CSX Transportation, Inc. v. Surface Transportation Board

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued April 14, 2014 Decided June 20, 2014 No. 13-1230 CSX TRANSPORTATION, INC. AND NORFOLK SOUTHERN RAILWAY COMPANY, PETITIONERS v. SURFACE TRANSPORTATION BOARD AND UNITED STATES OF AMERICA, RESPONDENTS On Petition for Review of an Order of the Surface Transportation Board G. Paul Moates argued the cause for petitioners. With him on the briefs were Paul A. Hemmersbaugh, Matthew J. Warren, Peter J. Shudtz, Paul R. Hitchcock, and John M. Scheib. Michael L. Rosenthal, Carolyn F. Corwin, and Louis P. Warchot were on the brief for amicus curiae Association of American Railroads in support of petitioners. 2 Erik G. Light, Attorney, Surface Transportation Board, argued the cause for respondents. With him on the brief were William J. Baer, Assistant Attorney General, U.S. Department of Justice, Kristen C. Limarzi, Chief, Appellate Section, Shana M. Wallace, Attorney, and Craig M. Keats, General Counsel, Surface Transportation Board. Robert B. Nicholson, Attorney, U.S. Department of Justice, entered an appearance. Before: TATEL, Circuit Judge, and SILBERMAN and SENTELLE, Senior Circuit Judges. Opinion for the Court filed by Senior Circuit Judge SENTELLE. SENTELLE, Senior Circuit Judge: In July 2013, the Surface Transportation Board issued a decision modifying its procedures for rate reasonableness cases. See Rate Regulation Reforms (“Decision”), STB Ex Parte No. 715 (served July 18, 2013). CSX challenges the decision on four grounds, three of which we reject. We reject CSX’s argument that the Board violated its statutory mandate when it made simplified procedures available for all cases. We also conclude that the Board adequately explained its adoption of a new revenue-allocation methodology as well as its rationale for adopting a new interest rate for reparations. As to the fourth challenge, we find merit in CSX’s argument that the Board acted arbitrarily and capriciously in raising the relief cap for its most simplified rate reasonableness procedure. Specifically, it appears that the Board double-counted costs in producing its estimate without explanation. Accordingly, we remand so that the Board can address this objection. 3 BACKGROUND The Surface Transportation Board regulates the rates of interstate railroads. See BNSF Ry. Co. v. Surface Transp. Bd., 526 F.3d 770, 773 (D.C. Cir. 2008). By statute, a party may bring a complaint before the Board challenging a railroad’s rate. See 49 U.S.C. § 10704(b). Upon receiving a complaint, the Board must determine whether the railroad in question possesses “market dominance.” See §§ 10701(d)(1), 10707(b)-(c). To have market dominance, a railroad’s revenue must meet or exceed 180 percent of its variable costs for the “transportation to which the rate applies.” See § 10707(d)(1)(A). “If the Board determines . . . that a rail carrier has market dominance over the transportation to which a particular rate applies, the rate established by such carrier for such transportation must be reasonable.” See § 10701(d)(1). In determining whether a rate is reasonable, the Board applies principles known as Constrained Market Pricing. Coal Rate Guidelines, Nationwide, 1 I.C.C.2d 520 (1985). Constrained Market Pricing provides three essential criteria to guide the Board’s analysis. First, a shipper “should not be required to pay more than is necessary for the rail carrier(s) involved to earn adequate revenues.” Id. at 520. Second, it should not “pay more than is necessary for efficient service.” Id. Third, it “should not bear the costs of any facilities or services from which it derives no benefit.” Id. Instead, “[r]esponsibility for payment for [shared] facilities or services . . . should be apportioned according to the demand elasticities of the various shippers.” Id. If the Board finds a railroad’s rate unreasonable, it may prescribe a maximum lawful rate and order the railroad to pay reparations. See 49 U.S.C. § 11704. 