Decision Information

Docket Number:  

Case Title:  

Decision Type:  

Deciding Body:  
Entire Board

    Decision Summary

Decision Notes:  

    Decision Attachments

166 KB

Approximate download time at 28.8 kb: 1 Minutes

If you do not have Acrobat Reader, or if you have problems reading our files with your current version of Acrobat Reader, the latest version of Acrobat Reader is available free at

    Full Text of Decision


42453 SERVICE DATE – JUNE 15, 2012






Docket No. NOR 42088







Digest:[1] In 2009, the Board found that the BNSF Railway Company was overcharging the utility in this case and ordered the railroad to lower its transportation rates and to reimburse the utility for past overcharges. A federal appeals court affirmed most of that decision. However, the court remanded the matter to the agency to explain more thoroughly why it used a contested method to allocate revenues from cross-over traffic in its rate case analysis. In this decision, the agency explains that it used the method to correct an unanticipated problem with the prior method that was allocating revenue below the Board’s measure of variable costs.


Decided: June 13, 2012


In 2004, Western Fuels Association and Basin Electric Power Cooperative (collectively, WFA) filed a complaint against BNSF Railway Company (BNSF) alleging that BNSF was charging unreasonably high rates for transporting coal to WFA’s power plant. WFA presented evidence under the then-existing version of the Board’s stand-alone cost (SAC) rate reasonableness test. While the case was pending, the Board changed certain aspects of the SAC test in a separate rulemaking. Major Issues in Rail Rate Cases (Major Issues), EP 657 (Sub-No. 1), slip op. at 25-31 (STB served Oct. 2006), aff’d sub nom. BNSF v. STB, 526 F.3d 770 (D.C. Cir. 2008). One change involved the method for determining the revenue earned by the defendant on so-called “cross-over traffic.” In the rulemaking, the Board adopted a new approach that it called the average total cost (ATC) methodology.


In a 2007 decision in WFA’s case, which was the first case to apply ATC since the conclusion of Major Issues, the Board determined that, to ensure the SAC test comported with fundamental economic principles, a modification of ATC was necessary. The Board also permitted WFA to revise its SAC presentation in response to the changes to the SAC test brought about by the rulemaking. Applying the updated SAC test, including the modified version of ATC, to WFA’s revised case in 2009, the Board determined that BNSF’s rates were unreasonable and ordered reparations for past overcharges and maximum rate levels for future shipments.


BNSF challenged that decision in court, contending, among other things, that the Board’s decision to adopt the modified ATC methodology was improper. The court rejected BNSF’s other arguments but concluded that the Board had not adequately addressed BNSF’s argument relating to the modification of ATC and remanded the matter to the Board. BNSF Ry. v. STB, 604 F.3d 602, 613 (D.C. Cir. 2010).


After considering the parties’ submissions on remand, we reaffirm here that the modified ATC methodology is preferable to the original ATC methodology. Although the Board adopted original ATC to account for economies of density in the cross-over traffic revenue allocation method, it had failed to take into account the fundamental economic principle that traffic must cover its variable costs before it can make any contribution to joint and common costs.



WFA owns a coal-fired electric generating station that depends on coal delivered by BNSF. In 2004, WFA filed a complaint challenging BNSF’s rates, seeking to prove its case under the Board’s SAC test.


By way of background, “the SAC test protects a captive shipper from cross-subsidizing other traffic, bearing inefficiencies, or paying more than the revenue needed to replicate rail service to a select subset of a carrier’s traffic base.” Duke Energy Corp. v. CSX Transp., 7 S.T.B. 402, 412 (2004); see also Arizona Elec. Power Coop., Inc. v. BNSF Ry., NOR 42113, slip op. at 4-5 (STB served Nov. 22, 2011). The SAC test requires hypothesizing a new, optimally efficient carrier—a stand-alone railroad (SARR)—operating in a market with no barriers to entry or exit that would serve the traffic to which the challenged rate applies (referred to as the issue traffic) and, at the complainant’s option, other traffic sharing lines and facilities with the issue traffic. Id. at 413. To make this modeling task easier, the Board permits complainants to design smaller SARRs that use something called “cross-over traffic,” which is non-issue traffic that would traverse the SARR for part of the movement and the residual incumbent railroad for the remainder of the movement. Bituminous Coal—Hiawatha, Utah, to Moapa, Nev., 10 I.C.C. 2d 259, 265-68 (1994). Based on a detailed operating plan, the SARR’s system-wide investment requirements and operating expense requirements must be estimated. From this, it is possible to calculate what the SARR would need to charge the selected traffic group to cover its costs, including a reasonable rate of return. If that amount is less than what the defendant carrier is actually charging the traffic group for the service replicated by the SARR, the traffic group as a whole is being overcharged. Subject to the Board’s method for apportioning relief among the movements in the traffic group, the complaining party may be entitled to rate relief. Pub. Serv. Co. of Colo. v. Burlington N. & Santa Fe. Ry., 7 S.T.B. 589, 598-99 (2004).


