Memorandum Opinion
Goldberg, Judge:This action originally came before the court on motions for judgment upon the agency record brought by plaintiffs Outokumpu Copper Rolled Products AB (“OAB”) and Outokumpu Copper (USA), Inc. (“OCUSA”), and Defendant-intervenors Hussey Copper, Ltd., et al.1 The parties challenged the final results of the second and third administrative reviews of imports of brass sheet and strip from Sweden by the United States Department of Commerce, International Trade Administration (“Commerce”). Brass Sheet and Strip from Sweden, 57 Fed. Reg. 2706 (Jan. 23, 1992) (“Final Results”). By order dated August 12, 1993, the court remanded this case to Commerce for reconsideration. The results of Commerce’s remand determination are presently before the court for review.
Background
In Outokumpu Copper Rolled Products AB v. United States, 17 CIT 848, 829 F. Supp. 1371 (1993), this court sustained Commerce’s determination in part, but remanded four parts to Commerce for reconsideration. The court instructed Commerce to:
1. Provide an explanation for its decision to eliminate model matching criteria where exact matches were unavailable, and its *205decision not to utilize defendant-intervenors’ proposed methodology where exact matches were unavailable (i.e. to retain each criterion and compare products with the next most similar characteristics within each grouping);
2. Specifically identify, for both the preliminary and final review results, all adjustments made to U.S. price and foreign market value, and the specific statutory and regulatory provisions supporting those adjustments;
3. Recalculate U.S. price without a reduction for U.S. inland freight charges for which Commerce had used best information available (“BIA”) in the underlying reviews; and
4. Correct ministerial errors that resulted in Commerce’s failure to make proper adjustments to foreign market value for differences in the physical characteristics of the U.S. and home market merchandise. .
On September 27,1993, Commerce filed with this court its Results of Redetermination Pursuant to Court Remand (“Remand Results”). Because the court finds that Commerce’s determinations with respect to these four issues are based on substantial evidence and in accordance with law, Commerce’s Remand Results are affirmed.
Discussion
Commerce’s final results filed pursuant to a remand will be sustained unless that determination is “unsupported by substantial evidence on the record, or otherwise not in accordance with law.” 19 U.S.C. § 1516a (b)(1)(B) (1988). Substantial evidence is “such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.” Consolidated Edison Co. v. NLRB, 305 U.S. 197, 229 (1938); N.A.R., S.p.A. v. United States, 14 CIT 409, 412, 741 F. Supp. 936, 939 (1990).
1. Commerce’s Product Comparison Methodology:
The court’s remand instructions noted that due to the brevity of Commerce’s explanation for its decisions to exclude certain groupings in its model match methodology, the court was unable to ascertain the reasons supporting Commerce’s decision, and whether those grounds were reasonable. Outokumpu, 829 F. Supp. at 1378. The court did not specify that Commerce alter its analysis on remand. Rather, the court afforded Commerce the opportunity to provide a more detailed explanation for its actions below. Id.
In establishing the foreign market value (“FMV”) of merchandise imported to the United States, Commerce is directed by 19 U.S.C. § 1677b(a)(l)(A) (1988) to restrict its comparison only to identical (i.e. “such”) or “similar” merchandise that is sold in the home market. In the underlying administrative reviews, Commerce utilized a four-criteria methodology for matching U.S. and home market merchandise; these criteria, in order of importance, are: the class or form of the product (sheet or strip); the metal alloy; the thickness, or gauge, of the product; and, the width of the product. If identical matches based on all four criteria were not available, Commerce eliminated the width characteristic *206and sought a match using the three remaining criteria. Similarly, if no match was found with the three remaining criteria, Commerce then attempted to establish product similarity based only on the criteria of class and alloy. Outokumpu, 829 F. Supp. at 1377.
Defendant-Intervenors objected to Commerce’s decision to sequentially eliminate the width and then the gauge matching criteria when exact matches using all four product characteristics were not possible. They asserted that Commerce’s methodology prevented Commerce from choosing the “most similar” products for comparison. Defendant-Intervenors claim that Commerce should have compared products with the next most similar widths, rather than eliminate that entire criterion altogether. Defendant-Intervenors’Memorandum of Points and Authorities in Support of Motion for Judgment Upon the Agency Record (“Defendant-Intervenors’Memorandum”) at 19.