4 Though Constrained Market Pricing provides complainants with a number of possible approaches to challenge a rate, almost all rate cases have proceeded under the Stand-Alone Cost test, sometimes referred to as the “SAC test.” In a SAC test, complainants design a hypothetical stand-alone railroad, sometimes referred to as an “SARR,” which is “a fully efficient hypothetical competitor railroad that serves the complaining shipper and other traffic sharing common facilities.” CSX Transp., Inc. v. Surface Transp. Bd., 568 F.3d 236, 238 opinion vacated in part on reh’g, 584 F.3d 1076 (D.C. Cir. 2009). The Board will find a challenged rate unreasonable if the stand-alone railroad would generate revenues that “exceed[] the costs (including a reasonable profit) of running the stand-alone railroad.” Id. at 238–39. In effect, SAC tests restrain railroads from exploiting market power, and prevent railroads from forcing captive shippers to pay for inefficiencies in the railroads’ investment operations. See Simplified Standards for Rail Rate Cases, EP 646 (Sub- No. 1), slip op. at 13 (STB served Sept. 5, 2007) (2007 Simplified Standards), aff’d CSX Transp., 568 F.3d at 236. A. SIMPLIFIED PROCEDURES Because SAC tests are complicated and costly, Congress directed the Board to “establish a simplified and expedited method for determining the reasonableness of challenged rail rates in those cases in which a full stand-alone cost presentation is too costly, given the value of the case.” 49 U.S.C. § 10701(d)(3). Over the years, the Board discharged this duty by developing a two-tiered system. For cases worth less than $5 million—the Board’s estimate for presenting a SAC case—the Board created a procedure it called the Simplified-SAC test. 2007 Simplified Standards, at 13–16, 30–31. Unlike Full-SAC, Simplified-SAC does not require the complainant to create a stand-alone railroad and does not 5 concern itself with uncovering inefficiencies in defendants’ railroads. See CSX Transp., 568 F.3d at 245. Instead, it focuses solely on whether the defendant railroad is abusing its monopoly power. Id. Simplified-SAC is limited to the predominant route of the issue traffic, it assumes that all infrastructure along the route is needed to serve the traffic, and it includes all traffic that traversed the route in the prior twelve-month period. 2007 Simplified Standards at 15–16. Simplified-SAC also uses the Board’s Uniform Rail Costing System to estimate total operating and equipment expenses. Id. at 16. Finally, under Simplified-SAC—as originally formulated—the Board used its prior Full-SAC cases to simplify the calculation of Road Property Investment costs— unlike in Full-SAC which includes a complete analysis of these costs. Id. at 15. For cases worth less than $1 million—the Board’s cost estimate for Simplified-SAC cases—the Board created the “Three Benchmark” approach. This method is simpler still; it assesses rates by comparing “the challenged rate to three benchmark figures, each expressed as a relationship between revenues and variable costs, i.e., those costs that increase as traffic over the railroad increases . . . .” See CSX Transp., 568 F.3d at 240 (internal quotations and citations omitted). B. THE RULEMAKING UNDER REVIEW In 2012, the Board instituted a rulemaking to revise its procedures for rate reasonableness cases. Four of the Board’s proposals are pertinent to this challenge. First, in the notice of proposed rulemaking (“NPRM”), the Board proposed to remove the relief cap on Simplified-SAC cases. See Rate Regulation Reforms, Ex Parte No. 715 (July 25, 2012) at 13. The Board observed that the “Full-SAC and Simplified-SAC approaches both appear to be . . . appropriate method[s] to 6 judge the reasonableness of the challenged rates, and there is no apparent reason to force the shipper to use the more expensive Full-SAC . . . in cases where the shipper seeks more than $5 million in relief.” Id. at 14. The Board also noted, on the other hand, that if a “complainant believes that there are enough inefficiencies . . . to justify the added expense and complexity of a Full-SAC presentation, it may pursue relief using [the] hypothetical [stand-alone railroad] analysis.” Id. The Board further proposed to adjust the procedures for Simplified-SAC cases by requiring complainants to develop full Road Property Investment costs, just as they do in Full-SAC cases. See id. at 14. This change would render the methodology both more expensive and more robust. See id. Commenters argued that Simplified-SAC was too imprecise for high-value cases, and that statutorily the Board could apply Simplified-SAC only to cases for “which a full stand-alone cost presentation is too costly, given the value of the case.” 49 U.S.C. § 10701(d)(3). The Board rejected both arguments. It found nothing in the governing statute that precluded its proposal. Decision, at 16. And noting that “all regulatory procedures to regulate rates necessarily entail some degree of imprecision,” the Board found nothing in commenters’ arguments that showed that the imprecisions in Simplified-SAC rendered it inappropriate for