flawed because the division of products subject to the TRB orders does not reflect commercial reality, but rather arose from the manner in which Timken defined merchandise covered by the petitions in the A-588-054 and A-588-604 TRB cases. Koyo further argues, in support of its assertions that the division between the orders is asymmetrical, that the over 4'' A-588-604 order covers components of TRBs which are typically further processed into under 4'' bearings. Second, Koyo contends that the statute does not grant the Department discretion to decide whether or not to invoke section 772(e)(1) and (e)(2) of the Act. Rather, Koyo argues, the Department is required to apply the special rule provided the respondent has demonstrated that the value added in the United States is likely to substantially exceed the value of the imported merchandise. Koyo contends that, given that the Department determined in its preliminary results that Koyo satisfied the ``substantially exceeds'' requirement, the Department was compelled by the statute to apply the special rule provision and determine the constructed export price (CEP) for A-588- 604 further-manufactured TRB sales using the price of either identical or other subject merchandise or, if it determined that there was not a sufficient quantity of sales using these two proxies, any other ``reasonable'' basis. Koyo argues that the Department cannot, as it did in its preliminary results, simply reject the special rule in its entirety simply because it determines there is an insufficient quantity of sales of identical or other subject merchandise. Rather, contends Koyo, the Department must calculate CEP using ``any other reasonable basis,'' as directed by the statute. Koyo proposes that, in light of the similarities of merchandise subject to the A-588-054 and A-588-604 orders, instead of evaluating whether the margins of finished over 4'' A-588-604 bearings were an appropriate proxy for further-manufactured merchandise, the Department should have relied on the margins of finished under 4'' A-588-054 TRBs as a surrogate for those over 4'' components which were further processed into under 4'' TRBs, and the margins of imported finished over 4'' A-588-604 TRBs as a surrogate for those over 4'' components which were further manufactured into over 4'' TRBs. In fact, Koyo argues, not only is such an approach another reasonable basis, but it adheres to the statutory preference for relying on the price of ``other'' or [[Page 2560]] ``identical'' subject merchandise. Koyo maintains that while the Department may be hesitant to use a margin from a different order or finding (A-588-054) and apply it to further-manufactured products subject to a different order (A-588-604), its concerns are not legally relevant for the special rule provision. Koyo contends that the statute contains no language suggesting that ``crossing orders'' would constitute an unreasonable basis for comparison, and that the Department's refusal to look beyond the confines of the over 4'' order is not consistent with the statue. Koyo also adds that the Department's failure to acknowledge the existence of the under 4'' finding cedes too much control to petitioners who might be encouraged to write petitions that create anomalous outcomes, as is the case, Koyo asserts, with respect to the division between the two TRB orders. Furthermore, Koyo argues, using the calculated margins for under 4'' finished TRBs as a proxy for that merchandise subject to the A-588- 604 order which was further manufactured is appropriate because of the physical similarities of the merchandise. It does not matter, argues Koyo, that each of these categories of merchandise is not subject to the same order because, as indicated above, the asymmetric division of the orders arose from historical happenstance. Koyo also asserts that, if the Department were to adopt such an approach, there would be a sufficient quantity of U.S. sales of imported finished A-588-054 TRBs to serve as a proxy for further-manufactured A-588-604 TRBs. Koyo also suggests that, as an alternative, the Department can compare the value of finished bearings subject to both the A-588-054 finding and A-588-604 order to the value of further-manufactured bearings subject to the A-588-604 order. This method would guarantee, Koyo argues, a sufficient quantity of sales to serve as a proxy and would be reasonable and appropriate. Having then determined a sufficient quantity test, Koyo suggests, the Department should weight- average the margins calculated for A-588-054 and A-588-604 finished bearings and apply the result to the A-588-604 further-processed components. Like Koyo, NSK argues that the Department must, in accordance with section 772(e) of the Act, apply the special rule for further-processed merchandise once it has determined that the value added in the United States is likely to substantially exceed the value of the imported product. NSK maintains that this is the only statutory requirement for the application of the special rule provision, and once this requirement has been met, the Department is mandated by the statue to apply the special rule for further-manufactured merchandise. NSK contends that the determination of whether such sales are appropriate or whether there is a sufficient quantity to provide a reasonable basis for comparison relates only to the two proxies for calculating CEP set forth in section 772(e)(1) and (e)(2) of the statute. NSK further argues that if the Department determines that neither of these two proxies can be used, it may calculate CEP on ``any other reasonable basis.'' However, NSK argues, once it finds that certain merchandise qualifies for the special rule under the condition set forth above, the Department cannot calculate CEP using the section 772(d)(2) standard methodology. Specifically, NSK maintains that reverting to this methodology is contrary to the language of the statute and that this method does not constitute another ``reasonable basis'' by which to calculate CEP because section 772(d)(2) of the Act provides that CEP be reduced by ``the cost of any further manufacture or assembly * * * except in circumstances described in subsection (e) of this section.'' Therefore, NSK asserts, the Department cannot reduce CEP by the cost of any further manufacturing within the realm of section 772(e). NSK argues that the Statement of Administrative Action (SAA) at 826 supports its assertions that the Department cannot apply the standard 772(d)(2) methodology once the condition for the special rule has been satisfied because it purposefully omits the standard methodology as an alternative to the two surrogates identified in section 772(e)(1) and (e)(2). In addition, NSK claims, the SAA makes every attempt to keep section 772(e) simple and a resort to the standard methodology is contrary to the intent of the special rule to reduce the burden of a further-manufacturing analysis for the Department. Finally, NSK argues that the Department's interpretation of section 772(e) is contrary to the objective of establishing a ``bright-line standard'' which allows the Department to inform respondents early during a review proceeding whether or not they must supply detailed further-manufacturing information. Because the Department reversed its decision regarding Koyo's further manufacturing months after it initially concluded this data would not be required, the Department, NSK argues, has defeated its stated objective of informing respondents if it will require further-processing data early in the review. Timken responds that the Department's application of adverse facts available with respect to Koyo's further-manufactured merchandise was supported by the record and in accordance with the law. Timken argues that the statute grants the Department broad discretion in the implementation of the special rule, and asserts that the Department's actions were consistent with the statute when it determined that Koyo's A-588-604 non-further-manufactured TRBs were an inappropriate proxy for Koyo's A-588-604 further-manufactured merchandise. Timken argues that the Department's use of the standard section 772(d)(2) methodology to calculate CEP for Koyo's further-manufactured sales constituted another ``reasonable basis,'' and that the information on the record did not provide any other ``reasonable basis'' for determining margins for further-manufactured sales. Department's Position: We disagree with respondents. The statute at section 772(e) provides that: Where the subject merchandise is imported by a person affiliated with the exporter or producer, and the value added in the United States by the affiliated person is likely to exceed substantially the value of the subject merchandise, the administering authority shall determine the constructed export price for such merchandise by using one of the following prices if there is a sufficient quantity of sales to provide a reasonable basis for comparison and the administering authority determines that the use of such sales is appropriate: (1) The price of identical subject merchandise sold by the exporter or producer to an unaffiliated person; (2) The price of other subject merchandise sold by the exporter or producer to an unaffiliated person. If there is not a sufficient quantity of sales to provide a reasonable basis for comparison under paragraph (1) or (2), or the administering authority determines that neither of the prices described in such paragraphs is appropriate, then the constructed export price may be determined on any other reasonable basis. Koyo asserts that the Department, having determined that the value added in the United States is likely to substantially exceed the value of the imported merchandise, is required to apply the special rule provision of the statute, and argues that the use of the word ``shall'' in the provision clearly demonstrates that the Department does not have discretion as to when to invoke the special rule. Koyo and NSK incorrectly assume that we rejected the provision entirely. In addition, Koyo incorrectly argues that we imposed an additional qualification [[Page 2561]] for the application of the special rule; namely, that we required Koyo not only to demonstrate that the value added to its further- manufactured subject merchandise substantially exceeded the value of the subject merchandise as entered but also to demonstrate that there was a sufficient quantity of its non-further-manufactured sales to serve as a proxy for the calculation of CEP for its further- manufactured merchandise. To the contrary, our decision to request further-processing data from Koyo was made within the confines of and according to the language of the special rule. The SAA provides that the special rule provision will come into play when it is estimated that the value added in the United States is substantially more than half of the price charged to the first unaffiliated purchaser of the finished merchandise (see SAA at 825-826). After a determination that the value added is likely to substantially exceed the value of the imported components, the statute specifies that the use of the options identified in section 772 (e)(1) and (e)(2) is contingent upon the existence of a sufficient quantity of sales to provide a reasonable basis for comparison and that the use of such sales is appropriate. In other words, even if the quantity of the proxy sales is sufficient, we will reject their use unless we determine that using them is appropriate. In determining whether the use of either of the two proxy methods is appropriate, the Department looks to the underlying purpose of the special rule, which is to avoid imposing an unnecessary burden on the Department, while still ensuring reasonably accurate results (see SAA at 825-826). As part of this determination, we consider such factors as whether their use may lead to inaccurate results. We believe that the greater the proportion of further-manufactured to non-further- manufactured merchandise, the greater the possibility of inaccurate results. If there is a concern about accuracy, we must consider whether an alternative method, especially the standard methodology, would be unduly burdensome. The burden of applying the standard methodology to calculate the CEP for further-manufactured merchandise may vary from case to case depending on factors such as the nature of the further- manufacturing process and the finished products. The accuracy gained by applying the standard methodology may also vary significantly from case to case, depending upon such factors as the amount of value added in the United States and the proportion of U.S. sales which undergo further processing. Where the burden of performing a further- manufacturing analysis is high, we may determine that the potential gains in accuracy do not outweigh the burden of applying the standard section 772(d)(2) methodology and that the use of one of the statutory alternatives set forth in 772 (e)(1) and (e)(2) is appropriate. However, if the burden is relatively low and the proportion of further- manufactured sales is sufficiently high to raise concerns about accuracy, we may consider use of the statutory alternatives inappropriate. In the instant case, the record does not lead us to conclude that the use of either of the two alternative methods described in section 772 (e)(1) and (e)(2) with respect to Koyo's further-manufactured merchandise is appropriate. The record indicates that Koyo's U.S. sales of further-manufactured subject merchandise represented a large portion of its total U.S. sales of subject merchandise during the POR. Therefore, the use of either of the proxy methods in this case--where the proportion of further-manufactured sales is relatively high--would have a relatively high potential for inaccuracy. In addition, as noted in our preliminary results, the finished merchandise sold by Koyo to the first unrelated U.S. customer was still in the same class or kind as merchandise within the scope of the TRB order and finding (i.e., imported TRB components were processed into TRBs). As a result, the calculation of the precise amount of cost of further manufacturing would not be nearly as burdensome as it would be for Fuji, another respondent who imported TRBs for incorporation in automobiles. Furthermore, in prior reviews we have calculated margins for Koyo's further-processed sales and have extensive experience with and knowledge of Koyo's further-manufactured sales and the calculation of the cost of further manufacturing in the United States with respect to these sales. Therefore, in this case we have determined that for Koyo the relatively small reduction of burden on the Department that would result from resorting to either of the two statutory proxy methods under the special rule is outweighed by the potential distortion and losses in accuracy as a consequence of their use. Accordingly, we have rejected the use of either of the two proxies as inappropriate and have sought to calculate the CEP for Koyo's further manufactured sales using another reasonable basis. This determination, however, does not indicate that, because we found the alternative methods in section 772 (e)(1) and (e)(2) to be inappropriate, we have abandoned the special rule, as Koyo and NSK suggest. For all respondents with further-manufactured merchandise, we first evaluated whether the value added in the United States was likely to substantially exceed the value of the imported components. We determined that Fuji, NTN, and Koyo met the ``substantially exceeds'' qualification for implementation of the special rule. However, while we have determined that the use of either of the two proxy methods is appropriate for Fuji and NTN, we have found that for Koyo, resorting to either of the alternatives set forth in the special rule provision is not appropriate. If we determine that the use of one of the two proxies set forth in section 772 (e)(1) and (e)(2) is inappropriate, as explicitly directed by the statute, we may use any other reasonable basis to calculate CEP for further-manufactured sales. Here the statute again grants us considerable latitude in determining precisely what constitutes ``any other reasonable basis.'' The SAA at 825 indicates that one possible method is basing the CEP of the further-processed merchandise on the transfer price from the exporter or producer to the affiliated importer. In general, however, if the two statutory alternatives cannot be used, we should identify and use a method which not only satisfies the overall purpose of the provision--the reduction of the burden on the Department--but also furthers the goal of accuracy. A reasonable alternative, then, may be our standard further-processing analysis if its use is not unduly burdensome and if it sufficiently reduces the potential for inaccuracy or distortion. As explained in detail above, the record in this case indicates that the use of the standard methodology for calculating CEP for Koyo's further-manufactured sales is a reasonable method. Therefore, we disagree with NSK that the standard methodology cannot serve as another reasonable basis. Not only would its exclusion as another reasonable basis effectively eliminate the Department's ability to use an accurate and valid alternative in situations such as this, but the plain language of the provision clearly does not preclude the standard methodology as a viable alternative. In addition, we disagree with NSK that, because the SAA does not specifically reference the standard methodology as another reasonable basis, we are unable to use it as such. In fact, the SAA does not specifically exclude the standard methodology as an option. In addition, both the statute and the SAA clearly grant the Department discretion with respect to the determination of what [[Page 2562]] constitutes another reasonable basis. While NSK and Koyo correctly point out that the intent underlying Congress' enactment of the special rule was the reduction of the burden on the Department, both respondents overlook the fact that the Department, nevertheless, has an overriding mandate to calculate accurate dumping margins (see Bowe- Passat v. United States, 17 CIT 335, 340 (1993) (Bowe-Passat)). While the special rule provides us with a method to eliminate the burden of calculating the cost of further processing, its intent was not to elevate the goal of burden reduction over the goal of accuracy. Finally, we note that while NSK argues against the use of the standard methodology as another reasonable basis, it provides no alternative for calculating CEP for Koyo's further-manufactured merchandise, nor does it point to any record evidence establishing that the standard methodology would be, in this instant case, inappropriate for calculating CEP for Koyo's further-manufactured merchandise. As discussed in the summary above, Koyo does propose, however, an alternative for calculating the CEP of its further-manufactured A-588- 604 TRB merchandise, which it believes constitutes ``another reasonable basis.'' Koyo proposes that the Department, instead of evaluating whether the margins for finished over 4'' A-588-604 bearings were an appropriate surrogate for A-588-604 further-manufactured merchandise, could have used the margins it calculated for under 4'' A-588-054 bearings as a proxy for that A-588-604 merchandise which was further processed into under 4'' bearings, and the margins calculated for over 4'' bearings as a proxy for that A-588-604 merchandise which was further processed into over 4'' bearings. While Koyo's proposal would be less burdensome than the use of the standard methodology, we believe that the standard methodology is not unduly burdensome and presents a higher probability of accurate results than using margins calculated for non-further-manufactured sales. Among other things, Koyo's proposal relies on information concerning a different class or kind of merchandise and therefore in this case does not sufficiently allay concerns about potential inaccuracy. The record indicates that the use of these proxy methods would have a relatively high potential for distortion; we believe that the gains in accuracy we would achieve using the standard methodology would outweigh the additional burden resulting from the use of the standard calculation. The record supports our continued use of the standard methodology as a reasonable basis for calculating the CEP for Koyo's further- manufactured merchandise. Based on its incorrect presumption that we found its sales of identical or other A-588-604 subject merchandise to be in an insufficient quantity to be used as a proxy for its further- manufactured A-588-604 merchandise, Koyo argues that the Department should have compared the value of all imported unfinished components to the value of all finished bearings (whether subject to the A-588-604 order or A-588-054 finding) in order to make our sufficiency determination. Since we rejected the use of Koyo's identical or other A-588-604 subject merchandise based on our determination that these alternatives were inappropriate, Koyo's argument is irrelevant. However, we nonetheless note that section 772 (e)(1) and (e)(2) of the Act refers to identical or other subject merchandise. As a result, when determining if such sales occurred in a sufficient quantity, the statute clearly limits our determination to the scope of the order and does not permit the inclusion of non-subject sales as Koyo suggests. In light of all of the above, we have determined that the facts in this case support the selection of the standard methodology as a reasonable basis. Furthermore, because Koyo failed to comply with the Department's request for further-processing data, for these final results we have applied as adverse facts available to Koyo's further- manufactured merchandise the highest rate ever calculated for Koyo in any segment of the A-588-604 proceeding (36.21 percent). Comment 2: Timken argues that Department should adhere to its normal practice and apply an adverse facts available rate of 36.21 percent to the total sales value of Koyo's further-manufactured sales rather than to the total entered value of these sales. Citing the United States Court of Appeals for the Federal Circuit's (CAFC) decision in Olympic Adhesives v. United States, 899 F. 2d 1565, 1572 (Fed. Cir. 1990), Timken contends that the Department is required to draw an inference that is ``reasonably adverse'' to the respondent. However, Timken asserts, Koyo's selective submission of information has seemingly worked to its advantage in that Koyo has apparently received a lower margin despite the application of facts available. Timken maintains that this is supported by the fact that while the preliminary margin for Koyo in the A-588-604 case for this review is 23.26 percent, in previous reviews in which the Department calculated margins for both Koyo's A-588-604 further-manufactured and non-further manufactured sales, preliminary margins for Koyo were 46.03 percent (1992-1993) and 41.21 percent (1993-1994). Koyo responds that while the Department should not have applied facts available at all, nonetheless it should reject Timken's argument. Koyo argues that it would be difficult to imagine how applying the highest rate ever calculated for Koyo is not ``reasonably adverse.'' Furthermore, Koyo asserts that just because Timken is able to devise a more adverse approach in applying facts available does not mean that the Department's application of facts available is not ``reasonably adverse.'' Koyo also contends that the Department's choice of the highest rate ever for Koyo as the adverse facts available rate was reasonable considering the fact that Koyo cooperated with the Department in every aspect of the review with the only exception being its decision not to file a response to Section E of the Department's questionnaire. Department's Position: We disagree with Timken. In accordance with section 776 (a) and (b) of the statute and our consistent practice, because Koyo failed to cooperate to the best of its ability in responding to our requests for information by declining to provide data on its further-processed sales, we applied adverse facts available in the absence of the further-manufacturing sales information. In selecting from among the facts available, we chose as the adverse facts available rate to apply to Koyo's further-manufactured sales the highest rate we ever calculated for Koyo in any previous review of the A-588-604 case. We then applied that rate to the total entered value of Koyo's further-manufactured sales. In choosing among the facts available, we are not required by the statute to select a method that is ``the most'' or ``more'' reasonably adverse. In choosing the highest margin ever calculated for Koyo in the A-588-604 case, we have adhered to the statutory language and selected information that is adverse to the interest of Koyo. Timken has failed to offer arguments or provide record evidence demonstrating that the rate selected is not reasonably adverse. Therefore, for these final results, we have not changed our application of facts available with respect to Koyo's sales of further- manufactured TRB components. Comment 3: For the preliminary results of these reviews, we applied facts available with respect to certain of MC's sales because it did not provide [[Page 2563]] complete model-matching data. The absence of this information prevented us from finding a suitable home market match for these U.S. sales. Because the Department did not issue a supplemental questionnaire to MC prior to the preliminary results, we provided the company with an opportunity to correct its deficiencies for the final results. Timken argues that if MC is unable to provide the information necessary for matching certain of its U.S. sales to sales of the foreign like product, the Department should apply as adverse facts available to these unmatched U.S. sales the highest margin for any respondent in any review of the A-588-054 finding. Department's Position: As stated in our preliminary results, MC's questionnaire response contained only limited model-match information which prevented us from finding contemporaneous sales of the foreign like product for comparison to a small number of U.S. sales of subject merchandise (see TRB Prelim at 47455). As a result, in accordance with section 776(a) of the Act, we resorted to facts available. Because MC was not previously afforded the opportunity to remedy or explain its deficiencies, on September 16, 1997, we issued a supplemental questionnaire to MC requesting this information. On September 30, 1997, MC responded by submitting the necessary data. The information provided by MC has allowed us to find contemporaneous sales of the foreign like product to compare to all of MC's U.S. sales. Therefore, we have determined that it is not necessary to apply facts available to any of MC's U.S. sales for these final results. 