In explaining its actions, Commerce stated in its Final Results:
The model match method the Department used for these reviews is consistent with the one used for the original investigation and for the first review. Neither respondent nor petitioners raised any objection to the model match until after the Department issued its preliminary results. Additionally, petitioners have not provided any compelling reasons why the Department should now change its model match method for these reviews.
Final Results, 57 Fed. Reg. at 2708.
In its Remand Results, Commerce explained in greater detail the rationale for its product matching methodology and concluded that the methodology chosen was “the most practical means of achieving reasonable comparisons of similar merchandise.” Remand Results at 5. Commerce noted that according to respondents’ data, the differences in dimensions and costs among the width and gauge groupings were minimal relative to the total cost of producing the subject merchandise. Because the differences among the various width and gauge groupings were not substantial, Commerce concluded that a weighted-average of prices at all widths or gauges would not differ substantially from a weighted-average FMV based on a single width or gauge grouping in the absence of identical matches. Id. at 5-6.
Commerce acknowledged that, in some instances, the record lacks information regarding differences in the physical characteristics of the U.S. and home market merchandise, and that it erred in failing to request this information. Remand Results at 6. Nevertheless, Commerce concluded that, because the record indicated no reason to believe that the cost differences between various width and gauge groups would have been substantial, an adjustment for differences-in-merchandise “would not likely have had a substantial impact on the calculation of respondents’ dumping margins.” Id.
Commerce is afforded broad discretion in selecting the methodologies it employs in the administration of the antidumping laws. NTN Bearing Corp. of America v. United States, 14 CIT 623, 633, 747 F. Supp. 726, 736 *207(1990). Commerce’s methodology need not be the most reasonable methodology available. Rather, Commerce’s product comparison will be upheld as long as it is reasonable. Koyo Seiko Co. v. United States, 16 CIT 539, 542, 796 F. Supp. 1526, 1529 (1992). Indeed, courts may “‘uphold [an agency’s] decision of less than ideal clarity if the agency’s path may reasonably be discerned.’” Ceramica Regiomontana, S.A. v. United States, 5 Fed. Cir. (T) 77, 78, 810 F.2d 1137, 1139 (1987) (quoting Bowman Transp., Inc. v. Arkansas-Best Freight Sys., Inc., 419 U.S. 281, 286 (1974)).
The court finds that Commerce’s decision to eliminate the width and gauge matching-criteria was a reasonable exercise of its discretion. Nothing in the record suggests that the differences in production costs associated with the width and gauge of the merchandise would have been substantial, and any adjustment to FMV for differences in these characteristics would not likely have had a substantial impact on the dumping margin calculations. In fact, defendant-intervenors conceded during the administrative hearings that Commerce’s methodology did not necessarily result in comparisons favorable to the respondents. More specifically, defendant-intervenors noted that “[s]ome of these [comparisons] inure to the benefit of the Petitioners; [and] some of these [comparisons] inure to the benefit of Respondents.” Plaintiffs’ Memorandum In Opposition To Defendant-intervenors’Rule 56.1 Motion For Judgment Upon The Agency Record (“Plaintiffs’ Response”) at 22; id. Appendix 11 at 32.
Thus, the court concludes that Commerce reasonably exercised its discretion by eliminating width and gauge matching-criteria when no identical matches were available. The court finds that Commerce’s determination is supported by substantial evidence and has sufficiently articulated a rational connection between the facts found and the choice made; Commerce’s chosen methodology is therefore affirmed.
2. Adjustment of U.S. Price for Commission Related Expenses:
The court directed Commerce on remand to explicitly identify, for both the preliminary and final review results, all adjustments made to U.S. price and foreign market value for commission-related expenses (“commissions”) paid by Outokumpu AB to Outokumpu USA. Out-okumpu, 829 F. Supp. at 1382. The court also requested that Commerce identify the specific statutory and regulatory provisions supporting its determinations regarding these commissions. The court required further explanation in order to determine whether Commerce improperly freed OAB from its burden of showing that the expenses were indirect and whether sufficient evidence supported Commerce’s determination that the commissions paid to Outokumpu USA were, in fact, indirect selling expenses. Outokumpu, 829 F. Supp. at 1381.
In its Remand Results, Commerce first discussed the relevant statutory and regulatory provisions, and Commerce’s practice in determining whether to make a circumstances-of-sale adjustment for com*208missions. Remand Results at 8-13. Commerce then described its treatment of the commissions in question at the preliminary and final administrative reviews. Remand Results at 13-18.
