BNSF Railway Co. v. Surface Transportation Board

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT Argued October 11, 2013 Decided January 31, 2014 No. 12-1327 BNSF RAILWAY COMPANY, PETITIONER v. SURFACE TRANSPORTATION BOARD AND UNITED STATES OF AMERICA, RESPONDENTS BASIN ELECTRIC POWER COOPERATIVE, INC. AND WESTERN FUELS ASSOCIATION, INC., INTERVENORS On Petition for Review of an Order of the Surface Transportation Board Richard P. Bress argued the cause for petitioner. With him on the briefs were Michael J. Gergen, Lori Alvino McGill, Paul T. Crane, Richard E. Weicher, Samuel M. Sipe, Jr., and Anthony J. LaRocca. Erik G. Light, Attorney, Surface Transportation Board, argued the cause for respondents. With him on the brief were 2 William J. Baer, Assistant Attorney General, U.S. Department of Justice, Robert B. Nicholson and Nickolai G. Levin, Attorneys, Raymond A. Atkins, General Counsel, Surface Transportation Board at the time the brief was filed, and Craig M. Keats, Deputy General Counsel. John H. LeSeur argued the cause for intervenors. With him on the brief were Christopher A. Mills and Peter A. Pfohl. Before: KAVANAUGH, Circuit Judge, and SENTELLE and RANDOLPH, Senior Circuit Judges. Opinion for the Court filed by Senior Circuit Judge SENTELLE. Dissenting opinion filed by Senior Circuit Judge RANDOLPH. SENTELLE, Senior Circuit Judge: BNSF Railway Company (“BNSF”) petitions for review of the decision of the Surface Transportation Board (“Board”) to adhere to a revenue-allocation methodology known as Modified ATC in determining that the rates BNSF charged Western Fuels Association, Inc. and Basin Electric Power Cooperative, Inc. (collectively “WFA”) were unreasonably high. BNSF first challenged this Modified ATC methodology in this Court in 2009. In 2010 we remanded the case to the Board so that it could address one of BNSF’s objections to Modified ATC in the first instance. On remand, the Board concluded that portions of BNSF’s arbitrary and capricious challenge fell outside the scope of the case given the specificity of our 2010 remand. This conclusion was in error. Because we never actually resolved BNSF’s arbitrary and capricious challenge to Modified ATC, we grant the petition, vacate the Board’s decision, and again remand the case to the Board. 3 I. BACKGROUND Until 2004, BNSF transported coal for WFA under a long-term contract. When the parties could not successfully negotiate a replacement contract, BNSF established a common carrier rate for WFA. Unsatisfied with this rate, WFA complained to the Board, alleging that the new rate was unreasonable. The Board employs a “Stand-Alone-Cost” (“SAC”) test to determine whether a railroad’s rates are unreasonable. BNSF Ry. Co. v. STB, 526 F.3d 770, 776–77 (D.C. Cir. 2008). Under the SAC test, complainants design a hypothetical optimally efficient stand-alone railroad (“SARR”) that serves a subset of movement in the railroad’s network, including the traffic to which the challenged rate applies. Id. at 777. The SAC test then calculates what a railroad would charge if operating the SARR. Id. The SARR’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. Id. at 782. Such “cross-over” traffic requires the Board to allocate revenue between the SARR and the real-world railroad. Id. When WFA first complained, the Board apportioned cross-over traffic revenues based on the percentage of miles a shipper used the SARR, a method known as Modified Straight-Mileage Prorate (“MSP”). Though simple, MSP “did not take into account ‘economies of density’—the principle that the more traffic on a given stretch of rail, the lower the average cost (and hence the lower the cross-over-traffic 4 revenue that should be attributed to it).” Id. In February of 2006—after WFA had submitted its SAC presentation—the Board held this matter in abeyance while it considered and ultimately adopted a new revenue allocation method called Average Total Cost (“ATC”). Under ATC, revenues are allocated to the hypothetical railroad based on the average total cost of a traffic pattern’s on-SARR movement. The Board explained that ATC successfully takes account of economies of density because it is centered on average total costs rather than average variable costs. In September 2007, the Board concluded that WFA had failed to make its case. In reaching that conclusion, the Board discarded ATC—in its first chance to apply it—and applied a new methodology: Modified ATC. The Board adopted Modified ATC to address an “illogical and unintended result” of ATC. Under ATC, WFA’s traffic patterns had produced scenarios in which revenue generated by some movements would not cover the variable costs of those movements on- SARR (“below-cost traffic”). To address this problem, Modified ATC proceeds in two steps. First, revenue is allocated to the on-SARR and off-SARR portions of a crossover movement to cover its respective variable costs. Second, remaining revenue is allocated between the SARR and defendant railroad in proportion to the relative average total costs of serving the on- and off-SARR segments. The Board allowed WFA to redesign its presentation in light of the changed rule. To best take advantage of Modified ATC, WFA reengineered its SARR, and all but eliminated below-cost traffic patterns. After reviewing WFA’s revised presentation, the Board concluded that BNSF’s rates were unreasonably high and ordered $345 million in relief. 