all cases. Decision, at 17–18. Accordingly, the Board adopted its proposal in full. Second, the Board proposed to raise the relief cap on its most simplified reasonableness procedure, the Three Benchmark approach, from $1 million to $2 million. As noted earlier, this procedure assesses reasonableness through a simple comparison between the challenged rate and “three benchmark figures, each expressed as a relationship between 7 revenues and variable costs, i.e., those costs that increase as traffic over the railroad increases . . . .” See CSX Transp., 568 F.3d at 240 (internal quotations and citations omitted). The Board proposed this increase because developing Road Property Investment costs under Simplified-SAC increased the litigation cost of that type of complaint. See NPRM, at 15. Commenters argued that the Three Benchmark method was too imprecise to be broadened to so much traffic, and they warned that expanding its applicability threatened to artificially “ratchet” down rates. Shippers, on the other hand, presented evidence that the estimate was still too low. The Board agreed with the shippers. It decided to raise the relief cap to $4 million dollars. It arrived at this figure by estimating the cost of a Simplified-SAC case under the old procedures at $2 million, and adding to that a $2 million estimate of the new costs of producing Road Property Investment evidence. See Decision, at 22–25. In its third proposal, the Board addressed the problem of allocating revenue in Full-SAC cases to so-called “cross- over” traffic. As we recently explained: The [stand-alone railroad’s] projected revenues are determined based on the real-world rates charged by the railroad servicing the traffic group included in the SAC presentation. This calculation is straightforward when complainants model the entire traffic group, but becomes more complex when SAC presentations include movements that travel a portion of their journey on the hypothetical SARR and a portion on actual railroads. Such “cross-over” traffic requires the Board to allocate revenue between the SARR and the real-world railroad. 8 BNSF Ry. Co. v. Surface Transp. Bd., 741 F.3d 163, 164 (D.C. Cir. 2014) (WFA II) (citations omitted). The Board has struggled for some time to produce a satisfactory solution to this problem. In 2006, the Board adopted the revenue allocation method known as Average Total Cost, or “ATC.” Id. Average Total Cost allocated revenues to the stand-alone railroad based on the average total cost of a traffic patterns’ movement on the stand-alone railroad. Id. In September 2007—in the midst of a dispute between BNSF Railway Company and Western Fuels Association, Inc.—the Board discarded ATC and applied a new methodology: Modified Average Total Cost, or “Modified ATC.” Id. at 165. The Board did so to address an “illogical and unintended result” of ATC. Id. Under ATC, the shipper’s “traffic patterns had produced scenarios in which revenue generated by some movements would not cover the variable costs of those movements on-SARR . . . .” Id. Under Modified Average Total Cost, revenue is first allocated to the portions of a cross- over movement to cover its respective variable costs. Id. Then, any remaining revenue is allocated in proportion to the relative average total costs of serving the on- and off-SARR segments. Id. The Board applied Modified Average Total Cost to the case and granted the shipper relief in 2009. BNSF petitioned this Court for review, arguing that the Board acted arbitrarily and capriciously by departing from ATC. Specifically, BNSF claimed that Modified Average Total Cost improperly double counted variable costs. See BNSF Ry. Co. v. STB, 604 F.3d 602, 604 (D.C. Cir. 2010) (WFA I). We granted BNSF’s petitions in part to allow the Board to address this objection on remand. 9 On remand, BNSF still advocated for reversion to Average Total Cost. It also argued, however, that “even if the below-cost allocations under ATC were problematic, Modified ATC represented a disproportionate response to this problem.” WFA II, at 165. BNSF “suggested a different approach that would proportionately adjust ATC to address the problem it created.” Id. “Under BNSF’s suggestion . . . , the Board would first apply ATC to all movements with revenues exceeding variable costs. Then, for below-cost traffic, the Board would allocate additional revenues to eliminate the shortfall.” Id. The Board upheld its use of Modified Average Total Cost and refused to apply BNSF’s suggestion to the case before it, concluding that BNSF’s proportionality