In 2006, WFA submitted its evidence, including its proposed SARR, which consisted almost entirely of BNSF’s Orin Subdivision, the highest density rail line in the nation. After BNSF and WFA completed their evidentiary presentations, the Board placed WFA’s case in abeyance pending the outcome of a rulemaking proceeding to change various aspects of the SAC test. Major Issues in Rail Rate Cases, EP 657 (Sub-No. 1), slip op. at 2 (STB served Feb. 27, 2006). One of the changes ultimately adopted by the Board involved its revenue-allocation methodology—the way it allocates revenue from cross-over traffic between the on- and off-SARR segments of the movement. Up until that time, the Board had consistently used some form of mileage-based methodology, the latest version being the Modified Straight Mileage Prorate (MSP); in the rulemaking, the Board replaced MSP with ATC, which allocates revenues in proportion to the defendant’s average total cost of providing service on the various segments in question. Major Issues, slip op. at 25-31. In moving to ATC, the Board sought to avoid bias that may arise from the use of cross-over traffic. That is, the Board sought to ensure that a “SAC analysis using cross-over traffic will approximate the outcome of a full SAC analysis, which provides origin-to-destination service for the entire traffic group.” Id. at 24. The Board determined that it would apply the new revenue-allocation methodology to pending cases like WFA’s. Id. at 75-76.


In September 2007, the Board issued a decision finding that, based on the record before it, WFA had not shown that BNSF’s rates were unreasonable. W. Fuels Ass’n v. BNSF Ry. (2007 Decision), FD 42088 (STB served Sept. 7, 2007). As instructed by the Board following Major Issues, the parties had applied ATC to the cross-over movements’ total revenues. For a substantial number of these movements, the result of doing so was to drive below 100% the revenue-to-variable cost (R/VC) percentages—as measured by the Board’s Uniform Rail Costing System (URCS)—for the on-SARR portion. This occurred because of two factors. First, WFA had included considerable cross-over traffic in its traffic group with total revenue either below or barely above the variable costs of handling the traffic. Second, the off-SARR segments of these movements had lower traffic densities, and thus higher average total costs. By allocating revenues from these movements in proportion to average total costs, as required by ATC, a proportionally larger percentage of that revenue was allocated to the off-SARR segment. Id. at 14. The upshot, the Board stated, was that “the on-SARR revenue allocation for those movements would be insufficient to cover the variable costs (as calculated by URCS) of handling traffic for the highest-density portion of a movement.” Id. This result, the Board said, was unintended and illogical because “[t]raffic must cover its variable costs before it can be expected to make any contribution to joint and common costs.” Id.


To avoid allocating revenues at levels below variable costs as measured by URCS, the Board determined that it had to refine the ATC approach. Rather than applying ATC to total revenue, the Board concluded that it would apply ATC to total revenue contribution, i.e., revenue in excess of variable costs, as measured by URCS. Id. Under this modified ATC approach, allocating revenue from cross-over traffic would involve a two-step process. First, sufficient revenue would be allocated to each segment to cover that segment’s variable costs of providing service, as measured by URCS. Second, remaining revenues, if any, would be allocated using the original ATC methodology. This is the approach that the Board used in the 2007 Decision when it determined that WFA had not proven its case.