2. Adjustments to Normal Value Comment 4: Timken asserts that because there is a discrepancy between NTN's computer tape and the total billing adjustment figure in a verification exhibit, NTN has incorrectly reported its home market billing adjustments. Thus, Timken argues that NTN's reporting is inconsistent with its narrative response and its verification exhibit and, given these inconsistences, the Department should convert NTN's negative billing adjustments to positive adjustments. NTN claims that there is no merit to Timken's request because the Department verified that its reported home market billing adjustments were accurate. Therefore, NTN argues that the Department should retain its billing adjustments as reported, and reject Timken's proposed adjustment. Department's Position: We agree with Timken in part. We examined NTN's home market database and found a significant discrepancy between the total billing adjustment for the POR located in exhibit three of our home market verification report and the total billing adjustment we derived from NTN's home market database. However, the difference between the two totals was significantly different from the difference Timken cited in its brief. Therefore, we reviewed the record to ascertain the accurate total for NTN's billing adjustments. During verification we thoroughly verified NTN's reported home market volume and value for the POR. As our verification report indicates, it was necessary for us to reconcile the volume and value NTN reported in its response to its Ministry of Finance (MOF) reports. As part of this reconciliation we examined an adjustment NTN made for its total HM billing adjustments for the POR (see Department's Home Market Verification Report for NTN, July 9, 1997, exhibit three) (NTN HM Report). Not only did we successfully trace this total to the computer program NTN used to calculate it, but we also traced NTN's reported volume and value for the POR for its home market sales directly to the MOF report with no discrepancies (see NTN HM Report at 6). We also verified NTN's reported, transaction-specific home market billing adjustments by examining a variety of sales documentation in the sales trace portion of our verification (see NTN HM Report at 17). Again we found no discrepancies. As a result of both verification exercises, one would assume that NTN's reported home market billing adjustments were accurate and that the total of its transaction- specific billing adjustments for the POR would equal the total reported on exhibit three of our verification report. However, this is not the case. In fact, when we calculated the overall POR total billing adjustment for NTN's home market database, this total was significantly different from that reported in the referenced exhibit three. Therefore, we needed to determine which billing adjustment figure was correct. For example, as we noted above, not only did we verify the accuracy of the total from exhibit three, but we also verified the accuracy of NTN's reported transaction-specific billing adjustments. After additional review, we have concluded that the exhibit three figure is the accurate total. We recognize that while our verification of NTN's reported transaction-specific adjustment yielded no discrepancies, this verification exercise constituted a ``spot check.'' In other words, we only examined selected billing adjustments. It is therefore possible that, while the samples we selected were correct, several of NTN's other reported transaction-specific billing adjustments were inaccurate. In fact, we have determined that other transaction-specific adjustments are inaccurate because the total of all billing adjustments does not match the total from exhibit three. Furthermore, because the total from exhibit three resulted in our successful trace to NTN's MOF reports, we find this total to be far more reliable than any other information on the record. Therefore, having determined that the exhibit three total billing adjustment amount is the accurate figure, we have adjusted NTN's reported transaction-specific billing adjustments to reflect this total. However, because our own analysis indicates an adjustment different than that calculated by Timken, while we agree with Timken that an adjustment is warranted, we have relied on our own calculated adjustment amount. Furthermore, because the record provides no information as to which transaction-specific billing adjustments are accurate, and because NTN has neither explained this discrepancy nor provided us with any information with respect to the correction of this discrepancy in its reported data, we have relied on facts available to correct NTN's reported home market billing adjustments. Because we are unable to identify which billing adjustments are inaccurate, as facts available, we systematically sorted through NTN's raw home market database and totaled the reported per-sale billing adjustments until we arrived at a total equal to our calculated adjustment. We then adjusted these sales' billing adjustments such that they reflected the total in exhibit three and disallowed the rest of NTN's reported billing adjustments. For a detailed description of this methodology please refer to the proprietary version of the Department's Final Analysis Memorandum for NTN, dated January 7, 1998. Comment 5: Timken contends that NTN used an incorrect denominator when calculating the ratio it used to allocate home market inventory carrying costs, resulting in inaccurate expense calculations. Timken also argues that even if NTN had calculated an accurate ratio, it nevertheless incorrectly applied this percentage to its home market sales. Therefore, Timken asserts, for purposes of the final results the Department should not only recalculate the inventory carrying cost expense ratio for [[Page 2564]] NTN's home market sales using the appropriate sales value (denominator), but it should also use the revised ratio to recalculate NTN's claimed adjustment. NTN argues that the Department thoroughly verified its calculation methodology for inventory carrying costs and found no discrepancies. NTN further asserts that not only is Timken's argument based on a misunderstanding of information on the record, but the Department has repeatedly accepted NTN's reporting methodology in prior TRB and antifriction bearings (AFB) cases. Tapered Roller Bearings and Parts Thereof, Finished and Unfinished, From Japan, and Tapered Roller Bearings, Four Inches or Less in Outside Diameter, and Components Thereof, From Japan, Preliminary Results of Antidumping Duty Administrative Reviews and Termination in Part, 61 FR 25200, 25202 (May 20, 1996) and Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, Germany, Italy, Japan, Singapore, Sweden, and the United Kingdom; Final Results of Antidumping Duty Administrative Reviews and Partial Termination of Administrative Reviews, 61 FR 66472, 66486 (December 17, 1996) (AFBs V). As a result, NTN argues, there is no basis for Timken's argument. Department's Position: We agree with NTN. Timken appears to have misunderstood the verification exhibits. However, it is clear from the information on the record that NTN accurately calculated and applied the appropriate ratio when allocating its home market inventory carrying costs. Not only did we verify NTN's inventory carrying cost allocation, including the denominator, without discrepancy (see NTN HM Report at 11), but NTN's response demonstrates that NTN applied an accurate ratio to all of its home market sales. Therefore, we have not recalculated NTN's reported home market inventory carrying costs for these final results. Comment 6: Timken argues that the Department should ensure that all of NTN's reported home market adjustments are accurate and claims that NTN's post-sale freight, pre-sale freight, packing labor, packing material, and indirect selling expenses (to include technical services, advertising, warehousing, and other indirect selling expenses) were incorrectly allocated using an inaccurate total sales value. Timken asserts that the Department should recalculate allocation ratios for all of NTN's expenses (except technical services) using the correct total sales value and apply these revised ratios to NTN's home market sales to calculate revised expense adjustments. NTN contends that these adjustments were successfully verified and that its methodology has been accepted by the Department in the past. NTN claims that for the reasons set forth in comments four and five, the Department need not reexamine the data verified in Japan. Further, NTN contends that, because Timken fails to put forth any rationale regarding their proposed modification of the ratio used to determine the amount of the adjustment, there is absolutely no grounds for the Department to reverify and modify data already closely examined. Department's Position: We agree with the respondent. In the instant review we conducted a thorough verification of NTN's reported home market adjustments to include post-sale freight, pre-sale freight, packing labor, packing material, and indirect selling expenses. We concluded that NTN's methodology yielded accurate results (see NTN HM Report at 8). After reviewing the record again for these final results, it is clear from the home market verification exhibits that all of NTN's adjustments were calculated correctly. Therefore, for these final results, we have accepted NTN's reported home market adjustments but have recalculated them without regard to levels of trade, as discussed in our response to comment 32 below. Comment 7: Timken claims that NTN failed to provide an adequate narrative response to the Department's supplemental questionnaire regarding its calculation of technical service expenses. Instead, Timken argues, NTN failed to demonstrate that its technical services were selling expenses, and, as a result, the Department should not make an adjustment to normal value (NV) for these expenses. NTN claims that Timken's argument is based on a misunderstanding of the NTN corporate brochure and that Timken misunderstands the relationship, role, and function of those entities which incurred technical service expenses during the POR. NTN argues that it has clearly demonstrated that all of its technical service activities are selling expenses and that, Timken's claims should be dismissed as groundless. Department's Position: We agree with NTN. In its brief Timken argues that a specific entity which NTN reported as incurring technical service expenses did not incur these expenses as selling expenses. However, it is clear from information on the record that the unit in question performed selling functions, including technical services, and thus incurred selling expenses, including technical service expenses. Furthermore, at verification we verified that the unit in question clearly performed selling functions, and clearly incurred selling expenses. Timken's claims are based on its failure to recognize the distinction between two separate divisions of the same unit which perform separate responsibilities. Therefore, for these final results, we agree with NTN and have continued to accept its reported technical service expenses. Comment 8: Timken argues that Koyo's transaction-specific home market billing adjustments (BILADJ1H) are already reflected in the reported gross unit prices and that, consequently, the Department should not adjust Koyo's unit prices for the BILADJ1H amounts. Timken asserts that the Department's June 20, 1997 home market verification report for Koyo (Koyo Home Market Report) supports its arguments that home market prices have already been revised to account for billing adjustments because the report notes that ``Koyo searched back through its database, located the matching sales transaction(s), canceled them out and re-entered the revised price'' (Koyo Home Market Report at 6). Additional evidence from the relevant verification exhibit, Timken argues, also supports its conclusions. For example, Timken notes, there is a particular TRB model which appears in Koyo's sales ledger at a given price (apparently the ``original'' price), but the reported gross unit price for a sales transaction involving that model is different from the ledger price. Timken further argues that the fact that the Department had to deduct post-sale price adjustments from the ledger totals to reconcile total value and volume supports its position that gross unit sales prices are already net of billing adjustments. Koyo responds that Timken has overlooked an error in the BILADJ1H computer reporting methodology that was corrected prior to verification. Koyo argues that the revised tapes it submitted to the Department after verification properly reflect gross unit prices prior to any billing adjustments. The reported gross unit price of the TRB model identified by Timken, Koyo argues, has been corrected on the revised tapes and now matches the ``original'' price (before any adjustments) appearing in its sales ledgers. Department's Position: We disagree with petitioner. We have reviewed the record, and have determined that the gross unit prices Koyo reported in its revised home market tape submitted on [[Page 2565]] May 30, 1997 are not net of billing adjustments. As Koyo explained in its rebuttal brief, it filed a revised home market sales file with the Department as a result of the error it discovered in the reporting methodology for this adjustment. The model identified by Timken was one of those affected by this reporting error (see Koyo Home Market Report at 6). Therefore, for these final results, we have not made any changes with respect to our treatment of Koyo's transaction-specific billing adjustments. Comment 9: Timken argues that the Department should reject Koyo's and NSK's claims for home market lump-sum post-sale price adjustments (PSPAs). With respect to Koyo, Timken asserts that the Department should deny an adjustment to NV for Koyo's customer-specific lump-sum billing adjustments (BILADJ2H) for the following reasons: (1) NV may not be modified by adjustments attributable to non-scope merchandise; (2) sales prices modified by adjustments not attributable to those particular sales cannot be used to calculate NV; (3) Koyo has not acted to ``the best of its ability'' in reporting its lump-sum billing adjustments; and (4) Koyo hasn't demonstrated that its allocation methodology is not distortive. Timken notes that Koyo has calculated its lump-sum billing adjustments by multiplying the total adjustment amount paid to a customer by the ratio of its TRB sales to that customer to the total sales to that customer. As a result, Timken argues, this adjustment is attributable to subject and non-subject merchandise and this allocation methodology attributes a portion of the adjustment to sales for which no adjustment was made. Timken claims that the Court of International Trade (CIT) in Torrington Co. v. United States, 17 CIT 199, 218, 818 F. Supp 1563, 1578 (1993) (Torrington) held that ``merchandise which is outside the scope of an antidumping duty order cannot be used in the calculation of antidumping duties,'' and its decision in this case is reason alone to reject Koyo's claim for lump-sum billing adjustments. Regarding its next claim, Timken asserts that the statute provides that ``normal value shall be the price'' at which the foreign like product is sold. However, Timken argues, as a result of Koyo's allocation methodology, there are some home market sales prices that have been modified by a portion of the lump-sum billing adjustment which is not properly attributable to those particular sales. Therefore, Timken argues, the modified price of such sales is not ``the price at which the foreign like product was first sold.'' Timken also argues that Koyo's claim for lump-sum billing adjustments should be denied because Koyo has not acted to the best of its ability in reporting them. Timken claims that there is information on the record which demonstrates Koyo could have devised a computer program to match lump-sum adjustments to the relevant sales. For example, Timken argues that while in its questionnaire response Koyo states that it does not maintain lump-sum adjustments in a customer- specific manner, it also states in the response that it ``matched these debit and credit notices to the relevant sales in order to report the billing adjustments on a transaction-specific basis.'' Timken maintains that this is exactly the type of computer programming that would allow Koyo to attribute lump-sum billing adjustments to the sales upon which they were granted. Timken also argues that additional evidence of Koyo's ability to match lump-sum adjustments to relevant sales is contained in the home market verification report, which describes how Koyo was able to design a program that searched its database to reenter revised unit prices. Finally, Timken asserts that Koyo has not demonstrated that its lump-sum billing adjustments are not distortive, and claims that Koyo's May 30, 1997 home market sales printout demonstrates this. Timken points to two examples of similar gross unit prices from the sales tape that have been modified by lump-sum billing adjustments, and argues that because each of the adjustments is a given percentage of the unit price, all those sales which have had the adjustment allocated to them, even though they were not in the group of sales to which the adjustment is correctly attributed, have been modified by that percentage. Timken further contends that such a difference is distortive given that the statute recognizes any margin over .5 percent as significant, and recommends that the Department subtract (rather than add) BILADJ2H from the gross unit price to correct this distortion. Petitioner also urges the Department to reject NSK's claims for lump-sum rebates. Timken argues that NSK failed to accurately report the amount of it customer-specific lump-sum PSPAs directly attributable to specific sales of scope merchandise, and that, accordingly, the Department should deny an adjustment to NV for these PSPAs. Timken argues that the Department's acceptance of NSK's lump-sum adjustments is contrary to Federal-Mogul Corp. v. United States, 834 F. Supp. 1391 (CIT 1993), in which the CIT, upon remand, ordered the Department to attempt to devise a methodology that removed PSPAs and rebates paid on sales of non-subject merchandise from FMV (NV), and to deny such an adjustment if they could not be removed. In that decision, Timken contends, the CIT held that PSPAs and rebates paid on both subject and non-subject merchandise were acceptable provided the amount paid per sale was the same throughout the POR. Timken further argues that because the CIT determined that NSK did not meet the standard for acceptance of its lump-sum PSPAs, and because NSK's allocation methodology for the current reviews is apparently unchanged, its lump- sum adjustments should not be allowed. Timken asserts that the CIT, in other decisions, has upheld the rejection of NSK's lump-sum adjustments. For example, in Torrington Co. v. United States, 881 F. Supp 622 (CIT 1995), Timken argues, the CIT ordered the Department on remand to develop a methodology for removing PSPAs paid on non-subject merchandise from the calculation of FMV (NV). Timken also cites Torrington Co. v. United States, 926 F. Supp. 1151 (CIT 1996), and claims that the CIT, in that decision, held that PSPAs which could not be tied specifically to the sales for which they were granted could not be treated as direct expenses and affirmed the disallowance of NSK's lumps-sum adjustments because they could not be tied to specific part numbers. Timken further contends that in Timken Co. v. United States, 930 F. Supp. 621 (1996), the CIT held that Commerce should not have allowed an adjustment to FMV (NV) for NSK's lump-sum PSPAs. Timken also cites to the CIT's more recent decision in NSK vs. United States, 969 F. Supp. 34 (CIT 1997), asserting that the CIT affirmed the Department's disallowance of NSK's lump-sum PSPAs. In sum, Timken argues that NSK's allocation methodology is distortive because, although every individual payment of a lump-sum adjustment ties to a certain group of sales transactions (and not to all sales), NSK nonetheless allocates the total of lump-sum payments to all POR sales. In addition, Timken claims that NSK has not acted to the best of its ability in reporting its lump-sum PSPAs. Koyo responds that its lump-sum billing adjustments are not distortive. Citing to the most recent AFB and TRB reviews, Koyo asserts that the Department has evaluated Koyo's lump-sum billing adjustment methodology and in each case has allowed the adjustment. The reporting methodology [[Page 2566]] for the current reviews, Koyo asserts, is the same as that which the Department has previously accepted. Koyo argues that Timken's reliance on Torrington to support its assertions that the adjustment should be denied because it was allocated over a total sales value which included non-subject merchandise is inappropriate because the case was decided before the enactment of the URAA. Furthermore, Koyo argues, even prior to the URAA, the CAFC rejected