In the preliminary results of both the 1988-89 and 1989-90 administrative reviews, Commerce determined that the respondents’ U.S. commissions were direct selling expenses. Brass Sheet and Strip from Sweden, 56 Fed. Reg. 29,619, 29,620 (June 28, 1991). In the final administrative review results, however, Commerce reversed its preliminary decision and decided not to treat U.S. commissions as direct selling expenses. Brass Sheet and Strip from Sweden, 57 Fed. Reg. 2706, 2707 (Jan. 23, 1992). Commerce determined that the commissions were not direct selling expenses because Commerce was unable to test the arm’s-length nature of the commission arrangement between OAB and OCUSA. Because Commerce was not satisfied that the payments were at arm’s length, Commerce did not make an adjustment for these payments as commissions. Id. Commerce continued, stating:
The Department, however, regards such payments to related parties as indirect selling expenses * * *. We have, therefore, deducted these from the exporter’s sales price (ESP) calculation.
Id. (emphasis added).
Defendant-Intervenors disputed Commerce’s determination that the commissions were indirect, rather than direct, selling expenses. Defendant-Intervenors’ Memorandum at 35. Most of OAB’s sales during the review periods received no downward adjustment to U.S. price, because most of OAB’s U.S. sales were purchase price transactions, rather than ESP transactions. Unlike ESP transactions, purchase price transactions do not merit an adjustment to U.S. price for indirect selling expenses.
In its Remand Results, Commerce explained that the Federal Register notice of final results contained an “incorrect statement” as to Commerce’s characterization of the commissions. Remand Results at 17. Specifically, the final results erroneously stated that Commerce had treated U.S. commissions as indirect selling expenses; in actuality, Commerce “did not treat [the U.S. commissions] as expenses at all.” Id. Consequently, Commerce did not need to make further revisions to its final calculations for FMV because Commerce did not actually make any adjustments or offsets for these commissions in its preliminary calculations.2
Given Commerce’s explanation in its Remand Results, the issue before the court is no longer whether the commissions paid to OCUSA *209were, in fact, indirect selling expenses. Commerce acknowledges that the commissions were inadvertently labeled as indirect expenses, and that despite this mislabeling, no adjustments were actually made for them. The court, however, still must review Commerce’s determination that the commission payments to the related party in the U.S. were not made at arm’s length, and, therefore, were not bona fide selling expenses.
Commerce generally does not make an adjustment for commissions paid to related parties. See, e.g., Porcelain-on-Steel Cooking Ware from Mexico, 51 Fed. Reg. 36,435, 36,438 (Oct. 10, 1986). Commissions paid to related parties are considered intra-company transfers of funds, and, as such, do not qualify for a circumstances-of-sale adjustment. Antifriction Bearings (Other Than Tapered Roller Bearings) and Parts Thereof from France, 57 Fed. Reg. 28,360, 28,407 (June 24, 1992). As intra-company transfers of funds, these payments are considered to be part of the general operating expenses of the company, and not costs directly related to particular sales. Porcelain-on-Steel Cooking Ware from Mexico, 51 Fed. Reg. at 36,438.
Commerce applies a two-prong test in order to determine whether an adjustment for related-party commissions is appropriate. First, the commissions must be directly related to the specific sales. Second, the commission arrangement must be the result of arm’s length negotiations. LMI-La Metalli Industriale, S.p.A. v. United States, 8 Fed. Cir. (T) 157, 160-61, 912 F.2d 455, 458-59 (1990); Certain Welded Carbon Steel Standard Pipes and Tubes From India, 57 Fed. Reg. 54,360, 54,367 (Nov. 18, 1992).
In its Final Results, Commerce found that the commissions were directly related to the specific sales in question. Brass Sheet and Strip from Sweden, 57 Fed. Reg. at 2707. However, Commerce was not satisfied that the commission payments from Outokumpu AB to Outokumpu USA were at arm’s length. Id.