Although the Board found that WFA had not proven its case “on this record,” the Board found persuasive WFA’s claim that, had it known that the Board would replace the MSP revenue allocation methodology with ATC, it would have submitted a fundamentally different case. The Board thus offered WFA the opportunity to revise its evidentiary presentation by changing the traffic group or configuration for its SARR. Id. at 20. WFA chose to avail itself of that opportunity.


BNSF sought reconsideration of the 2007 Decision. In addition to challenging WFA’s option to revise its SAC presentation, BNSF also argued that modified ATC is inconsistent with the fundamental principle underlying ATC, which, BNSF said, was to simplify the SAC analysis without introducing bias. Using two hypothetical examples, BNSF contended that modified ATC dilutes the impact of traffic density by essentially double-counting variable costs, thereby reintroducing the very bias that the Board had sought to avoid with ATC. By making high-density lines more attractive to replicate in a SARR than light-density lines, BNSF argued, modified ATC biases the SAC analysis in favor of complainants who design shorter, high-density SARRs that use cross-over traffic.


BNSF also disputed the Board’s justification for adopting modified ATC. BNSF maintained that by allocating revenues in a way that reflects the density of the traffic on the segments in question, ATC accomplishes exactly what the Board intended. BNSF acknowledged that allocating revenues in this way on certain low-rated traffic (i.e., traffic that was charged relatively low rates by the carrier) might not allow the SARR to cover the incumbent’s variable costs, as measured by URCS. But BNSF urged two reasons why this did not raise valid concerns. First, SARRs are designed to be optimally efficient, so it would be reasonable to expect a SARR’s variable costs to be lower than those of the incumbent. Second, if a complainant believes that a particular low-rated movement will not contribute to fixed costs, it has the option to exclude that movement from its traffic group. Finally, BNSF stated that, even if the Board were rightly concerned with ensuring a revenue allocation under which the SARR would cover the incumbent’s variable costs on each movement, the Board had no valid reason to apply modified ATC to those cross-over movements that, in fact, covered those costs under the original ATC methodology.


In 2008, the Board denied BNSF’s petition for reconsideration. W. Fuels Ass’n v. BNSF Ry. (2008 Decision), NOR 42088 (STB served Feb. 29, 2008). The Board stated that BNSF’s first hypothetical example demonstrated the problem with original ATC—i.e., that on certain low-rated movements, the allocation of revenue to the high-density, on-SARR portion of a movement would be less than the variable costs for that segment. Id. at 4. The Board rejected BNSF’s argument that this was not a problem, given the SARR’s presumed greater efficiency and the complainant’s ability to simply exclude such movements from its traffic group. The Board explained that the fairness of a revenue-allocation procedure should not depend on either the complainant having to design a SARR that is more efficient than the incumbent railroad or the complainant having to drop traffic from its traffic group that the defendant finds worthwhile to handle at the current rate. Id. at 4-5.


WFA submitted its revised SAC analysis in 2008. In its revised presentation, WFA changed both its traffic group and the physical layout of its SARR. As to the former, WFA dropped most, but not all, of the highly competitive, low-rated traffic that it had included in its initial traffic group and substituted less competitive, higher-rated traffic. As to the latter, WFA expanded the geographic scope of its SARR beyond the extremely high density Orin Subdivision. In reply, BNSF renewed its challenge to the modified ATC methodology. BNSF argued that WFA had taken the opportunity to change its traffic group to exclude low-rated movements that were unprofitable under ATC, thus making it unnecessary for the Board to apply modified ATC to the remaining traffic.


In a February 2009 decision, the Board found that the challenged rates were unreasonable and ordered reparations and set maximum rate levels through 2024. W. Fuels Ass’n v. BNSF Ry. (2009 Decision), NOR 42088, slip op. at 13 (STB served Feb. 18, 2009). In that decision, the Board also rejected BNSF’s renewed challenge to modified ATC. The Board acknowledged that WFA had changed its traffic group to include “less traffic” with revenue at or near its variable costs. Id. According to the Board, however, the modification was still necessary, noting that the Board sought a uniform approach to allocating revenue from cross-over traffic that could also be used in Simplified-SAC cases.[2] 2009 Decision, slip op. at 13.