critique fell outside the scope of our remand. Id. The Board did, however, recognize the merits of BNSF’s suggestion, and initiated this rulemaking to consider whether such a method might be a better allocation method than Modified ATC. Id. In the notice of proposed rulemaking, the Board explained that its proposed methodology, which it called Alternative ATC, had been brought to its attention in the Western Fuels Association remand. NPRM, at 18. It described the proposed methodology as having two steps. The first step would “follow . . . original ATC . . . .” Id. at 17. Then, “[a] second step would . . . be performed to ensure that the revenue allocated to both the facilities replicated by the SARR and those of the residual defendant carriers would not be driven below the defendant’s [Uniform Rail Costing System] variable costs for the movement over those segments.” Id. at 17–18. The method, the Board suggested, “might better address two competing principles in the selection of a cross-over traffic methodology”; namely, the need to reflect “economies of density” and the need to avoid 10 “the implausible result of driving the revenue allocation on any segment below variable costs.” Id. at 18. In their comments below, Petitioners argued, among other things, that “an on-SARR revenue allocation that generates a[] [revenue-variable-cost ratio] of less than 100 percent is [neither] implausible [n]or irrational.” Opening Comments of CSXT & NS, at 17 (filed Oct. 23, 2012). The Board was unpersuaded, and adopted Alternative ATC as proposed. Decision, at 30. Finally, the Board changed the interest rate it applies to reparations from the 90-day U.S. Treasury Bills (“T-Bill”) rate to the U.S. Prime Rate. See Decision, at 35–36. It concluded that the Prime Rate better reflected the opportunity costs a shipper loses through unreasonable rates. Id. ANALYSIS We review Board decisions under the Administrative Procedure Act, and will set aside a Board decision if it is “arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.” 5 U.S.C. § 706(2)(A). We evaluate the Board’s statutory interpretation under the framework of Chevron U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984). “If the intent of Congress is clear, that is the end of the matter; for the court, as well as the agency, must give effect to the unambiguously expressed intent of Congress.” Id. at 842–43. If, on the other hand, the statute is ambiguous, “the question for the court is whether the agency’s answer is based on a permissible construction of the statute.” Id. at 843. CSX challenges the Decision on four grounds. It argues (A) that the Board unreasonably construed § 10701(d)(3) 11 when it made the Simplified-Stand-Alone Cost test available for all cases; (B) that the Board acted arbitrarily and capriciously in raising the relief cap for Three Benchmark cases to $4 million; (C) that the Board failed to adequately explain its departure from ATC; and (D) that the Board acted unreasonably in replacing the T-Bill rate with the U.S. Prime Rate. We treat each in turn. A. REMOVAL OF THE RELIEF CAP FOR SIMPLIFIED-SAC CASES CSX argues that the Board’s decision to remove the relief cap on Simplified-SAC cases violates Congress’s clear intent to the contrary. Congress directed the Board to “establish a simplified and expedited method for determining the reasonableness of challenged rail rates in those cases in which a full stand-alone cost presentation is too costly, given the value of the case.” 49 U.S.C. § 10701(d)(3). CSX reads this language as a clear directive from Congress: the Board is to use simplified methods only when a full SAC test is too costly for the case at hand. Otherwise, CSX argues, Congress’s qualifying language—“given the value of the case”—would be mere surplusage. Why would Congress circumscribe the Board’s task if it did not intend for that circumscription to actually limit the Board’s discretion? CSX bolsters its argument by reference to the interpretive canon of expressio unius. According to CSX, the fact that Congress made clear the Board was to use a simplified approach for low-relief cases implies that it is not to use such an approach for high- relief cases. The Board violated this directive, CSX argues, when it allowed Simplified-SAC in any case. We disagree. The Board’s interpretation is perfectly consistent with Congress’s direction in the statute. “[T]he Board is the expert body Congress has designated to weigh 12 the many factors at issue when assessing whether a rate is just and reasonable.” CSX Transp., 568 F.3d at 