Commerce applies guidelines set forth in Coated Groundwood Paper From the United Kingdom, 56 Fed. Reg. 56,403, 56,406 (Nov. 4, 1991), to determine whether commissions paid to related parties in the United States are at arm’s length. Commerce first compares the commissions paid to related and unrelated sales agents in the same market. If there are no commissions paid to unrelated parties, Commerce then compares the commissions earned by the related selling agent on sales of merchandise produced by the respondent, to commissions earned on sales of merchandise produced by other unrelated sellers or manufacturers. In appropriate circumstances, Commerce will also “examine the nature of the agreements or contracts between the manufacturer(s) and selling agent(s) which establish the framework for payment of commissions and for services rendered in return for payment, in order to ensure that both related and unrelated agents perform approximately the same services for the commission.” Id. If, based on the above analysis, Commerce is satisfied that the commission payments are made at arm’s length, and *210are directly related to the underlying sales, Commerce will make an appropriate adjustment. Id.
In this case, Commerce was not satisfied that the related-party commission payments were at arm’s length. The court finds that substantial evidence in the administrative record supports Commerce’s determination. OAB and OCUSA are related parties because both are wholly-owned subsidiaries of Outokumpu Copper OY of Finland. Plaintiffs’ Response at 37. During the period of the reviews, OAB did not use unrelated commissionaires to sell the subject merchandise in either the U.S. or the home market.3 Additionally, OCUSA did not act as a commissionaire for unrelated producers of the subject merchandise. Therefore, Commerce was unable to establish a benchmark against which to compare the arm’s length nature of the “commission” payments from OAB to OCUSA.4
Defendant-intervenors assert that Commerce improperly failed to place the burden of proof on plaintiffs to demonstrate that the payments from OAB to OCUSA were not made at arm’s length. The court notes defendant-intervenors’ concerns that plaintiffs had an incentive to avoid showing that the commissions were at arm’s length, because plaintiffs could thereby avoid a detrimental adjustment to U.S. price. Defendant-intervenors’ Memorandum at 36-37. Nevertheless, although manipulation of commission payments to influence dumping margin calculations may be possible, speculation as to the existence of such manipulation must yield to the evidence presented. LMI-La Metalli Industriale, 8 Fed. Cir (T) at 162, 912 F.2d at 459-60. Here, no evidence in the administrative record supports defendant-intervenors’ assertion that the commission payments in question were made at arm’s length.
In applying the second prong of the test set forth in Coated Ground-wood Paper From the United Kingdom, which was developed pursuant to the LMI-La Metalli Industríale decision, Commerce seeks to determine whether the payments between related parties were made at arm’s length. Commerce does not seek to prove that such payments were not made at arm’s length. Commerce has chosen to operate under the assumption that commission payments in related party transactions are not at arm’s length. Defendant-intervenors essentially argue that Commerce should abandon this presumption when considering related party commissions in the U.S. market, and instead presume that the related party transactions were made at arm’s length. Commerce, how*211ever, has rejected the argument that related party commissions should be considered under different standards depending upon the market in which the commission is paid. See Nepheline Syenite from Canada, 57 Fed. Reg. 9237, 9241 (Mar. 17, 1992).
The court reiterates that Commerce is given considerable deference in its decision to grant a circumstances-of-sale adjustment. Smith-Corona Group, Consumer Products Div., SCM Corp. v. United States, 1 Fed. Cir (T) 130, 137, 713 F.2d 1568, 1575 (1983). As longas Commerce’s “decision is reasonable, then Commerce has acted within its authority even if another alternative is more reasonable.” Koyo Seiko Co. v. United States, 16 CIT 366, 796 F. Supp. 517, 523 (1992). Commerce considered all of the circumstances of the “commission” relationship between OAB and OCUSA. Upon considering all of the circumstances, Commerce concluded that it need not abandon the presumption that these payments were not made at arm’s length. The court finds that in this case Commerce reasonably exercised its discretion by relying on this presumption and by not requiring plaintiff to prove that the payments from OAB to OCUSA were not made at arm’s length.5
In sum, because the court finds that Commerce acted reasonably in determining that the related-party commission payments were not made at arm’s length, and, thus, were not bona fide selling expenses that merited a circumstances-of-sale adjustment to U.S. price, the court affirms this portion of Commerce’s Remand Results.
3. Eliminating the Deduction for Freight Charges From U.S. Price:
In the underlying administrative reviews, Commerce used BIA to make a deduction to U.S. price for U.S. inland freight charges. The court, however, found that Commerce’s use of BIA for the amount of freight charges was improper because the respondents had neither refused nor were unable to timely produce information that Commerce had allegedly requested. Outokumpu, 829 F. Supp. at 1387.