the distinction between scope and non-scope merchandise used by the CIT. In addition, Koyo argues that the Department determined that the Torrington case was of limited relevance in evaluating Koyo's billing adjustment allocation methods for 1994/95 AFB review (see Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, et. al.; Final Results of Antidumping Administrative Review, 62 FR 2091 (January 15, 1997) (AFBs VI)). With respect to Timken's claims that Koyo has not demonstrated that its allocation methodology is not distortive, Koyo responds, citing the 1995/1996 AFB Final Results (Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof From France, Germany, Italy, Japan, Romania, Singapore, Sweden, and the United Kingdom; Final Results of Antidumping Duty Administrative Reviews, 62 FR 54043 (October 17, 1997)(AFB's VII)), that the Department has established a test for determining whether or not allocations are distortive. Koyo argues that the Department's test does not depend on whether sales in the allocation pool were of non-subject merchandise, nor does it depend upon the difference between the allocated and actual adjustment. Koyo contends that the purpose of the test is to determine if merchandise in the allocation pool is significantly different in terms of value, physical characteristics, and the manner in which it was sold. Koyo further argues that evidence on the record in the instant reviews demonstrates that it has met the Department's test regarding whether an allocation is distortive. For example, Koyo claims that the merchandise over which it allocated billing adjustments was similar in terms of value and physical characteristics and that as a result, the sales patterns of these products were similar to the merchandise to which the adjustments were specifically attributable. Citing AFBs VI and AFBs VII, Koyo notes that the Department concluded in the three most recent AFB reviews that ``it is not feasible for Koyo to report this adjustment on a more specific basis'' and that ``we are satisfied that Koyo's allocation methodology across subject merchandise by sales value was not distortive.'' Koyo adds that the Department followed the intent of the URAA to liberalize reporting requirements in accepting billing adjustments and asserts that the SAA at 823-824 makes clear this intent. Furthermore, Koyo claims, the Department's new antidumping regulations allow the Department to consider allocated expenses and adjustments where transaction-specific reporting is not possible, and that Koyo's allocation methodology is consistent with the expressed goals of the new regulations. Koyo claims that calculating BILADJ2H on a transaction-specific basis is not possible. Because a portion of the adjustments are attributable to more than one sale, Koyo asserts that data linking a given sale to a particular adjustment does not exist within its database. Finally, regarding Timken's assertion that any adjustment which exceeds the de minimis threshold points to a distortive allocation, Koyo responds that the concept of the de minimis threshold is not related to whether or not an allocation is distortive. NSK responds that the judicial precedent relied on by Timken to support its assertions includes cases decided before the enactment of the URAA which do not govern the Department in post-URAA reviews. Furthermore, NSK argues that the SAA and the new antidumping regulations support the Department's acceptance of NSK's methodology for allocating lump-sum PSPAs. NSK adds that the Department, in recent AFB reviews, has accepted NSK's lump-sum PSPAs. Department's Position: We disagree with Timken. We have granted claims for PSPAs as direct adjustments to NV if we determined that the respondent, in reporting these adjustments, acted to the best of its ability in providing information and meeting the requirements we have established with respect to these adjustments, and that its reporting methodology was not unreasonably distortive (see section 782(e) of the Act). We did not treat such adjustments as direct or indirect selling expenses, but as direct adjustments to identify the correct starting price. While our preference is for transaction-specific reporting, we recognize that this is not always possible. It is inappropriate to reject allocations that are not unreasonably distortive where a fully cooperating respondent is unable to report the information in a more specific manner (see section 782(e) of the Act). Accordingly, we have accepted these adjustments when it was not feasible for a respondent to report these adjustments on a more specific basis, provided that the allocation method used does not cause unreasonable inaccuracies or distortions. In applying this standard, we have not rejected an allocation method solely because the allocation includes adjustments granted on non-scope merchandise. However, such allocations are not acceptable where we have reason to believe that respondents did not grant such adjustments in proportionate amounts with respect to sales of out-of- scope and in-scope merchandise. We have made this determination by examining the extent to which the out-of-scope merchandise included in the allocation pool is different from the in-scope merchandise in terms of value and physical characteristics, and the manner in which it is sold. Significant differences in such terms may increase the likelihood that respondents did not grant price adjustments in proportionate amounts with respect to sales of subject and non-subject merchandise. While we scrutinize any such differences carefully between in-scope and out-of-scope sales in terms of their potential for distorting reported per-unit adjustments on the sales involved in our analysis, it would be unreasonable to require that respondents provide sale-specific adjustment data on non-scope merchandise in order to prove that there is no possibility for distortion. Such a requirement would defeat the purpose of permitting the use of reasonable allocations by a respondent that has cooperated to the best of its ability. Based on our examination of the record in this and in past reviews, we are satisfied that Koyo's records do not allow it to report these billing adjustments on a transaction-specific basis and that Koyo acted to the best of its ability in calculating the reported adjustment on as narrow a basis as its records allowed. Therefore, for these final results we have made a direct adjustment to NV for Koyo's lump-sum billing adjustments. With respect to NSK, as explained in our preliminary analysis memorandum, we have accepted its claims for lump-sum rebates because we are satisfied that NSK's methodology, while it includes non-subject merchandise, does not shift rebates from non-scope to scope merchandise. In its response, NSK submitted information demonstrating that the ratio of scope to non-scope merchandise purchased by each customer who received this rebate was relatively constant throughout the POR. Furthermore, we have determined based [[Page 2567]] on our review of the record that NSK acted to the best of its ability in reporting these price adjustments and that reporting on a more specific basis was not possible given the manner in which NSK maintains its records. Comment 11: Timken argues that Koyo's home market rebates were not allocated over the actual sales for which they were incurred. Timken further asserts that Koyo did not act to the best of its ability in reporting this adjustment because the record indicates that Koyo had the capability to properly link rebates to specific sales. Finally, Timken maintains that Koyo has provided no evidence demonstrating that its allocation methodology is not distortive and asserts that the Department should consequently deny Koyo's claim for home market rebates. Koyo responds that, as with its billing adjustments, the measure of whether or not an allocation is distortive does not depend on the difference between an allocated and actual adjustment or whether the allocation pool includes merchandise for which the expense was not originally incurred. Koyo argues that when attempting to determine whether an allocation is distortive, the Department examines the extent to which merchandise in the allocation pool is different from merchandise for which the expense was incurred in terms of value, physical characteristics, and the manner in which it is sold. In fact, Koyo asserts, the Department, in the most recently completed TRB review, found that the subject and non-subject bearings included in the allocation pool for home market rebates did not differ significantly with respect to value, physical characteristics, or sales patterns. Koyo further argues, citing the home market verification report, that the Department's finding that ``[there were] no discrepancies in either the program or the calculation methodology'' demonstrates that Koyo's rebate allocation methodology is not distortive. Finally, Koyo argues that because the Department in the 1994/1995 TRB reviews determined that Koyo had acted to the best of its ability in reporting rebates, and because the allocation methodology for these reviews is unchanged, the Department should continue to make a direct adjustment to home market price for rebates for these final results. Department's Position: We disagree with Timken. During the POR Koyo granted rebates to certain of its home market customers. Koyo calculated rebate factors by dividing the total rebates paid to a given customer by the total POR sales to that customer. In our supplemental questionnaire, we asked Koyo to explain why it was unable to report home market rebates on a more specific basis. In its supplemental response Koyo stated that more specific reporting for a certain customer who received rebates was not possible because its records did not allow it to isolate sales of those bearings for which rebates were granted. Based on information Koyo provided, we are satisfied that Koyo acted to the best of its ability in reporting home market rebates. However, because Koyo's allocation methodology includes non-scope merchandise, we have nevertheless examined Koyo's allocation to determine if it is distortive. Our review of the record indicates that the non-scope merchandise included in Koyo's allocation are sales of bearings other than TRBs. Not only has our review and analysis of the record given us no reason to believe that Koyo is more likely to grant rebates on sales of bearings other than TRBs than on sales of TRBs, but we note that Koyo is primarily in the business of selling bearings, some of which are within the scope of the TRB orders and others which are not. While we recognize that there are differences among bearings, we have not found that the scope and non-scope bearings included in Koyo's allocation vary significantly in terms of value, physical characteristics, nor the manner in which they were sold such that Koyo's allocation would result in an unreasonably inaccurate or distortive allocation. Therefore, for these final results we have made no changes in our treatment of Koyo's home market rebates. Comment 12: Timken argues that Koyo's domestic pre-sale freight expenses should be allocated equally to sales in the home market and in the United States. Timken contends that Koyo's practice of allocating Japanese pre-sale freight expenses to U.S. sales on the basis of transfer prices is potentially distortive because such prices are not at arm's length. Koyo's allocation methodology, Timken argues, has the effect of shifting expenses attributable to U.S. sales to sales in the home market. Timken also argues that Koyo's response demonstrates that there are certain home market sales for which the company did not incur pre-sale freight expenses. Timken suggests that, because the record indicates that Koyo maintained warehouses at its plants during the POR, and because pre-sale freight expenses are not incurred for sales shipped directly from the plant warehouse to the customer, the Department should follow its practice from the 94-95 TRB review in which it removed the total sales value of Koyo's OEM home market sales from the denominator of the expense ratio. Koyo responds that Timken's argument regarding its allocation methodology for pre-sale freight has been rejected by the Department in past reviews and urges the Department to once again dismiss it. Koyo argues that it has reported its pre-sale freight expense in the same manner as in past reviews, and asserts that the Department has verified and accepted its use of Koyo Seiko's total sales value as the denominator for calculating pre-sale freight. Koyo further maintains that its total bearing sales amounts were based on its ``Sales and Cost of Goods Sold'' summary, and that the total bearing sales for its distributors were based on figures the Department tied to the audited financial statements. Furthermore, Koyo argues that these sales totals included those to affiliated and unaffiliated customers in the home market and export markets. Koyo argues that because its home market and export sales were to a mix of both affiliated and unaffiliated customers, its allocation methodology was fair and that the proper basis for allocation is its prices for all relevant sales. Department's Position: We agree with petitioner in part. While we agree that Koyo's questionnaire response does indicate that it did not incur pre-sale freight expenses for certain home market sales, we disagree with Timken that Koyo's allocation of these expenses is otherwise unreasonable. In its response Koyo reported home market pre- sale freight expenses which reflected those expenses it incurred when transporting TRBs destined for sale in both the U.S. and home markets from the home market plant to home market warehouses. While Koyo reported these pre-sale freight expenses for all of its home market and U.S. sales, its questionnaire response indicates that there are certain home market sales for which Koyo did not incur this expense because the merchandise was not transported from the plant to a warehouse at a location different from the plant. For example, on page 36 of its section B response to our questionnaire, Koyo explains that, prior to sale, not only did it store TRBs at its two home market central warehouses, warehouses at its branch and sales offices, and at the warehouses of its consolidated distributors, but it also stored certain merchandise at its plant warehouse. In the proprietary explanation following this description Koyo again indicates that there are certain types of home market sales for which the merchandise was stored at its plant warehouse. In [[Page 2568]] addition, on page 23 of its section B response, when explaining its post-sale home market freight expenses, Koyo states that it incurred post-sale freight expenses either in shipping merchandise from the plant directly to a customer or when transporting merchandise from a warehouse to a customer. Again, this indicates that there are certain home market sales for which the merchandise is shipped directly from the plant to a customer and, therefore, is not transported to a warehouse at a location different from the plant. Therefore, we agree with Timken that the record demonstrates that there are certain home market sales for which Koyo did not incur home market pre-sale freight expenses. We have determined that for these final results it is necessary to (1) reallocate Koyo's reported home market pre-sale freight expenses such that the total sales value of those home market sales for which the expense was not incurred is excluded from the allocation denominator, and (2) apply the expense only to those home market sales for which the expense was incurred. However, Koyo's response does not enable us to specifically identify within Koyo's home market database those sales for which the expense was not incurred. In light of this, we have determined to rely on facts available to determine those sales for which the expense was not incurred. Based on Koyo's proprietary narrative explanation on page 36 of its response, we have concluded that Koyo did not incur this expense on certain sales to home market OEM customers. While we recognize that it is likely that not all of Koyo's home market OEM sales were exempt from this expense, because we are unable to identify exactly which OEM sales were exempt, we have applied non-adverse facts available and recalculated the expense adjustment by (1) removing from Koyo's reported allocation denominator the total sales value of Koyo's home market OEM sales and (2) applying the recalculated expense adjustment to U.S. sales and only non-OEM home market sales. However, despite the fact that we have determined for these final results that Koyo's pre-sale freight allocation denominator is overstated and the expense was reported for home market sales for which it was not incurred, we disagree with Timken that Koyo's allocation otherwise fails to reflect the manner in which the expense was actually incurred. In general, when a respondent relies on an expense allocation to calculate its per-unit adjustment amounts, we require that allocation to reflect the manner in which the expense was actually incurred (see, e.g., 92-93 TRB Final at 57635 and Certain Fresh Cut Flowers From Columbia; Final Results of Antidumping Duty Administrative Reviews, 61 FR 42848 (August 19, 1996)). In addition, we examine the respondent's allocation methodology to determine if there is internal consistency between the numerator and denominator and in the methodology as a whole. For example, if an expense is allocated on the basis of total sales value, as is the expense at issue here, the expense amount (the numerator) and the total sales value (the denominator) should reflect the same pool of sales such that the total expense amount reported by the respondent is divided by the total value of the sales for which the expense was actually incurred. Likewise, the allocation ratio should be applied to the same sales price reflected in the denominator. For example, we would not accept the application of an allocation ratio to home market gross sales price if the denominator was calculated by totaling the value of all sales on the basis of a net price. In the instant case, Koyo Seiko, the Japanese parent, incurred the pre-sale freight expenses at issue for all merchandise, whether destined for sale to the U.S., third-country, or home market (with the exception of the home market OEM sales described above). Because Koyo does not maintain its records such that it is able to calculate the total expense amount incurred for each market, it was unable to separately calculate the specific pre-sale freight expense attributable to each market. Therefore, Koyo used as its allocation numerator the total expense amount incurred by Koyo Seiko for all merchandise, as derived from Koyo Seiko's sales records. The sales for which this expense was incurred were Koyo Seiko's sales to all its customers, which encompassed a mix of affiliated and unaffiliated entities in both the export and home markets. Thus, Koyo calculated its pre-sale freight allocation denominator by totaling the value for all of Koyo Seiko's sales to all its customers, as derived from Koyo Seiko's records. While for these final results we have adjusted this denominator to exclude the total sales value of home market OEM sales, we have nevertheless preserved Koyo's basic allocation methodology. Because Koyo Seiko's customers encompassed a mix of affiliated and unaffiliated parties in both the home and export markets, Koyo's denominator includes sales values which reflect both transfer and resale prices. Since Koyo Seiko's customer in the United States is Koyo Corporation of U.S.A. (KCU), its wholly-owned U.S. affiliate, the U.S. sales transactions relevant to Koyo's allocation are those between Koyo Seiko and KCU. Thus, Koyo correctly included within its denominator the total value of its sales to KCU, which were made at transfer prices. Similarly, in the home and third-country markets Koyo Seiko sold to both affiliated and unaffiliated customers. Therefore, Koyo properly included within its allocation denominator the total value of Koyo Seiko's sales to its home and third-country market customers, some of which were made at resale prices while others were at transfer prices. Koyo's methodology, therefore, not only relies on a numerator and denominator which reflect the same pool of sales, but its denominator is calculated on the basis of the value of those sales for which the reported total expense amount was actually incurred. When calculating the per-unit expense adjustment amount for each U.S. and home market transaction, Koyo applied its allocation ratio (which was the same for all sales) to the appropriate unit price. For U.S. sales it applied the ratio to the transfer prices Koyo reported between Koyo Seiko and KCU, which were the U.S. prices upon which the expense was incurred and the U.S. sales values reflected in Koyo's allocation denominator. For home market sales, Koyo applied the ratio to either a resale price (for unaffiliated customers) or transfer price (for affiliated customers) because these were the home market prices upon which the expense was incurred and the home market sales values reflected in the allocation denominator. Timken argues that, in order to properly reflect commercial reality and avoid distortion, Koyo should instead apply its expense ratio to U.S. resale prices, the price between KCU and the first unaffiliated U.S. customer. However, Timken overlooks the fact that this transaction is not the sale for which the expense was actually incurred. As a result, Timken's proposed methodology would neither reflect the manner in which, nor the sales upon which, Koyo actually incurred the expense. Timken's argument also ignores the fact that Koyo's allocation denominator includes not only U.S. transfer values but home market and third-country transfer values as well. Thus, Timken's assertion that Koyo always calculates the home market expense adjustment on the basis of resale prices is incorrect. Rather, the record demonstrates that, for sales to affiliated home market parties, Koyo calculated the adjustment on the basis of the transfer price between Koyo Seiko [[Page 2569]] and the affiliated home market customer. In addition, rather than argue that all transfer values included in Koyo's denominator should be excluded from the allocation methodology, Timken limits its argument to only U.S. transfer prices and fails to demonstrate why U.S. transfer values are an improper factor in the denominator's calculation while home market and third-country transfer values are not. Finally, the record does not contain, and Timken has not provided, any evidence demonstrating that the transfer prices Koyo reported between Koyo Seiko and KCU are unreliable. Rather, the record indicates that these transfer prices were maintained by KCU, for purposes other than antidumping proceedings, within the ordinary course of business. Furthermore, we note that antidumping proceedings are only one of the factors a respondent must account for in setting its transfer prices; transfer prices are also subject to possible Internal Revenue Service audits for U.S. tax purposes and to U.S. Customs' review. Therefore, based on the above reasons, we do not agree with the petitioner that Koyo's basic allocation methodology is unreasonable. Therefore, for these final results, while we have recalculated Koyo's originally reported allocation ratio to exclude home market OEM sales, we have made no other changes to Koyo's overall allocation methodology. Comment 13: Timken argues that the Department should recalculate Koyo's home market average short-term borrowing rate to exclude interest amounts which it maintains are aberrational and unsupported by the record. Timken asserts that home market verification documents detailing loans taken by Koyo during the POR contain two loan entries which do not list certain relevant information regarding the terms and details of these loans