240. And as a general matter, it enjoys broad discretion to design rate reasonableness tests. See 49 U.S.C. § 10701(d). It is true, as CSX points out, that Congress removed the Board’s discretion with respect to low-relief cases. But Congress made no direction whatsoever for other cases, and no direction can be implied from Congress’s silence. Ultimately, CSX’s position rests on a logical fallacy. “If P then Q” does not imply “If not P then not Q”—yet this is CSX’s entire argument. It infers from the fact that the Board must provide a simplified approach for low-relief cases, that it must not do so for any other case. On the contrary, the statute represents a floor, not a ceiling for the Board’s discretion. Congress required the Board, at a minimum, to develop a simplified approach for low-recovery cases; nothing in that requirement circumscribed the Board’s discretion concerning high-recovery cases. And reading the statute thus—that is according to its own terms—renders no part of it surplusage. We also find the Board’s interpretation reasonable. Since the Board retained discretion to adjudicate high relief cases as it saw fit, we defer to the agency so long as it provided a “reasoned explanation for why it chose” to make those cases eligible for Simplified-SAC. See Vill. of Barrington v. STB, 636 F.3d 650, 660 (D.C. Cir. 2011). The Board satisfied this burden. As it explained in its decision, the Board saw “no reason that Congress would order the agency to prevent captive shippers from using [an] alternative approach” that is “simplified” and “expedited,” so long as it is also “a robust method for determining the reasonableness of challenged rail rates.” Decision, at 17. This is not an unreasoned explanation, and we therefore defer to it. 13 In cursory fashion, CSX also argues that the Simplified- Stand-Alone Cost methodology introduces imprecisions that render its application to high-relief cases arbitrary and capricious. In a Full-SAC case, the parties develop operating costs for the traffic group specific to the stand-alone railroad. Simplified-SAC, on the other hand, uses Uniform-Rail- Costing-System numbers which do not account for higher- cost movements—for instance, those transporting highly toxic materials—and could result in significant distortions in high- value cases. We are not persuaded. The Board responded to CSX’s concerns in its decision. It noted that the Uniform Rail Costing System is its “general purpose costing model” and that “using [Uniform Rail Costing System] system-average costs should provide a reasonable approximation of the total operating expenses of the traffic group.” Id. at 17 (quotations omitted). The Board also concluded that commenters “never explain[ed] what feature of [Uniform Rail Costing System] introduce[d] so much imprecision in Simplified-SAC—as compared to Full- SAC . . . —to warrant a limitation on relief.” Id. at 17. “Indeed,” the Board noted, “although a Full-SAC presentation is more ‘precise’ than a Simplified-SAC presentation, it is so only in the sense that, through a highly complex and detailed presentation involving a hypothetical railroad, it ferrets out operating inefficiencies.” Id. at 17. Accordingly, the Board concluded, “[t]here is no basis to permit the railroads to earn excessive profits simply because, unlike the Full-SAC method, the Simplified-SAC method does not detect the inefficiencies in rail operations that may further raise rates.” Id. at 16. Given its thorough treatment of these comments, we find the Board’s decision neither arbitrary nor capricious. Finally, we note that amicus for Petitioner, Association of American Railroads, adds several non-statutory arguments of 14 its own, but because “we ordinarily do not entertain arguments not raised by parties” we decline to address them here. Narragansett Indian Tribe v. Nat’l Indian Gaming Comm’n, 158 F.3d 1335, 1338 (D.C. Cir. 1998). B. INCREASE OF THE RELIEF CAP FOR THREE BENCHMARK CASES CSX’s second challenge goes to the Board’s decision to raise the relief cap on its most simplified reasonableness procedure, the Three Benchmark approach, from $1 million to $4 million. CSX challenges the Board’s decision on two fronts. First, it argues that the Board’s rationale for the increase was predicated on an incomplete record and mathematical errors. Second, it argues that the Board’s decision—which dramatically broadened the availability of the Three Benchmark approach—threatens to artificially depress rates. The Board decided to raise the relief cap in light of its revised estimate that a