5 In 2009, BNSF petitioned this Court for review, challenging the Board’s decision on several grounds. As to Modified ATC, BNSF argued that the Board acted arbitrarily and capriciously by departing from ATC. BNSF argued that Modified ATC double counts variable costs—first to cover variable costs, and then again as a component of total costs— and thus fails to account for economies of density. BNSF Ry. Co. v. STB, 604 F.3d 602, 604 (D.C. Cir. 2010) (WFA I). Because the Board had not “specifically mention[ed] double- counting” in earlier proceedings, we granted in part BNSF’s petitions in order to allow the Board on remand to address this objection, but otherwise denied the petitions. Id. at 613. On remand, BNSF maintained that Modified ATC was an irrational response to the problem created by ATC. It advocated reversion to ATC, but also argued that even if the below-cost allocations under ATC were problematic, Modified ATC represented a disproportionate response to this problem. Thus BNSF suggested a different approach that would proportionately adjust ATC to address the problem it created. Under BNSF’s suggestion (“Alternative ATC”), 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. A divided Board upheld its use of Modified ATC and refused to consider BNSF’s proportionality critique, concluding that it fell outside the scope of our remand. It observed that we did not specifically direct the Board to address any proportionality problem with Modified ATC, or to consider Alternative ATC as a solution to this problem. At the same time, the Board recognized the merits of Alternative ATC, and noted that it would initiate a rulemaking to consider 6 whether Alternative ATC might in fact be a better allocation method than Modified ATC. The Board also considered placing this matter in abeyance again pending the rulemaking, but decided against it for three reasons. First, the Board feared that so doing would incentivize litigants to propose theories late in litigation. Second, it found that the interests of administrative finality weighed in favor of ending the matter. Finally, the Board noted that applying yet another method to this case would prolong it even further since WFA would then be entitled to revise its SARR again. After ruling on this case, the Board initiated rulemaking and ultimately adopted a version of Alternative ATC for future cases. See Rate Regulation Reforms, STB Ex Parte No. 715 (July 18, 2013), at 30. BNSF petitioned this Court for review. II. 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). BNSF brings two principal challenges on appeal. First, it contends that the Board erred in refusing to address its proportionality argument below. Second, it contends that the Board should have held this matter in abeyance while it evaluated (and ultimately adopted) a form of Alternative ATC. We agree with BNSF’s first contention. BNSF argues that the Board misconstrued our holding in WFA I. There, we remanded without addressing BNSF’s substantive challenges to Modified ATC. We summarized 7 BNSF’s argument that Modified ATC counts “variable costs . . . twice,” and therefore “fails appropriately to consider economies of density.” WFA I, 604 F.3d at 612. But because the Board had not addressed BNSF’s double-counting objection, we granted in part BNSF’s petitions so that the Board on remand could address this objection. Id. at 613. According to BNSF, because we did not address the arbitrary and capricious challenge to Modified ATC in WFA I, this challenge remained for the Board to address on remand. The Board’s reasoning to the contrary is terse: “In remanding the case,” the Board noted, we “did not direct the Board to address either . . . the disproportionate-remedy argument [or] the proposed alternative ATC method.” And because we had not specifically directed the Board to address proportionality, the Board concluded that it could only address the issue by expanding the remand on its own initiative. It appears the Board understood WFA I as essentially approving its use of Modified ATC so long as the Board could furnish a satisfactory response to BNSF’s double-counting objection. This conclusion was in error. Our decision in WFA I neither explicitly nor implicitly ruled on BNSF’s substantive challenge to Modified ATC. We simply never reached the merits of BNSF’s arguments. Instead we returned the case to the Board so that it could address the double-counting objection. This allowed the Board to justify Modified ATC as against BNSF’s objections in the first instance so that we could, if need be, later evaluate that justification. This instruction did nothing to insulate Modified ATC from any of the substantive charges BNSF brought against it in WFA I. Thus the only question is whether BNSF’s proportionality argument was in fact preserved and presented in WFA