for which the reported interest was incurred. Koyo argues that Timken's assertions are misplaced because they are based on a misunderstanding of the credit verification exhibit. Koyo argues that the interest amounts Timken identified as aberrational do not constitute payments on specific loans, but rather reflected interest paid by Koyo Seiko under some other arrangement. Citing the Department's home market verification report, Koyo asserts that the Department has already verified the accuracy of Koyo's reported credit expense ratio and found no discrepancies. Department's Position: We disagree with Timken. During verification we carefully reviewed the manner in which Koyo calculated its short- term interest rate and its credit expense ratios. After reviewing supporting documentation for each of several loans we selected from Koyo's credit calculation worksheets, we were satisfied that Koyo had accurately reported its credit expense. While those entries identified by Timken were not among those chosen, we emphasize that the purpose of verification is not to conduct an exhaustive review of a response. Rather, verification is intended to serve as a spot check to verify the overall integrity of a response (see, e.g., Bomont Industries v. United States, 14 CIT 208, 209, 733 F. Supp. 1507 (1990)). Absent any unreconciling information regarding Koyo's calculation of its short- term borrowing rate, we were satisfied that this expense was accurately reported. Furthermore, we are generally satisfied with Koyo's explanation of and the reliability of those interest amounts which Timken claims should be removed from the interest rate calculation and can find no evidence on the record that indicates these interest amounts should be excluded from the calculation of credit; accordingly, we have not done so for these final results. Comment 14: In its case brief Timken argued that Koyo allocated its home market inventory carrying costs (ICC) over a sales value which improperly excluded sales to its distributors. However, Timken subsequently withdrew that argument. In addition, Timken also suggested that the Department should recalculate Koyo's ICC ratio using the following methodology: (1) calculate separate ICC ratios for Koyo Seiko sales and sales by its wholly-owned distributors; (2) apply the Koyo Seiko ICC ratio to all of its sales; and (3) apply both Koyo Seiko's and the distributors' ICC rates to the affiliated distributors' sales. Timken contends that this methodology would produce a more accurate result because merchandise stored by Koyo Seiko and by its distributors remains in inventory for different average periods of time. Koyo did not specifically comment on Timken's proposed method for allocating ICC. Department's Position: We agree with Timken. Exhibit B-9 of Koyo's questionnaire response indicates that, to calculate its reported home market ICC, Koyo derived an overall average number of days using inventory balance and total sales figures for both Koyo Seiko and its consolidated distributors. However, because there were significant differences between the inventory balances and total sales values for Koyo Seiko as compared to those for its consolidated distributors, Koyo's calculation of a single average number of days in inventory has the effect of overstating the ICC incurred for those sales made by the consolidated distributors and understating the ICC incurred for sales made by Koyo Seiko. Therefore, because information exists on the record which would allow us to calculate home market ICC which more closely reflect the actual experience of Koyo Seiko and its consolidated distributors, we have recalculated Koyo's reported home market ICC for these final results. Please see our Final Results Analysis Memorandum for Koyo, dated January 7, 1998 for a detailed explanation of our recalculation. Comment 15: Timken argues that Koyo improperly excluded a certain category of expenses from its reported total export selling expenses. Koyo responds that its methodology properly excludes these expenses and contends that Timken's argument is based on a misunderstanding of the home market verification report and of Koyo's August 26, 1997 letter in response to Timken's pre-preliminary comments. Department's Position: We disagree with Timken. We have reviewed the home market verification report and the relevant exhibit and have determined that Koyo correctly excluded this expense category from its total export selling expenses. The proprietary nature of this argument prevents us from discussing it in further detail here. For more information, refer to the proprietary version of our final results analysis memorandum for Koyo, dated January 7, 1998. 3. Adjustments to United States Price Comment 16: Timken argues that the Department should apply facts available with respect to Koyo's pre-sale U.S. inland freight because Koyo failed to demonstrate that the expenses attributed to subject merchandise are reasonably accurate. Timken is concerned that Koyo's allocation methodology is distortive because (1) the total reported expenses include those associated with shipping merchandise from Japan, Europe, and from Koyo Corporation of U.S.A. Manufacturing Division (KCUM) to KCU, and (2) Koyo allocated freight costs based on the weight of all sales, including those sales for which KCU apparently did not incur a freight expense. Koyo maintains that Timken incorrectly assumes that this expense is allocated on the basis of sales value when it was actually allocated on the basis of the weight of the merchandise shipped. Koyo further maintains that [[Page 2570]] allocating this expense based on weight is not distortive because the cost to ship a given weight of non-scope and scope merchandise is identical. Koyo also argues that its allocation methodology is well- established, has been repeatedly verified, and was again verified without discrepancy at the Department's U.S. verification for these reviews. Department's Position: We disagree with Timken. Koyo calculated its U.S. inland pre-sale freight expense ratio by dividing KCU's total POR freight-in expenses (associated with shipping merchandise from railheads to KCU's warehouses) as reported in its financial statements by the total gross weight shipped to U.S. customers of all products. Koyo then multiplied the resulting freight-in factor by the unit gross weight to arrive at its reported pre-sale freight amounts. During verification, not only did we carefully examine Koyo's methodology for allocating its U.S. pre-sale inland freight expenses, but we tied KCU's reported total pre-sale freight expenses directly to its financial statements and found no discrepancy. In addition, we verified that the gross weight reported by Koyo was accurate (see Koyo Seiko U.S. Verification Report, August 7, 1997, at 13). As noted above, the expense total appearing in Koyo's numerator encompasses POR freight-in expenses incurred when shipping merchandise (whether scope or non-scope) from Europe and Japan to KCU's sales warehouses, and from KCUM to KCU. Similarly, the denominator includes the POR gross weight of all such sales for which these expenses were incurred. We have examined the record and are satisfied that Koyo's records do not allow it to report these expenses on a more specific basis. Additionally, while Timken asserts that KCU apparently allocates U.S. inland pre-sale freight expense totals to certain sales for which KCU did not pay for freight transfers from KCUM to KCU, we can find no evidence on the record indicating that KCU did not incur all expenses associated with shipping merchandise from KCUM to KCU's sales warehouse. Because we are satisfied that Koyo's allocation is as specific as possible, and because the numerator and denominator properly reflect all shipments and all expenses, we have not resorted to the use of facts available for these final results. Comment 17: NTN argues that the Department should have calculated CEP profit on a level-of-trade (LOT)-specific basis. NTN claims that the Department noted that prices differed significantly based on the LOT at which merchandise was sold. NTN claims that selling expenses also differed by LOT and had an effect on prices but that this difference does not account entirely for the different price levels. NTN further emphasizes that section 772 (a) and (f) of the Act expresses a preference for the profit calculations to be performed as specifically as possible and on the narrowest basis as possible. Finally, NTN asserts that because the Department calculated constructed value (CV) profit on a LOT-specific basis and matched U.S. and home market sales by LOT, the calculation of CEP profit should also take LOT into account. Timken argues that the Department rejected the identical argument by NTN in its final results of the sixth review of the AFBs case, stating that ``neither the statute nor the SAA require us to calculate CEP profit on a basis more specific than the subject merchandise as a whole. * * * [t]he statute and SAA, by referring to ``the'' profit, ``total actual profit,'' and ``total expenses'', imply that we should prefer calculating a single profit figure'' (see AFBs VI at 2081 and 2125). For these same reasons, Timken contends that the Department should again reject NTN's assertion in this TRB review. Department's Position: We agree with Timken. Neither the statute nor the SAA requires us to calculate CEP profit on a basis more specific than the subject merchandise as a whole. See AFBs VI at 2125. Respondent's suggestion would not only add a layer of complexity to an already complicated exercise with no increase in accuracy, but a portion of the CEP-profit calculation would be more susceptible to manipulation. Therefore, for these final results we have not changed our CEP profit calculation. Comment 18: NTN asserts that the Department had no basis for including EP sales in the calculation of the CEP profit adjustment and argues that section 772 (a) and (f) of the Act clearly state that the adjustment for profit to CEP sales is to be based on the expenses incurred in the United States as a percentage of total expenses. NTN contends that section 772(d) of the Act contains no provision for the inclusion of export price expenses and that the canon of statutory construction, expressio unius est exclusio alterius, indicates that the absence of such a provision precludes its inclusion. NTN further asserts that the SAA similarly states that ``the total expenses are all expenses incurred by or on behalf of the foreign producer and exporter and the affiliated seller in the United States with respect to the production and sale of. . . the subject merchandise sold in the United States and the foreign like product sold in the exporting country (if Commerce requested this information in order to determine the normal value and the constructed export price)''. Therefore, NTN claims that the Department has calculated CEP profit in a manner contrary to that specified in the statute. Department's Position: We disagree with NTN. The Department's September 4, 1997 policy bulletin regarding the calculation of CEP profit indicates that section 772(f)(2)(D) of the Act clearly states that the calculation of total actual profit is to include all revenues and expenses resulting from the respondent's EP sales as well as from its CEP and home market sales. The basis for total actual profit is the same as the basis for total expenses under section 772(f)(2)(C) of the Act. The first alternative under this section states that, for purposes of determining profit, the term ``total expenses'' refers to all expenses incurred with respect to the subject merchandise sold in the United States (as well as home market expenses). Thus, where the respondent makes both EP and CEP sales to the United States, sales of the subject merchandise would encompass all such transactions. Therefore, because NTN had EP sales, we have included these sales in the calculation of CEP profit. Comment 19: NTN argues that the Department's decision to ignore adjustments to its U.S. indirect selling expenses for expenses incurred when financing cash deposits for antidumping duties is contrary to both the Department's position in past reviews and judicial precedent, and that it inappropriately denies an adjustment for expenses incurred solely as a result of the existence of an antidumping order. NTN asserts that the CIT has previously held that these imputed interest expenses do not constitute selling expenses, and cites PQ Corp. v. United States, 11 CIT 53, 67 (1987) (PQ Corp), in which the CIT stated, ``if deposits of estimated antidumping duties entered into the calculation of present dumping margins, those deposits would work to open up a margin where none otherwise exists.'' NTN claims that the rationale in PQ Corp applies similarly to interest incurred when financing cash deposits, and asserts that if the Department were to allow interest expenses from previous reviews to affect the calculation of margins for current reviews, it would precipitate an unending cycle which would prevent the Department from revoking an order. [[Page 2571]] NTN maintains that the CIT, in Timken v. United States, Slip Op. 97-87 (July 3, 1997)(Timken), upheld NTN's adjustments to U.S. indirect selling expenses for interest incurred when financing cash deposits, and notes that the Department itself argued in support of such an adjustment. NTN argues that, as set forth in Timken, interest expenses attributable to cash deposit financing are not incurred in the course of selling merchandise in the United States. NTN also references the CIT's decision in Federal Mogul Corp. v. United States, Slip Op. 96-163 (December 12, 1996), claiming that the CIT explicitly rejected the petitioner's argument that interest expenses constituted selling expenses because they were incurred as a result of NTN's ``decision'' to engage in dumping. Additionally, argues NTN, the Court rejected the petitioner's argument that allowing such an adjustment was duplicative of interest paid on the refund of excess cash deposits. Timken responds that the Department correctly rejected NTN's claim for a downward adjustment to U.S. indirect selling expenses for interest incurred when financing cash deposits. Timken argues that allowing such an adjustment serves as an incentive to respondents to prolong litigation to avoid actual payment of duties, which is contrary to the purpose of the interest provision set forth in the Trade Agreements Act of 1979, which is to reduce incentives to delay payment of duties owed. Department's Position: We agree with Timken that we should deny an adjustment to NTN's U.S. indirect selling expenses for expenses which NTN claims are related to financing of cash deposits. The statute does not contain a precise definition of what constitutes a selling expense. Instead, Congress gave the administering authority discretion in this area. It is a matter of policy whether we consider there to be any financing expenses associated with cash deposits. We recognize that we have, to a limited extent, removed such expenses from indirect selling expenses for such financing expenses in past reviews of this finding, this order, and other orders. However, we have reconsidered our position on this matter and have now concluded that this practice is inappropriate. Further, we note that the Court's affirmance of our prior policy does not preclude us from following this new, reasonable policy. We have long maintained, and continue to maintain, that antidumping duties, and cash deposits of antidumping duties, are not expenses that we should deduct from U.S. price. To do so would involve a circular logic that could result in an unending spiral of deductions for an amount that is intended to represent the actual offset for the dumping (see, e.g., Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof from France, et. al.; Final Results of Antidumping Duty Administrative Reviews, 57 FR 28360 (June 24, 1992) (AFBs II); see also, e.g., Certain Cut-to-Length Carbon Steel Plate from Germany; Final Results of Antidumping Duty Administrative Review, 62 FR 18390, 18395 (April 17, 1995)). We have also declined to deduct legal fees associated with participation in an antidumping case, reasoning that such expenses are incurred solely as a result of the existence of the antidumping duty order (see AFBs II). Underlying our logic in both these instances is an attempt to distinguish between business expenses that arise from economic activities in the United States and business expenses that are direct, inevitable consequences of an antidumping duty order. Financial expenses allegedly associated with cash deposits are not a direct, inevitable consequence of an antidumping duty order. As we stated in the preliminary results at 47455: ``[m]oney is fungible. If an importer acquires a loan to cover one operating cost, that may simply mean that it will not be necessary to borrow money to cover a different operating cost.'' Companies may choose to meet obligations for cash deposits in a variety of ways that rely on existing capital resources or that require raising new resources through debt or equity. For example, companies may choose to pay deposits by using cash on hand, obtaining loans, increasing sales revenues, or raising capital through the sale of equity shares. In fact, companies face these choices every day regarding all their expenses and financial obligations. There is nothing inevitable about a company having to finance cash deposits and there is no way for the Department to trace the motivation or use of such funds even if it were. In a different context, we have made similar observations. For example, we stated that ``debt is fungible and corporations can shift debt and its related expenses toward or away from subsidiaries in order to manage profit'' (see Ferrosilicon from Brazil, 61 FR at 59412 (regarding whether the Department should allocate debt to specific divisions of a corporation)). So, while under the statute we may allow a limited exemption from deductions from U.S. price for cash deposits themselves and legal fees associated with participation in dumping cases, we do not see a sound basis for extending this exemption to financing expenses allegedly associated with financing cash deposits. By the same token, for the reasons stated above, we would not allow an offset for financing the payment of legal fees associated with participation in a dumping case. We see no merit to the argument that, since we do not deduct cash deposits from U.S. price, we should also not deduct financing expenses that are arbitrarily associated with cash deposits. To draw an analogy which shows why this logic is flawed, we also do not deduct corporate taxes from U.S. price; however, we would not consider a reduction in selling expenses to reflect financing alleged to be associated with payment of such taxes. Finally, we also determine that we should not use an imputed amount that would theoretically be associated with financing of cash deposits. There is no real opportunity cost associated with cash deposits when the paying of such deposits is a precondition for doing business in the United States. Like taxes, rent, and salaries, cash deposits are simply a financial obligation of doing business. Companies cannot choose not to pay cash deposits if they want to import nor can they dictate the terms, conditions, or timing of such payments. By contrast, we impute credit and inventory carrying costs when companies do not show an actual expense in their records because companies have it within their discretion to provide different payment terms to different customers and to hold different inventory balances for different markets. We impute costs in these circumstances as a means of comparing different conditions of sale in different markets. Thus, our policy on imputed expenses is consistent; under this policy, the imputation of financing costs to actual expenses is inappropriate. Comment 20: Timken contends that the Department should recalculate NTN's U.S. credit expense because NTN reported a customer-specific average credit expense rather than a transaction-specific credit expense. Timken argues that NTN has provided the necessary information on the record to recalculate a transaction-specific credit expense. Further, Timken claims that, based on a comparison of the credit amounts reported by NTN to those credit amounts which are derived when using NTN's reported transaction-specific sales payment dates, it is apparent that NTN's customer-specific methodology produces distortive results. [[Page 2572]] NTN argues that the Department has upheld its methodology in several past proceedings and has verified the accuracy of NTN's data, not only in this review, but in previous reviews as well. Department's Position: We agree with Timken with regard to NTN's CEP sales. We have data on the record which allows us to calculate a transaction-specific credit expense for CEP sales. Therefore, we have recalculated NTN's credit expense using the dates of payment which NTN reported. Comment 21: Timken contends that NTN improperly excluded certain expenses from its reported U.S. indirect selling expenses and states that, for the purpose of final results, the Department should deduct these expenses from CEP. NTN argues that not only has the Department rejected Timken's claim in past reviews, determining that NTN's reporting methodology was accurate, but in this current review the Department thoroughly verified this methodology and again found no discrepancies. Department's Position: We agree with the respondent. As NTN has explained and as we have repeatedly accepted, because certain of its U.S. expenses were incurred solely for non-scope merchandise, in order to ensure an accurate allocation of its U.S. expenses, NTN first removed all such expenses from its pool of U.S. expenses. The remaining expenses which were incurred for either scope or non-scope merchandise, but cannot be specifically linked to either scope or non-scope merchandise by NTN, were then allocated to scope and non-scope merchandise. We have consistently determined this methodology to be reasonable not only in past reviews of these TRB cases but in past reviews of AFB cases as well (see 92/93 TRB Final and AFBs VII). In addition, for this review, we verified NTN's U.S. expenses and found no discrepancies (see Department's U.S. Verification Report for NTN, June 3, 1997, at 10) (NTN U.S. Report). Because NTN has not altered its methodology for this current review and because the record in this review indicates no reason for a different methodology to be used, we have again accepted this methodology for these final results. Comment 22: Timken alleges that certain of NTN's claimed EP transactions are actually CEP transactions when examined in light of the criteria for defining EP transactions as outlined in the Department's Antidumping Manual. Petitioner states that EP sales must meet the following criteria; (1) the sales transaction occurs prior to importation; (2) the merchandise in question was shipped directly from the manufacturer to the unrelated buyer, without being introduced into the inventory of the related selling agent; (3) this was a customary commercial channel for sales of this merchandise between the parties involved; and (4) the related agent in the United States acted only as a processor of the sales-related documentation and a communication link with the unrelated U.S. buyer. Timken argues that, when the activities of the related selling agent exceed the functions normally associated with a related agent involved with EP sales, the sale cannot be classified as an EP sale. For example, petitioner asserts that the Department's Antidumping Manual (1994) states that ``the extent of the related selling agent's normal functions, such as the administration of warranties, advertising, extensive in-house technical assistance, and the supervision of further manufacturing, may indicate that the agent is more than the ``paper-pusher'' envisioned for purchase price sales'' (see Antidumping Manual, Chapter 7 at 4-5). Timken claims that evidence on the record indicates that NTN's U.S. subsidiary, NBCA, performed numerous functions which exceeded those normally associated with a related agent involved in EP sales transactions. As a result, Timken concludes, the Department should reclassify all of NTN's reported EP sales as CEP sales. NTN argues that, as the Department has verified in this current and in previous TRB reviews, (1) there were no sales negotiations between its unaffiliated EP customer and NBCA, (2) NBCA did not receive any purchase orders from the unaffiliated customer, (3) NBCA did not generate any invoices for unaffiliated customers, (4) NBCA never took title to the merchandise in question, (5) NBCA never carried the merchandise in its inventory, and (6) NBCA never acted as the importer of record. In summary NTN states, these sales were clearly made in Japan and clearly met the Department's definition of EP sales transactions. Furthermore, NTN adds, the record demonstrates that NBCA acted solely as a communications link and a processor of documents with respect to U.S. EP sales. Department's Position: We agree with NTN. Timken lists the criteria the Department considers when deciding whether sales should be classified as EP or CEP. Of the criteria outlined, however, the only area that Timken questions is the activities of NBCA's liaison office. As NTN notes, there is no information on the record suggesting that NBCA is the seller for the sales in question or that NTN performed activities that exceeded those normally associated with the role of a related agent in EP transactions. Moreover, we verified NTN's response for this review and found that NBCA's functions with respect to EP sales were limited to being a communications link and a processor of documents. Therefore, we have not reclassified NTN's EP sales for these final results. Comment 23: Due to the proprietary nature of the comments we received regarding NTN USA's expenses, we are unable to state the concerns expressed by both NTN and Timken and our position with response to this issue. Therefore, for a detailed explanation of this issue and our position, please see the proprietary analysis memorandum for NTN dated January 7, 1998. 