Simplified-SAC case would cost $4 million to present. It reached this number in two steps. First it estimated the cost of presenting a case under the old Simplified-SAC procedures, relying principally on the testimony of U.S. Magnesium, LLC, the only party to have brought a case under the old procedures. See Decision, at 22– 23. The Board then added to this the estimated costs of the new procedures to reach its result. CSX challenges both steps. CSX claims that U.S. Magnesium’s estimate was inherently implausible, and that the Board thus erred in accepting it. U.S. Magnesium testified that its litigation costs could have reached $2 million in a Simplified-SAC case under the old methodology. See id. But U.S. Magnesium 15 settled its case, and in light of its actual expenses, its math does not add up, according to CSX. U.S. Magnesium actually spent only $750,000 in preparing its opening evidence. CSX argues that there is simply no way it would have had to spend $1.25 million more after it had constructed its entire case. We need not linger on the details of CSX’s claims here, however, because its argument misses the larger picture. U.S. Magnesium was the only party ever to have pursued relief under Simplified-SAC. Thus its estimate represents the only actual data the Board had to work with in making its estimate of possible costs under the old procedures. And as the Board explained, U.S. Magnesium had brought a “relatively straightforward Simplified-[Stand-Alone Cost] case of single commodity from a single origin to 12 destinations.” Id. at 23. Moreover, U.S. Magnesium had incurred “no expense in establishing market dominance because the defendant had conceded that issue.” Id. The Board did not act arbitrarily or capriciously in estimating the cost of Simplified-SAC cases at $2 million, given the limited data at its disposal and the simplified nature of U.S. Magnesium’s case. CSX further argues that the Board also erred in the second step of its analysis. Once it had estimated the cost of a case under the old procedures, the Board added the new costs of producing a full Road Property Investment presentation. Id. Accepting expert testimony that developing Road Property Investment costs usually accounted for about a third of the total costs in presenting Full-SAC cases—or $1.9 million—the Board added that to its baseline estimate to reach its final relief cap of $4 million. Id. at 23–25. CSX argues that the Board erred by adding this estimate in its entirety without subtracting the cost of developing Road Property Investment evidence under the prior regime. Though Road 16 Property Investment calculations were streamlined under the former procedures, Simplified-SAC complainants were still required to prepare some of the Road Property Investment analysis required in Full Stand-Alone Cost cases. See Simplified Standards, at 38–48. The Board does not offer much in response to this objection. It merely claims that, even if it did double count these costs, CSX has not shown that “the Board would have had to choose a lower limit.” Resp. Brief, at 52. But this argument answers the wrong question. The APA places the burden on the Board to render a decision that “examine[s] the relevant data and articulate[s] a satisfactory explanation for its action including a ‘rational connection between the facts found and the choice made.’” Motor Vehicle Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto Ins. Co., 463 U.S. 29, 43 (1983) (quoting Burlington Truck Lines v. United States, 371 U.S. 156, 168 (1962)). Because the Board did not explain the apparent double counting of Road Property Investment costs—first in the baseline and then in the new cost addition—it did not rationally connect its choice of action to the facts. See id. Accordingly, we will remand for the Board to address CSX’s double-counting objection. We will not, however, vacate the Board’s decision. This is an instance in which the Board “may be able readily to cure a defect in its explanation of [its] decision” and the “disruptive effect of vacatur” would be high. Heartland Regional Medical Center v. Sebelius, 566 F.3d 193, 198 (D.C. Cir. 2009). CSX also challenges the broadening of the Three Benchmark approach as arbitrary on a separate rationale. It argues that the Board irrationally expanded the applicability of the Three Benchmark approach to over two-thirds of all regulated traffic without sufficient explanation. When the Board lowers a rate by using the averages of other rates, it 17 thereby lowers the average for future cases—which threatens to