I. See, e.g., W. Va. v. EPA, 362 F.3d 861, 871–72 (D.C. Cir. 2004) (rejecting as forfeited 8 claims that “petitioners never raised . . . to the agency in the administrative proceedings . . . or in the earlier challenges in this Court”). The Board argues that BNSF forfeited its proportionality argument. First, the Board characterizes BNSF’s proportionality argument as fatally at odds with the double- counting objections it presented in WFA I. Where BNSF’s double-counting objection embraces below-variable-cost allocations under ATC, the disproportionate argument starts from the opposite premise: such allocations are problematic. Thus, the Board concludes, the double-counting argument could not logically have encompassed the proportionality argument in WFA I, because the two flow from contradictory suppositions. We disagree. There is no incompatibility between BNSF’s double- counting and proportionality arguments. In fact, one flows logically from the other. Modified ATC only double counts variable costs with respect to cross-over traffic for which revenues exceed variable costs. For such movements, Modified ATC includes variable costs in its initial allocation and then again when allocating remaining revenues. The less below-cost traffic a complainant includes in its SARR, the more irrational Modified ATC becomes. In other words, Modified ATC over-corrects the hypothetical problems created by ATC in cases such as this in which WFA has all but eliminated the traffic that produces the problem. The Board also argues that BNSF forfeited its proportionality argument by failing to present it earlier. The record belies this contention. BNSF has presented the basics of its proportionality argument throughout these proceedings. In its Petition for Reconsideration of the Board’s 2007 Decision, BNSF argued that “even if the Board’s concern 9 about the effect of ATC on low rated traffic justified suspension of the average total cost approach on that traffic, there is no conceivable justification for applying the modified ATC methodology to all cross-over traffic.” Petition for Reconsideration at 3 (emphasis added). And BNSF noted that “for movements that generate [Revenue/Variable Cost] ratios well in excess of variable costs, there is no risk that ATC will allocate to the SARR less than the incumbent’s . . . variable costs.” Id. at 19. BNSF adjusted its argument when WFA changed its traffic group, but the point remained. The Board had justified its adherence to Modified ATC because “it was unwilling to apply a methodology that risked allocating revenues below the costs incurred by the SARR,” but “[w]ith that risk removed by the reformulated traffic group, there [was] no basis for continuing to apply a modified ATC methodology . . . .” BNSF’s Third Supplemental Reply Evidence at III.A-22 (July 14, 2008). BNSF preserved this argument in its petition in WFA I. In its opening brief, BNSF argued that the Board had failed to provide any reasoned “explanation for persisting to apply the Modified ATC approach after WFA had overhauled its case to eliminate all low-rated traffic.” The Board argued in response that WFA had not excluded all such movements— apparently three remained—and thus the Board was justified in maintaining Modified ATC. BNSF’s point survived nonetheless. “By . . . applying Modified ATC to the traffic WFA selected, the Board allocated to the SARR an unwarranted increase in revenue.” Though this challenge was not articulated in terms of “proportionality,” it represents the basics of BNSF’s argument on remand. And had we ruled on the merits of BNSF’s challenges in WFA I, we would have had to have approved the continued application of Modified ATC to WFA’s revised traffic group, both categorically, and as a proportionate response to the problems with ATC. We 10 never reached this argument in WFA I due to the remand, but it is and always has been inherent in BNSF’s double-counting critique. In short, the Board erred in its failure to address BNSF’s proportionality challenge on remand. As we noted above, we review the Board’s decision under the “arbitrary and capricious” standard drawn from the Administrative Procedure Act. While it is a forgiving standard, it does not create a rubberstamp. We remind the Board on remand that the APA requires that it “examine the relevant data and articulate 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 Inc. v. United States, 371 U.S. 156, 168 (1962)). We further remind the Board that “an agency’s ‘failure to respond meaningfully’ to objections raised by a party renders its decision arbitrary and capricious.” PSEG Energy Resources & Trade LLC v. FERC, 665 F.3d 203, 208 (D.C. Cir. 2011) (quoting Canadian Ass’n of Petroleum Producers v. FERC, 254 F.3d 289, 299 (D.C. Cir 2001) (other citations omitted)). If it is true, as Petitioner asserts, that the Board has adopted an alternative revenue allocation method applicable to all future cases, we would expect its opinion to advise why that method is not equally applicable to this case. While we do not suggest that all such changes must be made retroactively, we must at least know that the Board has exercised reason, not arbitrariness and capriciousness, in treating this Petitioner differently. 