BNSF appealed this decision to the U.S. Court of Appeals for the District of Columbia Circuit. The court rejected all but one of BNSF’s arguments. The court determined that the Board had not addressed BNSF’s argument that modified ATC improperly double counts variable costs. The court therefore remanded the case to the Board for it to address that argument. BNSF Ry. v. STB, 604 F.3d 602, 613 (D.C. Cir. 2010). On remand, the Board accepted comments from the parties on the double-count issue. BNSF has moved to strike a portion of WFA’s comments as constituting an impermissible collateral attack on the principles announced in Major Issues.[3]




Original ATC was designed to provide an unbiased revenue allocation approach that would, to the extent practicable, approximate the result that would be obtained if the complaining shipper had submitted a SAC analysis that modeled origin-to-destination service for the entire traffic group. But when the Board applied original ATC to WFA’s initial traffic group, it discovered an unanticipated problem. The problem was that, for certain competitive traffic with low R/VC ratios, original ATC led to revenue allocations that did not cover BNSF’s variable costs of handling the traffic over high-density segments. That result was illogical and unintended. The Board explained that traffic must cover its variable costs before it can be expected to make any contribution to joint and common costs. 2007 Decision, slip op. at 14. The Board further explained that it had not contemplated this situation and that such a result (a revenue allocation below variable costs) “would plainly conflict with our express purpose to find a non-biased, cost-based method.” Id. To resolve this problem, the Board modified the revenue allocation approach to allocate revenue contribution (total revenue less variable costs) instead of total revenue in accordance with ATC.


On remand, BNSF raises three objections to the use of modified ATC in this case. First, BNSF asks us to reconsider our finding that there is a problem with original ATC, arguing that a revenue allocation that falls below our measure of variable costs is not a problem. Second, it argues our solution is biased in favor of high-density SARR lines because it double-counts or overstates variable costs, thereby understating the role that economies of density should play in the revenue allocation approach. Finally, BNSF suggests that if a problem with original ATC exists, modified ATC deals with what is a limited problem in a way that is disproportionate to the size of the problem and that there was an alternative, superior solution. We address each objection in turn below, and our conclusion to continue to use modified ATC in this case


  1. The Problem with Original ATC


We continue to find BNSF’s first objection unconvincing. The Board found that there was a problem with original ATC that we must address. The purpose of a SAC analysis is “to determine whether a complainant is bearing the cost of any inefficiencies or the cost of any facilities or services from which it derives no benefits.” 2007 Decision, slip op. at 8. To make this determination, the Board must ascertain the extent to which other traffic in the traffic group (as relevant here, cross-over traffic) can contribute to the cost of facilities or services shared with the issue traffic. But any particular movement can make such a contribution only if it first covers its variable costs. As the Board said in adopting modified ATC, it is a fundamental economic principle that “[t]raffic must cover its variable costs before it can be expected to make any contribution to joint and fixed costs.” Id. at 14.


We repeat a simple illustration, drawn from an example that BNSF used in its petition for reconsideration of the 2007 Decision, to demonstrate the problem. In the example, BNSF posits a cross-over movement that generates revenue of $100, but for which the defendant carrier’s variable costs of providing service from origin to destination is also $100. BNSF thus posits a movement that can provide no contribution to any of the joint and common costs the defendant incurs to provide the origin-to-destination service. But, under the assumptions BNSF makes in the example about traffic densities and other matters, the revenue allocation under original ATC implies that this movement could make a contribution to joint and common costs.[4] Under original ATC, the revenue allocated to one section would exceed the variable costs, suggesting the movement could offset some of the joint and common costs along that segment; while the other segment would receive a revenue allocation below variable costs, suggesting not only that no revenue is available to offset joint and common costs along that segment (which is true), but that the movement is a revenue drain (which is not true). This example shows that original ATC can lead to a revenue allocation that defies facts, is illogical, and is not what the Board intended. It was this sort of illogical allocation of revenue that led the Board in the 2007 Decision to adopt modified ATC.