ratchet down rates artificially. This Court has rejected rate- comparison formulas due to their ratcheting potential in the past. See, e.g., Burlington N. R.R. Co. v. I.C.C., 985 F.2d 589, 597 (D.C. Cir. 1993). Previously the Board defended against this threat by claiming that few cases were eligible for this approach. It was irrational, CSX claims, for the Board to ignore this problem now that so many cases are eligible for the Three Benchmark approach. We find that the Board adequately answered this challenge in its decision. There it explained that (1) relief continued to be limited, (2) ratcheting would require an avalanche of successful cases, (3) the Board could reassess its approach in such an implausible scenario, and (4) the chosen limit represented a reasonable balance of “concerns about possible ratcheting with Congress’s clear intent that shippers with smaller disputes have a means of challenging their rates.” Decision, at 24. As we noted earlier, almost no parties had proceeded under a Simplified-SAC approach before this rulemaking. It was reasonable for the Board to conclude from this evidence that Simplified-SAC is too costly where the value of the case is less than the cost of producing such a presentation, and that, for such cases, the Three Benchmark approach would be the only viable option. See id. C. ADOPTION OF ALTERNATIVE ATC CSX’s third challenge goes to the Board’s modification of its cross-over-revenue allocation method. CSX argues that the Board failed to respond to important comments regarding the Board’s reasoning in adopting the Alternative Average Total Cost formula for revenue allocation. The Board justified its proposal, in part, on the rationale that Average Total Cost had produced the “illogical” and “implausible” 18 result of allocating revenues insufficient to cover segments’ variable costs on the stand-alone railroad. But according to CSX, commenters demonstrated that below-variable-cost allocations were consistent with the logic and application of the Alternative Total Cost formula. The Board erred in ignoring these comments in this rulemaking. To see why CSX’s argument fails, we must tour the convoluted procedural history from which this rulemaking sprang. As we noted earlier, the Board originally discarded Average Total Cost in the midst of the Western Fuels Association proceeding, and adopted in its place Modified Average Total Cost. WFA II, 741 F.3d at 164. It was in this original proceeding that the Board explained that ATC had the “illogical and unintended result” of producing “scenarios in which revenue generated by some movements would not cover the variable costs of those movements” on the stand- alone railroad. Id. at 165. When we remanded to the Board to address BNSF’s double-counting objection, BNSF maintained that Modified ATC was an irrational response to the problem created by ATC, but it also suggested a different approach: Alternative ATC. The Board, recognizing the merits of BNSF’s suggestion, initiated a rulemaking to consider whether Alternative ATC might in fact be a better allocation method than Modified ATC. It is this—the Board’s subsequent decision to adopt Alternative ATC—that CSX now challenges. But the character of this decision is not whether the Board should discard ATC or, relatedly, whether there was a problem with it. The Board based its proposal in the NPRM on its consideration of cross-over revenue allocation in the Western Fuels Association proceedings. See NPRM, at 8. There the Board had already rejected ATC, and by framing its proposal as an offshoot of these proceedings, the Board put 19 commenters on notice that it was not considering whether to revert to Average Total Cost, or whether below-variable-cost allocations were problematic. On the contrary, the rulemaking we now review assumed that below-variable-cost allocations were illogical, and operated on that assumption. More to the point, the Board addressed the same arguments in the Western Fuels Association proceedings that CSX accuses it of ignoring now:  First, CSX claims the Board ignored comments to the effect that below-variable-cost allocations under ATC are not illogical. But as the Board noted in the Western Fuels Association proceedings, allowing such allocations “creates the illusion that . . . more revenue is available to help pay for [fixed] costs . . . than is available in reality.” Western Fuels Association Remand, at 7.  