4. Cost of Production (COP) and Constructed Value (CV) Comment 24: NTN claims that the Department's preliminary results adjustment to COP and CV for affiliated-party inputs is distortive and should be eliminated. NTN argues that the Department's adjustment, which was calculated based on sampled transactions, does not accurately reflect the experience of all sales and, by applying the results of the sample to the total population of affiliated-party purchases, the Department, in essence, used facts available when sufficient information was clearly available on the record. NTN further argues that the Department misinterpreted section 773(f)(2) and (3) of the Act by determining that an adjustment was necessary. NTN claims that section 773(f)(2) of the Act addresses the circumstances under which the Department should disregard some transactions, but it does not mention the Department's practice of choosing the highest of either the cost of production, transfer prices, or market prices when calculating COP or CV. NTN additionally claims that Section 773(f)(3) of the Act requires the Department to have reasonable grounds to believe that inputs are being sold at less than the COP before it may use COP information. NTN contends that, because the record demonstrates that affiliates sold many inputs to NTN above COP, it is incorrect for the Department to adjust the costs for all TRBs which contained affiliated- party inputs. NTN also asserts that, assuming the Department was correct in making an adjustment for affiliated-party inputs, it should use a more reasonable recalculation methodology such as the weighted-average difference between [[Page 2573]] COP and transfer price for all inputs sold to NTN. According to NTN, by adjusting all sales which had an affiliated-party input, the Department added additional profit to those inputs which already included profit. Therefore, NTN concludes that the Department should use NTN's affiliated-party input data as reported. The petitioner contends that the Department acted in accordance with section 773(f)(2) and (3) of the Act and that NTN failed to demonstrate that the Department's adjustments produced distorted results. According to the petitioner, section 773(f)(2) and (3) of the Act states that the Department may use information available in circumstances such as those which exist with respect to NTN in this review and that the Department has the right to use discretion in selecting the highest of (1) the transfer price from an affiliated party, (2) the COP for the input, or (3) the price from the unaffiliated party. Timken disputes NTN's interpretation of section 773(f)(3) of the Act, stating that the statute does not require the Department to act only if it is able to determine that all inputs have been priced below COP. Rather, Timken argues, the Department may act when it has reasonable grounds to believe or suspect that an amount represented as the value of an affiliated-party input is less than the COP of the input. Moreover, the petitioner asserts, the Department acted reasonably in making its preliminary adjustment because it had limited data regarding NTN's affiliated-party inputs. Thus, the Department reasonably determined that the problem it had identified was likely to affect all models with affiliated-party inputs. Finally, Timken claims, for each TRB part number where the COP of the affiliated-party input was greater than the transfer price, the Department should increase COP and CV by an amount equal to the difference between transfer price and the COP. Department's Position: We disagree with NTN that our adjustment to increase certain transfer prices to equal a market price is flawed. The Department tested affiliated-party inputs on a sample basis, and applied the results of the sample to the total population of affiliated-party transactions. Our adjustment relied on affiliated- party factors provided by NTN in its COP and CV database and it accounted for the fact that only certain inputs obtained from certain affiliates did not reflect a market value. The preamble of section 351.407 of the Final Rule at 27296 and 27413 leaves conducting an arm's-length test of the transfer price to the Department's discretion depending on the facts and circumstances of the case. In this instance, NTN provided the transfer prices and cost information for its major inputs. We examined this information on a sample basis and determined that the company's reported amounts were not less than its respective COP, as required by section 773(f)(3) of the Act. NTN also provided a market value for identical or similar inputs obtained from or sold to non-affiliated parties to establish that the transfer price was comparable to the market price. We then examined this information on a sample basis and determined that in certain instances the company's reported transfer prices did not reflect a market price, as required by section 773(f)(2) of the Act. As noted on page 24 of the June 13, 1997 cost verification report, NTN could not explain the difference between the transfer price and the market price. Thus, for the preliminary results we used the results of our samples to increase the manufacturing costs of the control numbers NTN identified as including related-party inputs. We also disagree with NTN's contention that it is not appropriate for the Department to rely on section 773(f)(2) and (3) of the Act in this instance. We note that section 351.407 (a) and (b) of the Final Rule, at 27296 and 27413, sets forth certain rules that are common to the calculation of CV and COP. This section states that for the purpose of section 773(f)(3) of the Act the Department will determine the value of a major input purchased from an affiliated person based on the higher of: (1) the price paid by the exporter or producer to the affiliated person for the major input; (2) the amount usually reflected in sales of the major input in the market under consideration; or (3) the cost to the affiliated person of producing the major input. Furthermore, we have relied on this methodology in Final Results of Antidumping Duty Administrative Review; Certain Corrosion-Resistant Carbon Steel Flat Products and Certain Cut-to-Length Carbon Steel Plate From Canada, 62 FR 18449, 18457 (April 15, 1997), AFBs VI at 2115, and the 92/93 TRB Final. In each of these review's final results, the Department determined that in the case of a transaction between affiliated persons involving a major input, we will use the highest of the transfer price between the affiliated party, the market price between unaffiliated persons involving the major input, or the affiliated supplier's cost of producing this input. Accordingly, for the final results we adjusted NTN's reported costs to account for the difference between the transfer price and market value for inputs purchased from affiliated parties based on the adjustment factor used in the preliminary results. Comment 25: Petitioner states that the Department should ensure that the calculation of COP and CV includes certain non-operating expenses (e.g., certain write-offs, depreciation of idle equipment, foreign currency gains and losses, etc.) NSK, on the other hand, contends that the exclusion of these non-operating expenses is permissible and is based on past Department practice. Department's Position: For the final results we have relied on NSK's reported general expense factor that excludes certain non- operating income and expenses. We reviewed the information on the record and noted that NSK included depreciation of idle equipment in its COP. As for the other non-operating expenses identified by the petitioner, we note that NSK excluded them from the calculation of COP. However, these non-operating expenses are minor expenses. Thus, including them in the calculation of the dumping margin has a de minimis effect on the calculation of NSK's margin. Comment 26: NSK argues that, in accordance with section 773(f) of the Act, the Department may only substitute affiliated-party costs for a respondent's reported transfer prices for affiliated-party inputs for purposes of sections 773(b) and 773(e) of the statute. However, NSK asserts, in the preliminary results the Department also substituted affiliated-party cost data when it determined whether the foreign like product was commercially comparable to each U.S. model, when it calculated a difference-in merchandise (difmer) adjustment for non- identical U.S. and home market matches, and when it recalculated NSK's reported U.S. inventory carrying costs prior to deducting this expense from CEP. Citing Ad Hoc Comm, of AZ-NM-TX-FL Producers of Grey Portland Cement v. United States, 13 F.3d 398, 401 (Fed. Cir. 1994), NSK contends that where Congress has included specific language in one section of the statute but has omitted it from another, related section of the same statute, it is generally presumed that Congress intended the omission. Therefore, NSK argues, because the statutory authority to determine whether the foreign like product is commercially comparable to the U.S. merchandise, to adjust NV for difmer, and to adjust CEP for U.S. inventory carrying costs is found in sections 771(16), 773(a)(6), and 772(d) of the Act, respectively, and not in [[Page 2574]] section 773(f), and because section 773(f) specifically limits the substitution of related-party costs to sections 773(b) and 773(e) of the statute, the antidumping law clearly does not permit the Department to use affiliated-party cost data to determine commercial comparability, to calculate the difmer adjustment, or to calculate an adjustment to CEP for inventory carrying costs. Therefore, NSK concludes, the Department should rely on NSK's reported cost data without regard to affiliated-supplier cost data in all instances except where specifically authorized by the statute. NSK also asserts that the substitution of affiliated-party costs when determining commercial comparability constitutes an alteration of the Department's model-match methodology and prevents respondents from taking advantage of the Department's TRB Option II reporting methodology. NSK argues that it is not only difficult for a respondent to obtain affiliated-party cost data in time to integrate it into the model match, but it is often the case that an affiliated supplier refuses to provide the respondent with its cost data. As a result, NSK contends, through no fault of its own, a respondent's inability to obtain affiliated-party cost data may result in the inability to compare appropriate models and in the Department's use of total facts available. Timken argues that, contrary to NSK's assertions, there is nothing in the statutory provisions cited by NSK which restricts the Department's discretion to use adjusted cost data for purposes other than sections 773(b) and 773(e) of the statute. For example, Timken maintains, section 771(16) of the Act, the ``model-match'' provision, only instructs the Department to select comparison merchandise that is ``like'' the U.S. subject merchandise in component material and uses and is ``approximately equal in commercial value,'' and does not specify the methodology by which the Department is to select the similar comparison merchandise or determine commercial comparability. Rather, citing the CAFC's decision in Koyo Seiko v. United States, 66 F.3d 1204 (Fed. Cir. 1995), Timken contends that, because Congress has implicitly delegated authority to the Department to determine and apply a model-match methodology, it was not inappropriate or unlawful for the Department to rely on affiliated-party cost data in making its commercial comparability determination for NSK. Likewise, Timken argues that the provision which underlies the Department's difmer adjustment, section 773(a)(6) of the Act, does not detail the precise methodology that the Department must use to make such an adjustment. Hence, Timken states, Congress has again implicitly delegated authority to the Department to formulate an appropriate methodology and the Department reasonably determined that it was appropriate to use NSK's affiliated-party cost data when calculating this adjustment. Timken also asserts that section 772(d) of the Act does not detail the methodology the Department is to use to calculate ICC adjustments to CEP but only lists the kinds of expenses that may be deducted from CEP. Therefore, Timken argues, Congress has once again implicitly delegated authority to the Department to select an appropriate methodology to calculate ICC and other expenses. Finally, Timken argues, the Department's substitution of affiliated-party cost data when determining the commercial comparability of NSK's home market comparison merchandise is not likely to have a significant impact on the Department's model matches. Moreover, Timken concludes, not only are respondents required to supply data on multiple models for matching under the TRB Option II reporting methodology, but any respondent concerned about the potential effect on the model-match may revise its submission accordingly. Department's Position: We agree with Timken. In our preliminary results for NSK, in accordance with section 773(f) of the Act, we recalculated NSK's reported TRB-specific COP and CV to include the COP of an affiliated-party input if the transfer price NSK reported for that input was less than the COP for that input. We note that COP and CV are composed of several components. The adjustment we made for NSK's affiliated-party inputs is actually an adjustment to its reported material costs. Because material costs are a component of the variable cost of manufacture (VCOM) and the total cost of manufacture (TCOM), and these in turn are components of COP and CV, when we adjusted NSK's reported material costs we not only recalculated its COP and CV, but we effectively recalculated VCOM and TCOM components of COP and CV as well. NSK's assertions overlook the fact that the Department does not rely on a respondent's reported costs solely for the calculation of COP and CV. We also use cost information in a variety of other aspects of our margin calculations. For example, when determining the commercial comparability of the foreign like product in accordance with section 771(16) of the Act, it has been our long-standing practice to rely on the product-specific VCOMs and TCOMs for U.S. and home market merchandise. Likewise, when calculating a difmer adjustment to NV in accordance with section 773(b) of the Act, it has been our consistent policy to calculate the adjustment as the difference between the product-specific VCOMs for the U.S. and home market merchandise compared (see, e.g., 92-93 TRB Prelim at 57631). Furthermore, we have permitted respondents to calculate their reported ICC on the basis of TCOM. As a result, if we determine a component of a respondent's COP and CV is distortive for one aspect of our analysis, it is reasonable to make the same determination with respect to those other aspects of our margin calculations where we relied on the identical cost data. To do otherwise would not only produce distortive results but would be contrary to our mandate to administer the dumping laws as accurately as possible. NSK incorrectly asserts that section 773(f) of the Act specifically limits substitution of affiliated-party cost data to our analysis under sections 773(b) and 773(e). In fact, section 773(f) indicates that for purposes of subsections (b) and (e) we may substitute certain cost data but 773(f) does not prohibit this kind of substitution for other purposes. None of the sections of the statute (771(16), 772(d), and 773(a)(6)), for which NSK argues that we may not substitute affiliated- party costs, explicitly precludes the incorporation of corrected cost data. For example, the only guidance provided by section 771(16) of the Act is that the comparison merchandise be ``like'' the U.S. subject merchandise in terms of component material and uses and ``approximately equal in commercial value.'' Therefore, as Timken points out, section 771(16) of the Act does not specify a particular methodology for determining appropriate matches. Rather, the statute implicitly delegates the selection of an appropriate methodology to the Department. Likewise, section 773(a)(6) of the Act grants us the same discretion to determine a suitable method to calculate a difmer adjustment and does not restrict our selection of an appropriate methodology to any particular approach. In addition, with respect to our recalculation of NSK's U.S. ICC, section 772(d) of the Act only specifies what adjustments are to be made to determine CEP and does not provide details regarding the precise [[Page 2575]] calculations for each particular adjustment. Accordingly, we have not altered our model-match, difmer, or calculation of NSK's ICC for these final results. 5. Miscellaneous Comments Related to Duty Absorption, Sample Sales, Level of Trade, and the Arm's-Length Test Comment 27: Timken contends that the Department's decision not to make an adjustment to CEP to account for indirect selling expenses and ICC incurred in Japan because expenses were not related specifically to commercial activity in the United States was incorrect. Timken argues that under pre-URAA law the Department deducted all selling expenses incurred in exporting to the United States and that the new law was not intended to change the Department's practice. Timken contends that the SAA clearly indicates that Congress did not intend to change the old law insofar as the Department's prior treatment of selling expenses was concerned. Further, Timken asserts that under the Department's new regulations (19 CFR 351.402(b), 62 FR at 27411), CEP should be adjusted if the expenses in question are related to the sale to the unaffiliated customer in the United States but not if they are only associated with the sale to the U.S. affiliate. Therefore, Timken argues that the Department should implement the SAA and the understanding Congress intended by deducting export selling expenses incurred in Japan from the calculation of CEP. NTN, NSK, Koyo, and Fuji assert that the SAA fully supports the Department's decision not to adjust CEP to account for indirect selling expenses and ICC incurred in Japan and cite to section 772(d) of the Act which states that ``constructed export price will be calculated by reducing the price of the first sale to an unaffiliated customer in the United States by the amount of the following expenses (and profit) associated with economic activities occurring in the United States.'' SAA at 823. Further, respondents argue that the Department has used the same methodology in Tapered Roller Bearings and Parts Thereof, Finished and Unfinished, From Japan, and Tapered Roller Bearings, Four Inches or Less in Outside Diameter, and Components Thereof, From Japan; Final Results of Antidumping Duty Administrative Reviews and Termination in Part, 62 FR 11825, 11834 (March 13, 1997) (94-95 TRB Final) and AFBs VII at 54043 and 54055, in which the Department concluded that export selling expenses are not specifically associated directly with commercial activity in the United States. Department's Position: As we stated in the 94-95 TRB Final at 11825, 11834 and AFBs VI at 2124, we will deduct from CEP only those expenses associated with economic activities in the United States which occurred with respect to sales to the unaffiliated U.S. customer. We found no information on the record for this review period to indicate that the indirect selling expenses and ICC for the respondents that were incurred in their respective home markets were incurred on sales to the unaffiliated customer in the United States. In addition, it is clear from the SAA that under the new statute we should deduct from CEP only those expenses associated with economic activities in the United States. The SAA also indicates that ``constructed export price is now calculated to be, as closely as possible, a price corresponding to an export price between non- affiliated exporters and importers'' (see SAA at 823). Therefore, we have deducted from CEP only those expenses associated with commercial activities in the United States. Timken's reference to the SAA to support the proposition that the new law is not intended to change our practice in this regard is misplaced. Timken cites various provisions of the SAA which state that our practice with respect to ``assumptions'' would not change. The SAA explains that ``assumptions'' are selling expenses of the purchaser for which the foreign seller agrees to pay (see SAA at 824). Thus, if the home market producer agrees to pay for the affiliated importer's cost of advertising in the U.S. market, the Department would deduct such an expense as an ``assumption.'' It should be noted that assumptions are different than selling expenses incurred in the home market in selling to the affiliated importer, which are not incurred ``on behalf of the buyer'' (i.e., the affiliated importer). Rather, the exporter incurs such expenses on its own behalf, and for its own benefit, in order to complete the sale to the affiliated importer (see AFBs VI at 2124). In this case respondent's reported selling expenses at issue were not associated with commercial activity in the United States. Rather, the expenses at issue were incurred prior to the commercial activity in the United States. Therefore, because the respondents' reported export selling expenses and ICC did not represent commercial activities performed in the United States, we did not deduct these expenses from CEP for these final results. Duty Absorption As indicated in the introduction to this notice, section 751(a)(4) of the Act provides for the Department, if requested, to determine whether antidumping duties have been absorbed by