11 III. CONCLUSION For the reasons set forth above, we grant the petition, vacate the order, and remand the matter to the Surface Transportation Board for further proceedings consistent with this opinion. RANDOLPH, Senior Circuit Judge, dissenting: I would deny the petition for review. The last time BNSF Railway Company petitioned for judicial review in this matter, our court remanded the case to the Surface Transportation Board. We disposed of BNSF’s petitions in these words: “Accordingly, we grant the petitions in part, so that the Board on remand can address BNSF’s double- counting objection to modified ATC, and we otherwise deny the petitions.” BNSF Ry. Co. v. Surface Transp. Bd., 604 F.3d 602, 613 (D.C. Cir. 2010). Anyone reading this sentence would think the court was remanding the case for the Board to perform one function, and one function only—rule on BNSF’s double- counting objection. “[W]e otherwise deny the petitions” can only mean that the court was rejecting the other arguments BNSF made in its petitions. Yet the majority tells us that the opinion’s closing paragraph opened the door to allow BNSF to raise a host of new objections to Modified ATC. Maj. Op. at 7- 8. With respect, I think that is mistaken. The language of disposition, usually found at the end of a judicial opinion, should be clear enough that the parties do not have to guess about what the court has decided. In this case, if the court had remanded to the Board to consider or reconsider several issues one would have expected the court to have said so. That is, if the court had wanted the Board to address BNSF’s other substantive arguments it would have specified them, as the court did for example in Somerset Welding & Steel, Inc. v. National Labor Relations Board, 987 F.2d 777, 782 (D.C. Cir. 1993). I admit that we are sometimes less precise, as when we remand for “proceedings consistent with [our] opinion,” see, e.g., City of Cleveland v. Fed. Power Comm’n, 525 F.2d 845, 857 (D.C. Cir. 1976), disapproved on other grounds, Ark. La. Gas Co. v. Hall, 453 U.S. 571 (1981), or when we direct the agency to address “first” a particular issue, which implies that the agency may also consider other issues, see, e.g., John Cuneo, Inc. v. Nat’l Labor Relations Bd., 792 F.2d 1181, 1184 (D.C. 2 Cir. 1986). But there was no imprecision here—the court remanded to the Board for it to “address” one argument, and otherwise denied the petitions. The Board was therefore given a “narrow task.” Wash. Gas Light Co. v. Fed. Energy Regulatory Comm’n, 603 F.3d 55, 56 (D.C. Cir. 2010). The Board fully complied when, on remand, it addressed and rejected BNSF’s double-counting objection. W. Fuels Ass’n, Inc. v. BNSF Ry. Co., Dkt. No. NOR 42088, 2012 WL 2194142, at *9-10 (Surface Transp. Bd. June 13, 2012) (decision). BNSF now barely challenges that ruling. Even if the limited remand order required the Board to address more than double-counting I would still deny the petition for review. BNSF forfeited its current slate of arguments when it failed to raise them in its previous petitions for review. See Nw. Ind. Tel. Co., Inc. v. FCC, 872 F.2d 465, 470-71 (D.C. Cir. 1989). The Board therefore had the discretion to decline to evaluate BNSF’s newly-minted arguments or to hold them to the statutory reopening standard. The majority nowhere shows where or when BNSF had raised its Alternative ATC proposal before the proceedings on remand. Nothing in BNSF’s previous petitions for review even hints at the Alternative ATC approach. Instead, the majority points to BNSF’s argument that Modified ATC became unnecessary once Western Fuels eliminated low-rated traffic from its model. Maj. Op. at 9-10. That might be called a proportionality argument of sorts, but it is not the same one BNSF raises here. BNSF’s earlier argument would allocate revenue differently depending on the traffic selection in a shipper’s model. BNSF now suggests, for all models, that there is a revenue allocation method that better balances the Board’s competing concerns. By changing the object of comparison, BNSF created an altogether new argument because the Board’s decisions about revenue allocation rise or fall with the quality of 3 the suggested alternative. See BNSF Ry. Co. v. Surface Transp. Bd., 453 F.3d 473, 483-84 (D.C. Cir. 2006). I also believe the Board acted within its discretion when it declined to hold the case in abeyance. Maj. Op. at 6. I have serious doubts whether an agency’s decision to hold a case in abeyance is even judicially reviewable. Such docket- management choices are the sort of discretionary “decision[s] to structure the proceedings” that are left in the agency’s hands. See Vt. Yankee Nuclear Power Corp. v. Natural Res. Def. Council, 435 U.S. 519, 543-47 (1978). And even if an abeyance decision is reviewable, I would sustain the Board’s decision. The Board was sensibly concerned with rising litigation costs and the need for finality. These concerns would grow even more pressing if this decades-old case dragged on.