Second, CSX highlights comments arguing that it is meaningless to compare stand-alone railroad revenues to actual variable costs, because “the Stand-Alone Railroad is optimally efficient and would have different variable costs on the on-Stand-Alone- Railroad segment tha[n] [the] defendant carrier would have in the real world.” Pets. Br., at 52–53. The Board addressed this concern in the Western Fuels Association proceedings. There it considered the stand-alone railroad’s efficiency irrelevant “because the fairness of a revenue-allocation procedure should not depend on . . . the complainant having to design a [[s]tand-[a]lone [r]ailroad] that is more efficient than the incumbent railroad.” Western Fuels Association Remand, at 4. 20  The Board also rejected a third argument CSX now highlights. Commenters argued that the Board’s use of Uniform Rail Costing System costs is “inappropriate both because that segment may have significantly different costs than the carrier’s overall average costs, and because some Uniform Rail Costing System costs are unattributable [fixed] costs . . . , not variable costs.” Pets. Br., at 53. As the Board noted in the Western Fuels Association proceedings, use of Uniform Rail Costing System in a SAC analysis is appropriate because the Uniform Rail Costing System is “a measure of intermediate-variable costs,” which includes costs that are “fixed in the short term . . . but variable over the longer term.” Western Fuels Association Remand, at 8.  Fourth, in the Western Fuels Association proceedings, the Board rejected the contention that the burden should be on the complainant to avoid below-variable- cost allocations, an argument revived in this proceeding. See Western Fuels Association Remand, at 8.  Finally, commenters argued that the Board need not be concerned with below-variable-cost allocations for artificially segmented portions of a rail that would in the real world be priced in its entirety. In the Western Fuels Association proceedings, however, the Board noted that carriers will “in general, estimate revenues attributable to the segment in an amount at least equal to the long-run variable costs of providing service over that segment.” Western Fuels Association Remand, at 7. 21 In short, the Board addressed the arguments CSX now accuses it of ignoring in the original proceeding in which it discarded the Average Total Cost formula. The Board had no reason to repeat its responses in this proceeding, which addressed only whether to replace Modified ATC with Alternative ATC. D. MODIFICATION OF THE INTEREST RATE CSX’s final challenge goes to the Board’s adoption of a new interest rate for reparations. Prior to this rulemaking, the Board used the T-Bill rate for reparations. In the NPRM, the Board noted its concern that the T-Bill rate was insufficiently compensatory (0.1% at the time) and proposed to replace it with the U.S. Prime Rate (then 3.25%). NPRM, at 18. According to the Board, the interest rate should “correlate[] to market interest rates over a comparable time frame,” and the Board asserted that the U.S. Prime Rate satisfied this test because it was “the interest rate that the banks charge to their most creditworthy customers.” Id. CSX argues that it presented evidence that the Board’s stated premises for changing the interest rate were incorrect, yet the Board failed to address this evidence in its decision. CSX challenged the Board’s statement that the Prime Rate measures actual market interest rates; as CSX pointed out, the Prime Rate is merely a base rate or pricing index. Second, CSX pointed out that the Prime Rate is based on the Federal index, and is not a rate actually given to creditworthy customers. Yet, according to CSX, the Board ignored these comments and simply repeated its belief that the Prime Rate “correlates to market interest rates” and that it is the “rate that the banks charge to their most creditworthy customers.” Decision, at 35–36. 22 We find CSX’s arguments unpersuasive. CSX does not dispute that shippers’ opportunity cost is the appropriate measure for interest on reparations, that the T-Bill rate does not accurately reflect that cost, or that the U.S. Prime Rate represents a rate more attuned to that cost. In short, CSX does not dispute the essential reasoning on which the Board rested its decision to replace the T-Bill rate with the Prime Rate. Accordingly, we find that the Board adequately explained its decision to adopt a new interest rate. CONCLUSION We grant the petition in part, so that the Board on remand can address CSX’s claim that the Board double-counted costs in producing its estimate for the Three Benchmark relief cap, and we otherwise deny the petition.