a foreign producer or exporter subject to the order if the subject merchandise is sold in the United States through an importer who is affiliated with such foreign producer or exporter, and authorizes this type of investigation during an administrative review initiated two years or four years after publication of an order. For transition orders as defined in section 751(c)(6)(C) of the Act (i.e., orders in effect as of January 1, 1995), section 351.213(j)(2) of the Department's regulations provides that the Department will make a duty-absorption determination, if requested, for any administrative review initiated in 1996 or 1998. See 62 FR 27296, 27394 (May 19, 1997). Although these antidumping regulations are not binding upon the Department for these TRB reviews, they do constitute a public statement of how the Department expects to proceed in construing section 751(a)(4) of the Act. This approach ensures that interested parties will have the opportunity to request a duty-absorption determination prior to the time of the sunset review of the order under section 751(c) of the Act on entries for which the second and fourth years following an order have already passed. Because this finding and order on TRBs have been in effect since 1976 and 1987, these are transition orders in accordance with section 751(c)(6)(C) of the Act; therefore, based on the policy stated above, the Department will consider a request for a duty-absorption determination during a review initiated in 1996 or 1998. On December 11, 1996, Timken requested that the Department determine, with respect to various respondents, whether antidumping duties had been absorbed during the POR. Since these reviews were initiated in 1996 and such a request was made, we have made a duty-absorption determination as part of these administrative reviews. In our preliminary results of review we calculated the percentage of sales by a U.S. affiliate with dumping margins for each exporter. We stated that, with respect to those companies (with affiliated importer(s)) that had dumping margins, we would rebuttably presume that the duties will be absorbed for those sales which were dumped. Subsequent to the preliminary results, we received comments regarding our duty-absorption determination but have [[Page 2576]] not changed our presumption for these final results. Comment 28: NSK, NTN, and Koyo claim that the Department has interpreted section 351.213(j) of its regulations incorrectly as providing for duty-absorption inquiries in the second and fourth years following a sunset review after which an order is continued and in periods such as the seventh and ninth reviews for transition orders. Citing the principle of statutory construction ``expressio unius est exclusio alterius,'' wherein there is an inference that all omissions should be understood as exclusions, respondents conclude that the lack of explicit Congressional approval for duty-absorption inquiries for the latter transition orders shows that Congress did not intend for duty-absorption inquiries to be initiated more than four years after publication of an antidumping order. Finally, respondents contend that the Department is incorrect in justifying the duty-absorption inquiry by considering the TRB order and finding as transitional in accordance with section 751(c)(6)(C) of the Act. According to respondents, section 751(c)(6)(C) of the Act only applies to ``sunset'' reviews. Timken claims that not only does narrowing the applicability of the duty-absorption inquiries to only the second and fourth years of sunset reviews unduly limit the effectiveness of the statute, but there is no indication that sections 751(a)(4) or 751(c)(6)(D) of the Act intended such a narrow application. Timken's response to the legal principle of ``all omissions should be understood as exclusions'' is that it has little force in the administrative setting because deference is granted to an agency's interpretation of a statute, unless Congress has directly spoken to the question at issue (citing Mobile Communications Corp. Of America v. F.C.C., 77 F.3d 1399, 1404-1045). Timken further argues that ``whether the specification of one matter means the exclusion of another is a matter of legislative intent for which one must look at the statute as a whole'' (citing Massachusetts Trustees of Eastern Gas & Fuel Associates v. United States, 312 F.2d 214, 220 (1st Cir. 1963) (citing authority), aff'd, 377 U.S. 235 (1964)). Department's Position: As for the time frame in which we are conducting these reviews, section 351.213(j)(1) of our regulations, in accordance with section 751(a)(4) of the Act, provides for the conduct, upon request, of duty-absorption inquiries in reviews initiated two and four years after the publication of an antidumping duty order (see e.g., AFBs VII at 54043 and 54044). The preamble to the proposed antidumping regulations explains that reviews initiated in 1996 will be considered initiated in the second year and reviews initiated in 1998 will be considered initiated in the fourth year (see Final Rule at 7317). Because the TRB order and finding have been in effect since 1987 and 1976, respectively, these are transitional in accordance with section 751(c)(6)(C) of the Act (see e.g., AFBs VII at 54044 and 54075). This being a review initiated in 1996 and a request having been made, we have made duty-absorption determinations as part of these administrative reviews. Comment 29: Respondents argue that measuring duty absorption based on information not known until the completion of an administrative review is unfair. More specifically, they claim that the nature of the review process prevents them from determining the U.S. price increase necessary to pass dumping duties on to customers because the ultimate liability is not determined until the end of a review. Respondents argue further that, other than dumping duties paid at the time of entry, they have no means of estimating the price increases necessary to pass dumping duties on to the customers. Finally, respondents argue that the Department cannot presume ``rebuttably'' that duty absorption on sales to a U.S. affiliate exists if the record does not contain evidence of the U.S. purchaser's assumption of liability for ultimate assessment. Respondents claim that the Department's rebuttable presumption ignores commercial reality in that no U.S. buyer would agree to assume liability for an unascertainable amount of duties. Respondents claim that the Department has not provided any reason for adopting the presumption of duty absorption and that the presumption is not allowable by law. Timken agrees with the Department's approach in using the rebuttable presumption that the duties for sales that were dumped will be absorbed. Timken argues that the Department's examination of whether duty absorption occurred by reviewing data on the volume of dumped imports and dumping margins follows the guidelines of the SAA. Timken argues that the Department's decision was reasonable, given the lack of record evidence that the first unrelated customer will be responsible for paying the duty that is ultimately assessed, the consistency of the Department's dumping determinations, and the fact that the Department gives the respondents the opportunity to provide evidence that the unaffiliated purchasers will pay the assessed duty. Department's Position: We agree with Timken. An investigation as to whether there is duty absorption does not simply involve publishing the margin in the final results of review. As the Department noted in the preliminary results of these reviews, the determination that duty absorption exists is also based on the lack of any information on the record that the first unaffiliated customer will be responsible for paying the duty that is ultimately assessed. Absent such an irrevocable agreement between the affiliated U.S. importer(s) and the first unaffiliated customer, there is no basis for the Department to conclude that the duty attributable to the margin is not being absorbed (see, e.g., AFBs VII at 54043 and 54044). As was the case with the most recently completed review of AFBs, this is an instance where the existence of a margin raises an initial presumption that the respondent and its affiliated importer(s) are absorbing the duty. As such, the burden of producing evidence to the contrary shifts to the respondent (see Creswell Trading Co., Inc. v. United States, 15 F.3d 1054 (CAFC 1994)). Here, the respondents have failed to place evidence on the record, despite being given ample time to do so, in support of their position that they and their affiliated importer(s) are not absorbing the duties (see, e.g., AFBs VII at 54043 and 54044). Comment 30: Koyo and NSK argue that, even if a duty-absorption inquiry is lawful, the Department's duty-absorption methodology fails to measure duty absorption on respondents' U.S. sales database as a whole. Respondents claim that by not considering sales made at non- dumped prices the Department fails to get an accurate measure of whether duty absorption has occurred. Timken responds that taking into consideration negative margins in a duty-absorption inquiry may indirectly lead to increased levels of dumping. Timken asserts that while sales priced above ``dumping levels'' may in fact allow an importer to engage in duty absorption, this does not change the likelihood that dumping will increase upon revocation of an order. Department's Position: We disagree with respondents that we should aggregate negative and positive margins in our duty-absorption determination. The Department treats so-called ``negative'' margins as being equal to zero in calculating a weighted-average margin because otherwise exporters would be able to mask their dumped [[Page 2577]] sales with non-dumped sales (see Final Determination of Sales at Less Than Fair Value; Professional Electric Cutting Tools and Professional Electric Sanding/Grinding Tools from Japan, 58 FR 30149 (May 26, 1993)). It would be inconsistent on one hand to calculate margins using only positive-margin sales, which is the Department's practice, and then argue, in effect, that there are no margins for duty-absorption purposes because a deduction from the total duties determined should be made for sales without margins (see AFB VII at 54043 and 54076, citing Certain Hot-Rolled Lead and Bismuth Carbon Steel Products From the United Kingdom; Final Results of Antidumping Duty Administrative Review, 62 FR 18744, 18745 (April 17, 1997)). However, non-dumped sales affect the percentage of sales through affiliated importers which are dumped and therefore affect the results of the absorption inquiry. Level of Trade (LOT) As set forth in section 773(a)(7) of the Act and in the SAA at 829- 831, to the extent practicable we have determined NV based on sales at the same LOT as the LOT of the EP and CEP sales. When we were unable to find comparison sales at the same LOT as the EP or CEP sales, we compared the U.S. sales to sales at a different LOT in the comparison market. We determined the LOT of EP sales on the basis of the starting prices of sales to the United States. We based the LOT of CEP sales on the price in the United States after making the CEP deductions under section 772(d) of the Act but before making the deductions under section 772(c) of the Act. Where home market prices served as the basis of NV, we determined the NV LOT based on starting prices in the NV market. Where NV was based on CV, we determined the NV LOT based on the LOT of the sales from which we derived SG&A and profit for CV. In order to determine the LOT of U.S. sales and comparison sales, we reviewed and compared distribution systems, including selling functions, classes of customer, and the extent and level of selling expenses for each claimed LOT. Customer categories such as distributor, original equipment manufacturer (OEM), or wholesaler are commonly used by respondents to describe LOTs but are insufficient to establish a LOT. Different LOTs necessarily involve differences in selling functions, but differences in selling functions, even substantial ones, are not alone sufficient to establish a difference in the LOTs. Different LOTs are characterized by purchasers at different stages in the chain of distribution and sellers performing qualitatively or quantitatively different functions in selling to them. See AFBs VI at 2105. As in the preliminary results, where we established that the comparison sales were made at a different LOT than the sales to the United States, we made a LOT adjustment if we were able to determine that the differences in LOTs affected price comparability. We determined the effect on price comparability by examining sales at different LOTs in the comparison market. Any price effect must be manifested in a pattern of consistent price differences between foreign market sales used for comparison and foreign market sales at the LOT of the export transaction. To quantify the price differences, we calculated the difference in the average of the net prices of the same models sold at different LOTs. We used the average difference in net prices to adjust NV when NV was based on a LOT different from that of the export sale. If there was a pattern of no price differences, the differences in LOTs did not have a price effect and, therefore, no adjustment was necessary. Section 773 of the Act provides for an adjustment to NV when NV is based on a LOT different from that of the CEP if the NV level is more remote from the factory than the CEP and if we are unable to determine whether the difference in LOTs between the CEP and NV affects the comparability of their prices (see, e.g., AFBs VII at 31566 and 31572). This latter situation can occur when there is no home market LOT equivalent to the U.S. LOT or where there is an equivalent home market level but the data are insufficient to support a conclusion on price effect. This adjustment, the CEP offset, is identified in section 773(a)(7)(B) of the Act and is the lower of the following:The indirect selling expenses on the home market sale, or The indirect selling expenses deducted from the starting price used to calculate CEP. The CEP offset is not automatically granted each time we use CEP (see, e.g., Notice of Final Determination of Sales at Less Than Fair Value; Certain Cut-to-Length Carbon Steel Plate from South Africa, 62 FR 61731, 61732 (November 19, 1997)). The CEP offset is made only when the LOT of the home market sale is more advanced than the LOT of the CEP sale and there is not an appropriate basis for determining whether there is an effect on price comparability. We determined that for respondents Koyo and NSK there were two home market LOTs and one U.S. LOT (i.e., the CEP LOT). For Fuji we determined that one LOT existed in the home market and three distinct LOTs existed in the U.S. market (the CEP LOT and two EP LOTs). Because there was no home market LOT equivalent to any of the U.S. LOTs for Fuji, NSK, and Koyo, and because NV for these firms represented a price more remote from the factory than the CEP, for these firms we made a CEP offset adjustment to NV in our CEP comparisons (see Certain Internal-Combustion Industrial Fork Lift Trucks from Japan; Final Results of Antidumping Duty Administrative Reviews, 62 FR 5592, 5608 (February 6, 1997)). We determined that for MC a single LOT existed in the third-country market and that a single EP LOT existed in the U.S. market. Based on our comparison of the U.S. EP LOT to the third-country LOT, we determined that the third-country LOT was the same as the EP LOT. As a result, we made no LOT adjustment. For NTN we found that there were three home market LOTs and two (EP and CEP) LOTs in the United States. Because there were no home market LOTs equivalent to NTN's CEP LOT, and because NV for NTN represented a price more remote from the factory than the CEP, we made a CEP offset adjustment to NV in our CEP comparisons. We also determined that NTN's EP LOT was equivalent to one of its LOTs in the home market. Because we determined that there was a pattern of consistent price differences, we made a LOT adjustment to NV for NTN in our EP comparisons where the U.S. EP sale matched to a home market sale at a different level of trade. Comment 31: Koyo, NTN, and NSK contend that the Department's practice with regard to LOTs effectively precludes a LOT adjustment to NV for CEP comparisons and is thus contrary to law and Congressional intent. NSK contends that there is no statutory requirement that a LOT adjustment be based on the full difference in prices between the home market comparison LOT and the HM LOT equivalent to the CEP LOT and suggests that a partial LOT adjustment is contemplated by the statute. NSK contends that the plain reading of the statute requires that the Department must adjust NV for CEP sales for the difference between price levels at the LOTs which do exist in the home market. Therefore, NSK argues the Department should at least make a partial LOT adjustment when comparing NSK's CEP sales to home market aftermarket (AM) sales which, it contends, are more advanced than HM OEM sales because prices are higher at [[Page 2578]] the HM AM LOT. Finally, NSK contends that the Department should grant NSK a partial LOT adjustment equal to the price difference between home market AM sales and OEM sales. Koyo asserts that it and other respondents have proposed to the Department alternative methods by which the Department could construct an appropriate home market LOT by deducting from NV those home market expenses that correspond to the expenses that are deducted from CEP, but that the Department has failed to provide a reasonable explanation for rejecting the proposals. NTN states that the Department should make a price-based LOT adjustment when the LOT of the CEP sale is different from the LOT of the comparison foreign like product, and that the LOT of the CEP sale should be based on the sale to the first unaffiliated U.S. customer prior to the deduction of expenses pursuant to section 772(d) of the Act. NTN asserts that such an approach is not only consistent with the Department's model-match methodology, but evidence on the record demonstrates that NTN's performance of different selling activities at each LOT affected price comparability. NTN argues that it is unreasonable for the Department to refuse to make a price-based adjustment when there are significant differences in prices between home market LOTs and U.S. sales are matched to home market sales at LOTs different than the U.S. sale. Timken contends that under section 773(a)(7)(A)(ii) of the Act, Congress intended for a LOT adjustment to be made only if it was ``demonstrated to affect price comparability, based on a pattern of consistent price differences between sales at different LOT's in the country in which normal value is determined.'' Timken contends that the adjustment cannot be made unless a LOT equivalent to the U.S. LOT exists in the home market. Therefore, Timken claims, if the data available to the Department does not allow the demonstration required by section 773(a)(7)(A)(ii) of the Act, the statute does not permit a LOT adjustment and allows only a CEP offset. Further, Timken argues that NSK's assertion that the Department could have calculated a ``partial LOT adjustment'' for the difference between the CEP LOT and the home market AM LOT on the basis of a consistent pattern of price differences between the home market OEM and AM LOTs is unfounded. Timken contends that where there is no home market LOT comparable to the U.S. LOT, the statute does not authorize the use of price differences between different home market LOTs to substitute for calculated LOT adjustments. As a result, Timken concludes, the Department should reject NSK's claim for the same reasons it rejected the identical argument in AFBs VII at 54043 and 54056-57. Department's Position: We disagree with respondents. Our methodology does not preclude LOT adjustments to NV for CEP sales. Rather, we do not make a LOT adjustment where the facts of the case do not support such an adjustment. Based upon our examination of the information on the record, for this review we found that no respondent had a home market LOT equivalent to its CEP LOT. As a result, because we lacked the information necessary to determine whether there is a pattern of consistent price differences between the relevant LOTs, we did not make a LOT adjustment for any of the respondents when we matched a CEP sale to a sale of the foreign like product at a different LOT. We disagree with NSK that we should make a ``partial LOT adjustment'' because there is no provision in the statute for making such a partial adjustment. We make a LOT adjustment when there is ``any difference between the export price or constructed export price and the [NV] that is shown to be wholly or partly due to a difference in LOT between the export price or constructed export price and the normal value.'' See section 773(a)(7)(A) of the Act. While NSK has interpreted the phrase ``wholly or partly'' to justify a partial LOT adjustment, we interpret this phrase to mean that we may make a LOT adjustment only if part of the differences in prices between LOTs is attributable to the difference in LOT. In other words, we need not demonstrate that no factor other than LOT influenced a pattern of price differences. Thus, we do not read into this language of the statute the authority to make a LOT adjustment between two home market LOTs where neither level is equivalent to the LOT of the U.S. sale. We also disagree with Koyo that we should adopt one of the proposed alternative methods by which we would ``construct'' home market LOTs. We base home market LOTs on a respondent's actual experience in selling in the home market. Therefore, because there is no statutory basis for us to ``construct'' levels in the home market or elsewhere, we have not used Koyo's claimed constructed NV LOT in order to calculate a LOT adjustment for Koyo's CEP sales (see AFBs VII at 54040, 54047). Furthermore, we disagree with NSK that its CEP sales should be matched to its home market OEM sales before they are matched to home market AM sales. Based upon our examination of the information on the record, we found that no home market LOT for NSK had more selling functions than another home market level. Rather, the home market LOTs each involved different degrees of various selling functions. We conclude that, for NSK and for respondents generally, while the reported home market LOTs are different from one another, no home market LOT is more advanced than any other based upon the evidence on the record. We also disagree with NSK's assertion that, because its OEM prices are generally lower than its AM prices, its OEM LOT is less advanced than the distributor/aftermarket LOT. We determine whether one LOT is more advanced than another on the basis of the selling functions performed by a respondent with respect to the two LOTs. NSK's home market OEM and AM sales are more advanced than the LOT of the CEP sales because comparatively fewer selling functions are associated with the CEP sales than are associated with sales to either of the other LOTs. Therefore, we have not altered our LOT methodology. Finally, we disagree with NTN. The definition of ``constructed export price'' contained at section 772(d) of the Act indicates clearly that we are to base CEP on the U.S. resale price, as adjusted for U.S. selling expenses and profit. As such, the CEP reflects a price exclusive of all selling expenses and profit associated with economic activities occurring in the United States. See SAA at 823. These adjustments are necessary in order to arrive at, as the term CEP makes clear, a ``constructed'' EP. The adjustments we make to the starting price, specifically those made pursuant to section 772(d) of the Act (``Additional Adjustments for Constructed Export Price''), normally change the LOT. Accordingly, we must determine the LOT of CEP sales exclusive of the expenses (and associated selling functions) that