It remains the case that the use of original ATC could create the illusion that either more revenue is available to help pay for joint and common costs over light density segments than is available in reality, or that the movement is a revenue drain (with revenue below variable costs) on the heavy density segments, when it is not. This is because the original ATC methodology presumes the revenue allocated to a particular segment can fall below the variable costs of providing service over that segment. We do not believe that a railroad would rely upon such a presumption in real world operations. Consider a rail carrier trying to place a value on a segment of its system. Part of such a task would involve estimating the revenues that are attributable to the segment in question. For each movement over the segment, presumably the carrier would, 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.


BNSF claims there is no problem with original ATC because the Board is using URCS to measure variable costs for the SARR, and the Board has previously stated that “an R/VC ratio below 100% does not necessarily reflect improper pricing or money-losing service.”[5] BNSF also notes that the Board has acknowledged that “URCS costs may include a significant portion of what may actually be unattributable joint and common costs.”[6] Hence, the suggestion by BNSF is that, even if the cross-over traffic does not appear to cover its variable costs as measured by URCS, that does not mean that that traffic does not actually cover its true variable costs.


URCS is the Board’s general purpose costing model for all regulatory costing purposes. As such, we reject BNSF’s argument that we should permit the revenue allocation to fall below that measure of the variable costs in a SAC analysis on the alleged ground that URCS does not properly measure true variable costs. Granted, for a short period of time, a carrier might price certain traffic below 100% of its URCS variable costs. But this is true in part because “URCS is not a measure of short-run variable costs or the marginal cost of hauling rail traffic.” E.I. DuPont De Nemours & Co. v. CSX Transp., Inc., NOR 42099, slip op. at 17 (STB served June 30, 2008). Instead, URCS has a longer time horizon. Although some of the costs included in URCS, such as rails and ties, do not vary with the amount of traffic in the short term, they do vary with differing traffic levels over a longer time period, and URCS reflects this fact. Thus, URCS is “a measure of intermediate variable costs, on a system-average basis, that includes costs (such as return on road property investment) that are fixed in the short term,” id., but variable over the longer term.[7] Likewise, a SAC analysis seeks to determine the maximum reasonable rate over the long term. W. Tex. Utils. Co. v. Burlington N. & Santa Fe Ry., NOR 41191, slip op. at 6 (STB served Sept. 10, 2007) (noting the “long-term nature of a SAC analysis”). A revenue allocation that does not cover our regulatory estimate of variable costs of a particular movement—as established by URCS—presents a problem in a SAC analysis. Once that problem was identified, the Board acted reasonably in addressing it by adopting modified ATC.


Alternatively, BNSF repeats its argument that we can ignore the distortion created by the original ATC because the complainant in a SAC case is permitted to exclude from its traffic group any traffic that, under original ATC, would not receive an allocation that covers its URCS variable costs.[8] Indeed, according to BNSF, WFA exercised that ability and excluded almost all such traffic from its traffic group.


We remain unpersuaded by this argument. Where the Board identifies a problem that it can address directly, it is not appropriate instead to put the burden on the complainant to remedy the problem by having to choose either to include the traffic and receive a below-URCS revenue allocation, or exclude from its group traffic movements that the incumbent railroad finds worthwhile to handle at the current rate. See 2008 Decision, slip op. at 4-5. Notably, although WFA revised its traffic group, it did not eliminate all traffic that, under original ATC, would have a revenue allocation below 100% of variable costs. Id. at 13. Accordingly, the Board’s use of modified ATC was still necessary to avoid that problem.


As important, the selection of the revenue allocation methodology should not be decided in individual cases based on how many problem movements (i.e., cross-over movements with a revenue allocation below variable costs) are in the traffic group. Implicit in BNSF’s argument is that our choice of revenue allocation methodology should depend in an individual case on the number of problem movements with a revenue allocation below variable costs. However, it would be arbitrary for the agency to flip flop between original ATC and modified ATC from one case to another, when a uniform approach creates a comprehensive solution. It likewise would create improper incentives to place the choice of revenue allocation in the hands of the complainant, who already selects the traffic group in SAC cases; a complainant that includes considerable problem movements would use modified ATC while a complainant that excludes such traffic would use original ATC. We are also concerned about the issue of determining how much problem traffic is too much—i.e., how much traffic with revenue allocations below variable costs need to arise in a particular case before the Board needs to decide to use the modified ATC approach instead of original ATC.