Commerce has complied with the court’s instructions to recalculate U.S. price eliminating the reduction for freight charges. Remand Results at 18-19. Commerce had used BIA to determine U.S. freight expenses only in purchase price sales during the 1988-89 and 1989-90 reviews. Commerce did not make a BIA freight-expense adjustment to ESP sales in the 1989-90 review. Thus, Commerce’s modification affects only the purchase price calculations for the 1988-89 and 1989-90 reviews, and not the ESP calculations for the 1989-90 review. Commerce’s recalculations are in accordance with the court’s instructions and are therefore affirmed.
*2124. Computer Programming Errors:
On remand, Commerce has corrected computer programming errors as instructed by this court. Outokumpu, 829 F. Supp. at 1384. First, Commerce has assigned metal values to the proper month and year. Second, Commerce has recalculated the difference between the values of the metal alloys being compared. Third, where applicable, Commerce has included differences in fabrication costs in its recalculation of the adjustment for differences-in-merchandise. Finally, Commerce has made an adjustment to FMV to account for differences in the physical characteristics of the merchandise. Remand Results at 19. Because Commerce has complied with the court’s instructions, this portion of the Remand Results is similarly affirmed.
Conclusion
The court holds that the Remand Results from Commerce’s second and third administrative reviews are supported by substantial evidence on the record and are otherwise in accordance with law. Commerce has revised its final results and determined that Outokumpu’s weighted-average margins are 5.41 percent for the 1988-89 review period, and 6.32 percent for the 1989-90 review period. These Remand Results, issued September 27,1993, are hereby affirmed, and this case is dismissed.
Defendant-intervenors in this action include: Hussey Copper, Ltd.; The Miller Company; Olin Corporation-Brass Group; Revere Copper Products, Inc.; International Association of Machinists and Aerospace Workers; International Union, Allied Industrial Workers of America (AFL-CIO); Mechanics Educational Society of America (Local 56); and United Steelworkers of America (AFL-CIO/CLC). Defendant-intervenors are either domestic manufacturers of brass sheet and strip, or are domestic unions representing workers who produce such merchandise in the United States.
In the final computer program for the 1988-89 review, in which all U.S. sales were purchase price sales, Commerce did not add the U.S. commissions to FMV nor did Commerce make a deduction of home market indirect selling expenses from FMV as an offset to U.S. commissions.
Similarly, in the final computer program for the 1989-90 review, Commerce did not add U.S. commissions to FMV in comparisons with U.S. purchase price sales, nor did they deduct such commissions from U.S. price in ESP comparisons.
Furthermore, because Commerce did not offset U.S. commissions in either purchase price or ESP comparisons in the preliminary results of the 1989-90 review (because respondents did not provide data on home market indirect selling expenses), Commerce did not need to malee any further revisions to its final FMV calculations; in short, there were no commission offsets to eliminate upon recalculation of FMV
Originally, Outokumpu reported that a [ ] commission was paid to an unrelated sales agent in the United States. Questionnaire Response of July 3,1989 at 22. Subsequently, however, Outokumpu stated that no sales were made during the period of investigation that were subject to commissions paid to unrelated parties. See Supplemental Questionnaire Response of September 15,1989 at 13.
The administrative record indicates that Outokumpu submitted data regarding the “commission” payments. The data shows that OCUSA received from OAB the following payments during the review periods:
1. For ail non-closed consignment sales, [ ] during 1988; [ ] during 1989; [ \ during 1990; and
2. For all closed consignment sales, approximately [ 1.
Supplemental Questionnaire Response of May 16,1991 at 3-5.
Furthermore, in a separate antidumping matter, Brass Sheet and Strip From the Netherlands, 57 Fed. Reg. 9534 (Mar. 19,1992), Commerce also confronted the issue of whether an adjustment should be made for commissions paid by OAB to OCUSA. The respondent in that case, Outokumpu Copper Products B.V (“OBV”), like OCUSA, was a subsidiary of OAB (formerly known as Metallverken AB). In the 1988-89 and 1989-90 reviewperiods, OBV used an unrelated commissionaire to sell the subject merchandise in the United States. OBV also reported the amounts of payments made by OAB to OCUSA in connection with U.S. sales during the same periods. Based on a comparison of the data provided on commissions from OBV to unrelated commissionaires in the U.S. and OAB’s payments to OCUSA, Commerce determined that the payments from OAB to OCUSA were not made at arm’s-length rates. Id.