we deduct pursuant to this section (see, Certain Cold-Rolled Carbon Steel Flat Products from the Netherlands; Final Results of Antidumping Administrative Review, 62 FR 18475, 18480 (April 15, 1997)). As stated earlier, because none of NTN's home market LOTs were equivalent to the LOT of its CEP sales, we were unable to make a LOT adjustment for such sales. Comment 32: NTN contends that the Department should have relied on its U.S. and home market selling expenses, [[Page 2579]] which were based on LOT, as reported, instead of reallocating these selling expenses without regard to LOT. NTN argues that the Department incorrectly relied on the CIT's decision in The Timken Company v. United States, Slip Op. 96-86 (May 31, 1996)(Timken 1) as the basis for its reallocation because the standards set forth in the Timken 1 decision are not only met by NTN's allocated expenses, but its original reporting methodology is less distortive than the Department's reallocation without regard to LOT. NTN further asserts that the Department's reliance on Timken 1 is misplaced due to the fact that the Department has previously indicated that NTN's reporting methodology is within the parameters of the Timken 1 determination. For example, NTN asserts, in 92/93 TRB Final at 57629 and 57636, the Department determined that NTN's LOT-based reporting was not acceptable based ``solely on our discovery of a discrepancy in NTN's reported total U.S. sales value for scope merchandise during the POR.'' NTN maintains that it is clear from the language of the determination that the only reason the Department rejected NTN's reported expenses was an alleged discrepancy in reported numbers. NTN claims that not only is the reporting methodology in this review identical to that in the above- cited final results, but the Department found no discrepancies in this methodology during its U.S. sales verification. In addition, NTN contends that the Department determined that different LOTs existed in the U.S. and Japanese markets for its sales (see TRB Prelim at 47458-9), and that the decision to allocate certain U.S. and home market expenses without regard to LOT voids the LOT determination made in the preliminary results, insofar as the effect of the different LOTs on price is lessened by this reallocation. Furthermore, NTN argues that the Department's mandate is to administer the antidumping laws as accurately as possible (see Bowe-Passat at 335 and 340). Because the Department's reallocation of these expenses without regard to LOT eliminates the affect of LOT on price, NTN asserts, the Department's decision to reallocate these expenses is a direct violation of this mandate. Therefore, NTN concludes, the Department should rely on the LOT-specific expense allocation ratios and its LOT-specific expenses as originally reported in its questionnaire response. Timken contends that in Timken 1 the CIT stated that the issue raised by NTN's LOT-specific expense allocation methodology was ``whether the reported expenses demonstrably vary according to levels of trade.'' Timken argues that while NTN asserts that its LOT-specific allocation methodology meets this standard, NTN provides no explanation on the record of how its methodology met this standard nor is there any other evidence on the record supporting NTN's methodology. Timken further argues that in Timken 1, after identifying the issue in question, the CIT remanded the case to the Department to determine whether NTN had demonstrated that its expenses varied according to LOT. However, Timken states, while the Department was working on its response to that remand, it issued its 1992-93 final results, the final results cited by NTN, in which it rejected NTN's allocation of U.S. expenses due to a discrepancy in sales value. Timken states that it was only after publication of the 1992-93 final results for NTN that the Department completed its remand results pursuant to Timken 1 and determined that the record lacked the evidence necessary to demonstrate that NTN's expenses varied by LOT. See The Department's Final Results of Redetermination Pursuant to Court Remand (December 17, 1996), at 9. Timken contends that, given that these remand results have been affirmed by the CIT (see Timken v. United States, Slip Op. 97-87 (July 3, 1997), and that the Department has requested a remand in the litigation arising from the 1992-93 final results to consider this issue in light of its remand redetermination pursuant to Timken 1 and the CIT's affirmation thereof, the Department correctly rejected NTN's LOT-specific expense allocations in this instant review. Department's Position: We agree with Timken in part. We have determined that, for a majority of the expenses in question, NTN's LOT- specific selling expense allocation methodology bears no relationship to the manner in which NTN actually incurred these selling expenses. In Timken 1 the CIT ordered the Department to accept NTN's LOT-specific allocations and per-unit LOT expense adjustment amounts only if NTN's expenses demonstrably varied according to LOT. By ordering us to ascertain whether these expenses actually varied according to LOT, the CIT, in essence, indicated that NTN's use of its calculation of LOT- specific per-unit expense adjustments did not necessarily mean that NTN incurred the expenses differently due to differences in LOTs. Rather, additional evidence must also exist which demonstrates that NTN actually sold differently to each LOT by performing different activities/functions or by performing the same activities/functions to a different degree when selling to each LOT. In accordance with this order, in our remand results pursuant to Timken 1 we did not allow NTN's allocation of its expenses by LOT due to the lack of quantitative and narrative evidence on the record demonstrating that the expenses in question demonstrably varied according to LOT. In the instant review, we applied the same standards articulated by the CIT in Timken 1. In other words, we have examined the record to determine if evidence exists demonstrating that those home market and U.S. expenses NTN allocated by LOT did demonstrably vary according to LOT. For this review NTN provided two exhibits which outlined its derivation of LOT-specific per-unit expense adjustments for certain of its U.S. and home market expenses. Exhibit C-7 detailed NTN's calculations of LOT-specific per-unit expense adjustment ratios for its U.S. inland freight (warehouse to customer) expenses, other U.S. transportation expenses, U.S. Customs duty, U.S. packing material, overhead, and labor expenses, U.S. advertising expenses, U.S. inventory carrying costs, and other U.S. indirect selling expenses. Exhibit B-4 detailed NTN's LOT-specific per-unit adjustment ratios for its home market pre-sale and post-sale freight expenses, home market advertising expenses, home market packing labor and material expenses, home market technical service expenses, and other home market indirect selling expenses. Both exhibits indicate that, except for certain U.S. and home market packing material and packing labor expenses, none of the expenses were unique to a single LOT in that NTN incurred each of the above expenses when selling to each LOT. However, rather than calculate a single allocation ratio to be applied to all sales, NTN instead allocated a portion of each total expense amount to each LOT such that it was able to derive LOT-specific allocation ratios. When applied to the reported unit prices, NTN's LOT-specific allocation ratios resulted in the calculation of significantly different per-unit expense adjustment amounts such that NTN actually reported an expense adjustment amount for a TRB sale to one LOT which was significantly different than the amount of the same expense it reported for a sale of the identical TRB to another LOT. NTN determined the portion of each of the above expenses (except for certain U.S. and home market packing material and labor expenses) to be allocated to [[Page 2580]] each LOT by means of allocation methodologies which were based on (1) the differences in total sales value for each LOT, (2) the differences in the total number of invoices generated for each LOT, (3) the differences in the total number of employees involved in sales at each LOT, or (4) a combination of the above. As a result, these differences caused the differences in the expense amounts NTN reported for each LOT and in its LOT-specific ratios. While the record for these reviews contains detailed worksheets demonstrating NTN's allocation methodologies, it does not contain any narrative or quantitative evidence demonstrating why or to what degree a TRB sale to one LOT would generate a greater or lesser amount of the above expenses than a sale of the same TRB to another LOT. Rather, NTN's sole support for its allocations are the allocations themselves. While we recognize that total sales values, the total number of invoices, and even the total number of employees may vary according to LOT, these aggregate differences do not demonstrate whether NTN sold differently to its LOTs and fail to indicate what activities or functions NTN may have performed differently when selling to each LOT such that it actually incurred per-unit expense amounts differently due to differences in LOTs. The record, therefore, lacks the evidence necessary to demonstrate that all of NTN's expenses varied according to LOT. Therefore, for these final results, we have not accepted NTN's LOT-specific allocations and its use of LOT-specific adjustment ratios for its U.S. inland freight (warehouse to customer) expenses, other U.S. transportation expenses, U.S. Customs duty, U.S. advertising expenses, U.S. inventory carrying costs, and other U.S. indirect selling expenses, or for its home market pre-sale and post-sale freight expenses, home market advertising expenses, home market technical service expenses, and other home market indirect selling expenses. Rather, we have recalculated NTN's allocation ratios such that we derived a single ratio applicable to all sales regardless of LOT. We then applied these recalculated allocation ratios to NTN's reported U.S. and home market unit prices to calculate per-unit expense adjustment amounts which did not vary by the LOT to which the U.S. or home market sale was made. We disagree with Timken that all of NTN's U.S. and home market expenses should be recalculated without regard to LOT. In our preliminary analysis memorandum (see Preliminary Analysis Memorandum for NTN, September 2, 1997, attachments I and II), we did, in fact, recalculate NTN's U.S. selling expenses without regard to LOT. However, in contrast to the above, for certain of NTN's U.S. packing material and packing labor expenses, exhibit C-7 of NTN's response indicated that NTN incurred these expenses only when selling to one specific U.S. LOT. In addition, NTN's narrative explanation clearly indicated that certain of NTN's packing expenses individually differed by LOT. Because these expenses were unique to a single LOT, NTN (1) allocated each total expense amount solely to this LOT, (2) calculated a single allocation ratio for this LOT, and (3) applied this ratio only to those U.S. sales at this LOT. NTN's response clearly indicates that these expenses demonstrably varied according to LOT (see NTN questionnaire response, January 27, 1997, at exhibit C-7) (NTN Response). Furthermore, in the instant review, we verified these expenses in detail and concluded that NTN's allocation methodology regarding U.S. packing material and U.S. packing labor was accurate (see NTN U.S. Report, at 13). Therefore, for our preliminary results we applied our recalculated ratios for certain of NTN's U.S. packing and U.S. labor expenses only for sales to the one LOT for which these expenses were incurred. In addition, after further review of the record, we have also determined that NTN's home market packing labor and packing material expenses demonstrably varied according to LOT. Section A and exhibit B- 4 of NTN's response clearly demonstrate that different methods of packing are required depending upon LOT. As indicated above, NTN has allocated all of its home market expenses by LOT, but has not provided record evidence (except for home market packing) demonstrating that they were incurred differently by LOT. Therefore, for these final results we have only accepted NTN's allocation for home market packing expenses according to LOT. Therefore, with the exception of NTN's home market and U.S. packing expenses, due to the lack of quantitative and narrative evidence on the record demonstrating that certain of NTN's expenses demonstrably varied according to LOT, for these final results we have reallocated these expenses without regard to LOT. Arm's Length Test Comment 33: NTN asserts that the Department's 99.5 percent arm's- length test is not a reasonable basis for determining whether affiliated-party sales were at prices comparable to those to unaffiliated parties. NTN argues that in applying the arm's-length test the Department only considers the average percentage difference in pricing between affiliated and unaffiliated-party sales and ignores other factors which greatly influence price such as the terms and quantities of each affiliated-party sale. NTN further contends that the Department's 99.5 percent threshold is not really a ``test'', since it fails to provide an objective standard to determine whether affiliated sales are at arm's-length. Instead, NTN claims, the test weighs sales against an average which does not reflect the full range of prices paid in the transactions examined. Therefore, NTN asserts, the use of the 99.5 percent figure as a baseline to decide if sales are at arm's length does not address the fact that some arm's-length sales fall outside this narrow range. As a result, NTN claims, the percentage used would better reflect the range of arm's-length prices if it were lowered to a 95 percent threshold. Timken claims that in accordance with section 773(a)(1)(B) of the Act, the Department properly excluded those home market sales to affiliated parties which were not at arm's length. Timken argues that not only is it wholly within the Department's discretion to derive a methodology to determine whether home market sales to affiliates are at arm's length, but NTN has provided no evidence supporting its claim that the Department's 99.5 test was contrary to law. In addition, Timken points out, the record indicates that one of the factors suggested by NTN for inclusion in the 99.5 percent test, terms of sale, was reported the same for all of NTN's home market sales. Thus, Timken concludes, even if the Department agreed with NTN, the adoption of NTN's suggestion would have no effect. Department's Position: We disagree with NTN. Our 99.5 percent arm's-length test is a reasonable method for establishing a fair basis of comparison between affiliated and unaffiliated-party sales. NTN asserts that additional factors, such as quantity and payment terms, should be taken into consideration when comparing affiliated and unaffiliated-party sales, but fails to establish that the Department must abandon its existing test. NTN also argues that our use of the 99.5 percent threshold is distortive but provides no quantitative evidence demonstrating that a lowering of the threshold would yield more accurate results. Furthermore, the CIT has upheld the validity of our arm's-length test on numerous occasions. For example, in Usinor Sacilor v. United States, 872 F. [[Page 2581]] Supp 1000 (1994), the CIT clearly stated that ``[g]iven the lack of evidence showing any distortion of price comparability, the court finds application of Commerce's arm's-length test reasonable.'' Likewise, in Micron Technology, Inc. v. United States, 893 F. Supp 21, 38 (CIT 1995), because the CIT found that the plaintiff/respondent failed to ``demonstrate that Commerce's customer-based arm's-length is unreasonable'' and failed to ``point to record evidence which tends to undermine Commerce's conclusion,'' the CIT sustained the 99.5 percent arm's-length test, given a lack of evidence showing a distortion of price comparability. Further, in NTN Bearing Corp. of America, American NTN Bearing Manufacturing Corp., and NTN Corp. v. United States, 905 F. Supp. 1083 (CIT 1995), NTN argued, as here, that there were numerous factors influencing the price of a related-party transaction and the Department cannot make a meaningful price comparison without examining them. The CIT disagreed with NTN and stated that, in accordance with section 19 CFR 353.45(a) of our regulations, the Department has broad discretion in devising an appropriate methodology to determine whether particular related-party prices are, in fact, comparable to unrelated- party prices. Therefore, because NTN has failed to demonstrate that the 99.5 percent threshold produces distortive results and that the Department's methodology is unreasonable, in accordance with the CIT decisions cited above, we have not altered our 99.5 percent arm's-length test for these final results. Sample Sales On June 10, 1997, the CAFC held that the term ``sold'' requires both a transfer of ownership to an unrelated party and consideration. NSK Ltd. v. United States, 115 F.3d 965, 975 (Fed. Cir. 1997) (NSK). The CAFC determined that samples which NSK had given to potential customers at no charge and with no other obligation lacked consideration. Moreover, the CAFC found that, since free samples did not constitute ``sales,'' they should not have been included in calculating U.S. price. In light of the CAFC's opinion, we have revised our policy with respect to samples. The Department will now exclude from its dumping calculations sample transactions for which a respondent has established that there is either no transfer of ownership or no consideration. This new policy does not mean that the Department automatically will exclude from analysis any transaction to which a respondent applies the label ``sample.'' In fact, for these reviews we determined that there were instances where it is appropriate not to exclude such alleged samples from our dumping analysis. It is well-established that the burden of proof rests with the party making a claim and in possession of the needed information (see, e.g., NTN Bearing Corporation of America v. United States, 997 F.2d 1453, 1458-59 (CAFC 1993), (citing Zenith Elecs. Corp. v. United States, 988 F.2d 1573, 1583 (CAFC 1993), and Tianjin Mach. Import & Export Corp. v. United States, 806 F. Supp. 1008, 1015 (CIT 1992)). As discussed below, one respondent failed to demonstrate that its claimed sample sales lacked consideration. When respondents failed to support their sample claim, we did not exclude the alleged samples from our margin analysis. With respect to HM sales, in addition to excluding sample transactions which do not meet the definition of ``sales,'' we may exclude sales designated as samples from our analysis, pursuant to section 773(a)(1) of the Act, when a respondent has provided evidence demonstrating that the sales were not made in the ordinary course of trade, as defined in section 771(15) of the Act. With regard to assessment rates, in order to ensure that we collect duties only on sales of subject merchandise, we included the entered values and quantities of the sample transactions in our calculation of the assessment rates and set the dumping duties due for such transactions to zero. We have done this because U.S. Customs will collect the ad valorem (or per-unit, where applicable) duties on all entries of subject merchandise whether or not the merchandise was a sample transaction. However, to ensure that sample transactions do not dilute the cash deposit rates, we excluded both the calculated U.S. prices and quantities for sample transactions from our calculation of the cash deposit rates. Comment 34: Timken argues that for these final results the Department should include in NSK's U.S. database its zero-priced sample sales. Timken contends that although the CAFC's decision in NSK held that zero-priced sample sales which lacked consideration did not constitute ``sales'' for purposes of the antidumping law, the decision did not establish a per se exclusion for all zero-priced sample sales. Timken argues that such sales do not qualify for automatic exclusion from the U.S. database because the burden is on the respondent to demonstrate that sample sales did not involve the transfer of ownership or that they lacked consideration. Timken maintains that NSK did not provide information for the record affirmatively demonstrating that its U.S. sample sales were transferred without consideration or ownership. Timken further argues that the CIT in J.C. Hallman Mfg. Co. v. United States, 13 CIT 1073, 1076, 728 F. Supp. 751, 753 (1989) (J.C. Hallman) stated that samples must be reported under a ``temporary importation bond.'' Timken asserts that the CIT in that case also held that in the absence of such a bond, the Department has no way of knowing that the merchandise is not imported for sale. Timken contends that because NSK has provided no information demonstrating whether its zero-priced sample sales were imported under a temporary importation bond, the Department should reverse its preliminary determination and include NSK's zero-priced sample sales in its margin calculations for these final results. NSK responds that the Department correctly excluded zero-priced U.S. sample sales from its analysis. NSK contends that Timken's reliance on J.C. Hallman is misplaced because this case predated the court's NSK decision and because the CIT, in its omission of any reference to J.C. Hallman in its decision, effectively determined that the case was irrelevant for its decision. Furthermore, NSK argues, the only standard set forth by the CAFC in NSK is whether a sale occurred (i.e., involved consideration). NSK contends that as long as sample sales lacked consideration, then all other issues, such as whether the recipient took title to the merchandise, are irrelevant. NSK further argues that it reported its free samples as outside the ordinary course of trade and indicated that zero-priced samples were not sales because they lacked consideration. Because the Department did not ask any questions regarding the company's sample sales in its supplemental questionnaire, NSK argues, the Department concluded that it had all necessary information to determine whether or not zero-priced sample sales should be considered ``sales'' for purposes of its analysis. Department's Position: We disagree with petitioner. The record indicates that NSK's reported sample transactions did not involve consideration (see, e.g., NSK Section C Questionnaire Response, January 27, 1997, at 4). Accordingly, pursuant to the CAFC's decision in NSK, we have excluded NSK's reported U.S. sample sales from the U.S. sales database. Comment 35: NTN argues that the Department should exclude from its margin calculations those sample sales [[Page 2582]] it made in both the U.S. and home markets. With respect to its home market database, NTN asserts that its home market sample sales which it claims are outside the ordinary course of trade should be excluded from margin calculations in accordance with section 773(b) of the Act and in accordance with the CIT's decision in NSK v. United States, Slip Op. 97-74 (June 17, 1997), in which the CIT held that the Department improperly included NTN's sample sales. NTN also asserts that its U.S. sample sales should be excluded from the Department's analysis in accordance with the CAFC's ruling in NSK, in which the CAFC ordered that zero-priced sample sales be excluded for purposes of calculating margins. Timken responds that the CAFC in NSK did not establish a per se exclusion for so-called sample sales. Rather, Timken claims, the CAFC held that sales which lacked consideration did not constitute sales for purposes of the antidumping law. Timken notes that the Department's preliminary margin program at lines 92 and 704 already excludes zero- priced sales, and claims that the NSK decision does not support the exclusion of sales NTN alleges are samples. Finally, Timken argues that NTN has not adequately demonstrated that its home market sample sales are outside the ordinary course of trade, and that such sales therefore do not warrant exclusion from the home market database. Department's Position: We disagree with NTN. We examined the record to determine whether NTN's U.S. sample sales lacked consideration, and were unable to find any information whatsoever in either NTN's narrative or sales database regarding sample transactions. As noted above, the party in possession of the information has the burden of producing that information, particularly when seeking a favorable adjustment or exclusion. Because NTN did not provide any information in its response or elsewhere that would have aided us in determining whether NTN received a bargained-for exchange from its U.S. customers, we cannot conclude that NTN received no consideration for these alleged samples. While NTN's database does include sales which are zero-priced, we are unable to determine from the record if these transactions represent those sales which NTN apparently argues should be excluded from the U.S. database in accordance with NSK. Furthermore, the mere fact that a sale has a reported unit price of zero does not indicate that a transaction lacked exchange of consideration. Our preliminary margin program incorporated language to exclude all zero-priced sales in the home and U.S. markets. However, for the reasons stated above, we have altered our treatment of NTN's zero-priced U.S. sales and have included them in NTN's U.S. database for these final results. NTN also argues that we should exclude its alleged home market sample sales from its home market sales database. As noted previously, one of the circumstances under which we may exclude sample sales from the home market database is when a respondent has demonstrated that such sales were made outside the ordinary course of trade. Accordingly, we have examined the record with respect to NTN's alleged home market sample sales to determine if these sales qualify for such an exclusion. In its original questionnaire response NTN only states that ``samples are provided to customers for the purpose of allowing the customer to determine whether a particular product is suited to the customer's needs'' and that ``the purpose * * * would not be the same as those purchased in the normal course of trade'' (see, NTN Response at B-15). NTN has provided no other information demonstrating that its alleged home market sample sales were outside the ordinary course of trade. The fact that a respondent identified sales as samples does not necessarily render such sales outside the ordinary course of trade (see AFBs VI at 2124). For these reasons, we disagree with NTN that its home market ``sample'' sales should be excluded from our margin calculations. We have also evaluated whether NTN's alleged home market sample sales qualify for exclusion from the home market database in light of the NSK decision. As noted above, we exclude sample transactions from the dumping calculation only if a respondent has demonstrated that there is either no transfer of ownership or no consideration. Evidence on the record clearly indicates that NTN received consideration for all home market sales it claims are samples. As such, none of its home market sample sales meet the criterion for exclusion established by NSK. Therefore, because NTN's alleged U.S. and home market sample sales do not qualify for exclusion under NSK, and because NTN has failed to demonstrate that its home market sample sales are outside the ordinary course of trade, we have included these sales in our U.S. and home market databases for these final results. Comment 36: NSK argues that Timken's general issues should be stricken from the record because the petitioner failed to include these arguments in the case briefs it served to respondents. NSK contends that because the general issues section is free of proprietary material, it should have been served with the proprietary portion of Timken's brief rather than one day later. NSK claims that the Department should not allow Timken to abuse the ``one-day lag'', for the purpose of the rule is to permit counsel the opportunity to review proprietary portions of submissions and to confirm that (a) all proprietary information has been properly bracketed, and (b) that the public version correctly removes, ranges or indexes the proprietary information. 19 CFR 353.32(a)(2). Therefore, NSK asserts, because neither of these purposes is served by Timken's decision to withhold an entire portion of its case brief, the Department should reject the general issues portion of Timken's case brief. Department's Position: During our October 30, 1997 TRB hearing NSK raised these concerns. After adjourning to review the details of Timken's brief and the issues raised by NSK, we determined that Timken improperly served the general issues portion of its case brief to the other parties to this proceeding but nevertheless properly filed its brief with the Department. After further discussion with the parties in attendance we found that NSK, NTN, and Fuji all responded to Timken's general comments section in their rebuttal briefs, but that Koyo had not. Therefore, with the agreement of the parties in attendance, because Koyo did not have the opportunity to rebut this section of Timken's brief due to the service of the brief, we granted Koyo an additional week to respond to the general issues section of Timken's case brief and allowed Timken's general comments to remain part of the administrative record. Clerical Errors Comment 37: NSK argues that, when calculating home market net prices, rather than deducting NSK's reported REBATE1H, the Department incorrectly added these rebates to home market gross unit price. Timken states that, to the extent that the Department intended to deduct NSK's rebates when calculating NV, it agrees that the rebates were improperly added to gross price. Department's Position: We agree with NSK and have amended our computer program for these final results such that NSK's reported home market rebates are subtracted from, rather than added to, home market gross unit prices. Comment 38: NSK and Koyo assert that, when calculating CEP profit, the [[Page 2583]] Department incorrectly based its derivation of total home market revenue on gross home market prices rather than on home market prices net of discounts and rebates. Timken agrees that the calculation of home market revenues should be based on net price. Department's Position: We agree. The Department's September 4, 1997 policy bulletin regarding the calculation of CEP profit clearly indicates that total home market sales revenue should be calculated net of home market discounts and rebates. Therefore, for these final results we have adjusted our calculation of NSK's and Koyo's home market revenue such that our computer programs calculate home market revenues net of rebates and discounts. In addition, while NTN did not comment on this issue, we note that we made the identical error in our preliminary results computer program for NTN. Therefore, to ensure the calculation of the most accurate final results margin for NTN, we have corrected this error in our computer program for NTN as well. We also note that, while reviewing our preliminary results calculation of CEP profit for each of the respondents, we discovered that we inadvertently made an additional error. After calculating total actual profit and deriving a profit ratio, we multiply this ratio by the respondent's total U.S. selling expenses. Our September 4, 1997 policy bulletin clearly states that ``when allocating a portion of the actual profit to each U.S. CEP sale, we will include imputed (U.S.) credit and inventory carrying costs as part of the total U.S. expenses allocation.'' However, in our preliminary results computer programs we inadvertently excluded U.S. credit and inventory carrying costs from our calculation of the U.S. selling expenses upon which profit was allocated. Therefore, although no party to this proceeding commented on this issue, to ensure the calculation of accurate margins we have nevertheless corrected this error, where appropriate, for these final results. Comment 39: NSK argues that, although it is the Department's long- standing policy when calculating CV to deduct credit from CV as a home market circumstance-of-sale (COS) adjustment and to deduct ICC as part of the CEP offset, the Department's preliminary results computer program for NSK did not make these adjustments. NSK contends that not only should the Department make these adjustments for these final results, but when deriving the expense ratios for credit and ICC, the Department should ensure that these ratios are calculated on the same basis as the value to which the ratios are applied. Timken asserts that these imputed credit and inventory expenses are already included in the Department's calculation of CV as part of SG&A and that to add them again would result in double counting. Department's Position: We agree with NSK. When calculating CV in our preliminary results computer program for NSK we inadvertently failed to make a COS adjustment to CV for NSK's reported home market credit expenses and failed to deduct ICC from CV as part of the CEP offset. Therefore, for these final results we have modified certain language within our computer program to ensure that these deductions are made when we calculated NV using CV. In addition, in order to derive the actual credit and ICC amounts used in our CV calculation, we calculated our home market credit and ICC ratios on the same basis as the value to which we applied these ratios. Furthermore, while only NSK commented on this issue, we have determined that we made the identical error in our preliminary results computer programs for NTN and Koyo. Therefore, to ensure the calculation of accurate final results margins for these two respondents, we have corrected this error in our computer programs for NTN and Koyo as well. Comment 40: NSK contends that the Department improperly downloaded NSK's U.S. computer data by failing to define the Y2FACTU variable as having two decimal places. As a result, NSK asserts, the U.S. Y2 factors used by the Department in its preliminary results model-match for NSK erroneously relies on a U.S. Y2 factor which is overstated by 100. Timken agrees that NSK's U.S. Y2 factor variable appears to lack decimal places. Department's Position: We agree with NSK. However, rather than re- downloading NSK's U.S. data to correct this error, for these final results we have corrected this error by dividing all of NSK's U.S. Y2 factors within our database by 100 prior to conducting our model matches. Comment 41: Timken and NSK assert that an error exists in the Department's preliminary results computer programs which causes certain U.S. sales to be matched with the second or third most similar foreign like product in those instances where the identical or most similar foreign like product was determined to be below COP. Timken and NSK argue that, because it is the Department's long-standing practice to base its calculation of NV on CV whenever the identical or most similar foreign like product is below cost, for the final results the Department should correct this error such that whenever contemporaneous sales of an identical or most similar foreign like product is determined to be below COP, the computer program calculates NV on the basis of CV rather than continuing the search for a contemporaneous match of the next most similar foreign like product. Department's Position: We agree with Timken and NSK. Therefore, for these final results we have modified our multi-level array model-match computer programming language to correct this error and to ensure that all sales of a U.S. model for which the identical or most similar foreign like product is below COP are compared to CV. Comment 42: NTN argues that the Department's preliminary results computer program contains two errors which should be corrected for the final results. First, NTN claims, when creating the data sets NEGDATA1, HMREL, and HMUNREL from the data set HMOVER, the Department's computer program for NTN drops several observations which should have not been excluded from the margin calculations. Likewise, NTN argues, when the Department created the data sets HMSETS, HMCUPS, and HMCONES from the data set HMMM, the computer program again dropped several observations which should not have been excluded from the margin calculations. While Timken does not specifically agree or disagree with NTN's clerical error allegations, with respect to NTN's first alleged error it notes that the Department's computer programming language causes sales observations with a customer relationship code other than 1 or 2 to be excluded from the Department's calculations. Similarly, Timken notes that, with respect to NTN's second alleged error, the Department's computer programming language results in observations for which the home market part type was reported as other than 1, 2, or 3 also to be excluded from the margin calculations. Department's Position: With respect to NTN's first alleged error, we agree. In preparation for our arm's-length test we divided NTN's home market sales (data set HMOVER) into two groups on the basis of whether the sale was made to an affiliated or unaffiliated customer (data sets HMREL and HMUNREL). In our questionnaire we asked respondents to identify for each home market sale whether it was to an affiliated or unaffiliated customer, using a code of ``1'' for unaffiliated customers and a code of ``2'' for affiliated customers. [[Page 2584]] While our questionnaire does not instruct respondents to use any additional codes, NTN nevertheless separately identified its sales to home market affiliated customers which were consumed rather than resold using a code of ``3.'' In our preliminary results computer program we inadvertently excluded the code of ``3'' from the programming language we used to separate home market sales into the affiliated and unaffiliated sales groups. Therefore, for these final results we have corrected this error by identifying all home market sales to affiliated customers by means of both codes ``2'' and ``3.'' With respect to NTN's second alleged error, we disagree that the discrepancy NTN notes is an error. In our TRB questionnaire we ask respondents to identify TRB sets, cups, cones, and parts using numerical codes (``1'' for sets, ``2'' for cups'', ``3'' for cones, and ``4'' for parts), and we used these numerical codes when we created the data sets HMSETS, HMCUPS, and HMCONES in our computer program. The sales NTN identifies as being incorrectly excluded from the margin calculations were sales of home market parts (code ``4''). We did not create a separate HMPARTS data set and did not retain these sales in our margin calculation because we had already determined that NTN did not make any sales of TRB parts in the United States. Because our TRB model-match methodology does not permit the comparison of U.S. TRB sets, cups and cones to home market parts (we only match U.S. TRB sets to home market sets, U.S. cups to home market cups, U.S. cones to home market cones, and U.S. parts to home market parts), and because there were no U.S. sales of TRB parts, it was unnecessary for us to retain NTN's reported sales of home market TRB parts (code ``4'') in our data base. Therefore, because NTN's sales of home market TRB parts were not needed for comparison purposes, our exclusion of these sales from the margin calculations was appropriate and does not constitute an error as NTN alleges. Comment 43: Timken argues that, while the Department's computer program for Koyo properly sets the inside diameter (ID) for home market TRB cups and the outside diameter (OD) for home market TRB cones to zero, the program fails to do the same for Koyo's reported U.S. sales of TRB cups and cones. Timken asserts that, because the inside and outside diameters are two of the five physical criteria relied upon in the Department's model-match methodology, this error will cause distortions when the Department matches U.S. sales to sales of the foreign like product. Koyo contends that it is unnecessary for the Department to purposely set the ID for its reported U.S. TRB cups and the OD for its reported U.S. TRB cones to zero. Koyo argues that, regardless of whether there is an erroneous ID or OD reported for a U.S. TRB cup or cone, the Department's computer program nevertheless ranks the home market foreign like products for each U.S. model accurately. Department's Position: We disagree with Timken that distortions will result because the computer program does not set the OD for U.S. cones and the ID for U.S. cups to zero. Two of the physical criteria for TRB sets are the ID and OD. The ID reflects the measure of the TRB cone while the OD reflects the measure of the TRB cup. While a TRB set, which contains both a cup and cone, has both an ID and OD measurement, individually sold TRB cups do not have an ID and individually sold TRB cones do not have an OD. As a result of our home market set-splitting methodology, in which we derive separate cup and cones sales from the respondents' reported home market TRB set sales, it is necessary for us to purposely set the ID for split cups to zero and the OD for split cones to zero. In the past, we have found it unnecessary to include similar programming language with respect to a respondent's U.S. sales because we do not split U.S. sets into individual cup and cone sales. Timken's comments reflect its concern that, if a respondent incorrectly reports an ID value greater than zero for any U.S. cups and an OD value greater than zero for any U.S. cones, the Department's programming language would result in inaccurate model matching. Therefore, for these final results we have examined whether Koyo reported any ID values for its U.S. cups or OD values for its U.S. cones which were greater than zero. We found that Koyo had indeed reported values greater than zero for both the OD of its U.S. cones and the IDs of its U.S. cups. As a result, we have set the value of any positive inside cup diameters or positive outside cone diameters to zero in Koyo's U.S. summary sales databases. Final Results of Reviews Based on our review of the arguments presented above, for these final results we have made changes in our preliminary margin calculation programs. We determine that the following percentage weighted-average margins exist for the period October 1, 1995 through September 30, 1996: ------------------------------------------------------------------------ Margin Manufacturer/exporter/reseller (percent) ------------------------------------------------------------------------ For the A-588-054 Case:..................................... Koyo Seiko................................................ 9.60 Fuji...................................................... .34 NSK....................................................... 1.45 MC International.......................................... 1.92 For the A-588-604 Case:..................................... Fuji...................................................... (\1\) MC International.......................................... (\1\) Koyo Seiko................................................ 29.02 NTN....................................................... 27.80 NSK....................................................... 9.60 ------------------------------------------------------------------------ \1\ No shipments or sales subject to this review. These firms have no rate from any prior segment of this proceeding. The Department shall determine, and the Customs Service shall assess, antidumping duties on all appropriate entries. We will calculate importer-specific ad valorem duty assessment rates for the merchandise based on the ratio of the total amount of antidumping duties calculated for the examined sales made during the POR to the total customs value of the sales used to calculate those duties. This rate will be assessed uniformly on all entries that a particular importer made during the POR. (This is equivalent to dividing the total amount of antidumping duties, which are calculated by taking the difference between NV and U.S. price, by the total U.S. price of the sales compared and adjusting the result by the average difference between U.S. price and customs value for all merchandise examined during the POR.) While the Department is aware that the entered value of sales during the POR is not necessarily equal to the entered value of entries during the POR, use of entered value of sales as a basis of the assessment rate permits the Department to collect a reasonable approximation of antidumping duties which would have been determined if the Department had reviewed those sales of merchandise during the POR. The Department will issue appropriate appraisement instructions directly to the Customs Service. Furthermore, the following deposit requirements will be effective after the publication date of these final results for all shipments of TRBs from Japan entered, or withdrawn from warehouse, for consumption on or after the publication date of the final results of these administrative reviews, as provided by section 751(a)(1) of the Act: (1) The cash deposit rates for the reviewed companies will be those rates established in the final results of these reviews; (2) For previously reviewed or investigated companies not listed above, the cash deposit rate will continue to be [[Page 2585]] the company-specific rate published for the most recent period; (3) If the exporter is not a firm covered in these reviews, a prior review, or the less-than-fair-value (LTFV) investigations, but the manufacturer is, the cash deposit rate will be the rate established for the most recent period for the manufacturer of the merchandise; and (4) If neither the exporter nor the manufacturer is a firm covered in these or any previous reviews conducted by the Department, the cash deposit rate for the A-588-054 case will be 18.07 percent, and 36.52 percent for the A-588-604 case (see Preliminary Results of Antidumping Duty Administrative Reviews; Tapered Roller Bearings, Finished and Unfinished, and Parts Thereof, from Japan and Tapered Roller Bearings, Four Inches or less in Outside Diameter, and Components Thereof, From Japan, 58 FR 51061 (September 30, 1993)). The cash deposit rate has been determined on the basis of the selling price to the first unaffiliated U.S. customer. For appraisement purposes, where information is available, the Department will use the entered value of the merchandise to determine the assessment rate. This notice serves as a final reminder to importers of their responsibility to file a certificate regarding the reimbursement of antidumping duties prior to liquidation of the relevant entries during this review period. Failure to comply with this requirement could result in the Secretary's presumption that reimbursement of antidumping duties occurred and the subsequent assessment of double antidumping duties. These administrative reviews and this notice are in accordance with section 751(a)(1) of the Act (19 U.S.C. 1675(a)(1)) and 19 CFR 353.22. This notice also serves as a reminder to parties subject to administrative protective orders (APO) of their responsibility concerning the return or destruction of proprietary information disclosed under APO in accordance with 19 CFR 353.34(d) or conversion to judicial protective order is hereby requested. Failure to comply with the regulations and terms of an APO is a violation which is subject to sanction. These administrative reviews and this notice are in accordance with section 751 (a)(1) of the Act (19 U.S.C. 1675(a)(1) and 19 CFR 353.22. Dated: January 7, 1998. Robert S. LaRussa, Assistant Secretary for Import Administration. [FR Doc. 98-944 Filed 1-14-98; 8:45 am] BILLING CODE 3510-DS-P