We remain convinced that there is a problem with the original ATC approach. BNSF has not persuaded us that we can ignore the unanticipated and illogical feature of the revenue allocation methodology that may drive the revenue allocation below variable costs, as measured by URCS. In adopting modified ATC, we also sought a solution that addresses the problem, while maintaining a uniformly applied revenue-allocation methodology for both SAC and Simplified-SAC cases, where complaining shippers have no discretion over the composition of the traffic group. [9]


  1. The Alleged Double Count


Second, BNSF contends that modified ATC introduced bias into the SAC analysis by overstating the role of variable costs and thus understating the role economies of density should play in the revenue allocation. BNSF often refers to this issue as a “double count.” As did the reviewing court, we interpret this “double count” claim to be that “modified ATC fails appropriately to consider economies of density and artificially inflates the revenues attributable to the SARR.” BNSF Ry. v. STB, 604 F.3d at 612. We reject BNSF’s argument. The Board has determined that modified ATC strikes a more appropriate balance than original ATC between sound revenue allocation and accounting for economies of density.


According to BNSF, the old mileage-based methodology, MSP, produced distorted and biased results in favor of complaining shippers who designed short, high-density SARRs, and ATC removed those distortions and biases.[10] However, BNSF contends, modified ATC “double counts” variable costs and thereby “reintroduces the bias” in favor of short, high-density SARRs that existed under MSP.[11] As such, BNSF maintains that modified ATC “understates[s]” the role that economies of density should play in allocating revenues from cross-over traffic[12] and thereby departs from the goal of permitting a truncated SARR to approximate a Full-SARR analysis. BNSF argues that allocating total revenue (and not total contribution) in accordance with average total costs is the sole measure of whether the approach is unbiased. BNSF’s logic is as follows: the Board said in Major Issues that an unbiased approach should allocate total revenue in accordance with ATC; modified ATC does not do that because it first backs out the variable costs and then allocates the revenue contribution in accordance with ATC; and, therefore, modified ATC is biased.


We find BNSF’s arguments unpersuasive. Original ATC allocates total revenues from cross-over traffic in accordance with the relationship between the total costs of providing service over the various segments at issue. When it first adopted original ATC, the Board believed that such a revenue allocation would provide an unbiased method that would account for economies of density and diminishing returns, but, as discussed above, has now identified and taken steps to remedy problems that became apparent. In other words, the decision to modify original ATC in this case was premised on the agency’s conclusion that there was an unanticipated flaw in original ATC.


Based on our additional experience in this proceeding, we see now that there are two competing principles in play. A revenue allocation method should take into account the important role that economies of density play in any cost-based revenue allocation approach. At the same time, however, it should not create the implausible result of driving the revenue allocation below variable costs. But the example we offered above illustrates that elevating the first principle at the expense of the second is unreasonable. In that example, where the revenue from the traffic equaled the variable costs, the only reasonable allocation of revenue would be in strict accordance with the relative variable costs, with no weight given to economies of density whatsoever. This is because, as explained above, in that example there was no revenue from that traffic that can pay for joint and common costs along the segments in question. The movement would be paying only for its own variable costs. In our judgment, strict adherence to the first principle at the expense of the second does not provide the unbiased, cost-based approach needed to achieve our ultimate goal of permitting a SAC analysis with cross-over traffic to replicate as reasonably as possible the results of a SAC analysis without cross-over traffic. Although modified ATC does not give the same weight to economies of density as original ATC did, it still accounts for that element, while striking a more appropriate balancing of the principles discussed above.


In sum, modified ATC was the Board’s attempt to accommodate these two, at times, conflicting principles. By first allocating revenue to each section to cover variable costs, we accommodate the first principle by ensuring that we not drive revenue allocation below variable costs. This avoids the potential illusion that either more revenue was available to help pay for joint and common costs over light density segments than is available in reality, or that the movement is a revenue drain (with revenue below variable costs) on the heavy density segments when it is not. By then allocating the remaining contribution, we accommodate the important role that economies of density should play in the revenue-allocation method.


The process of ratemaking is inherently one of periodic review and improvement. While we may discover at some later point that other principles should be factored into our analysis, until our adjudicatory or rulemaking process yields a demonstrably superior alternative, we conclude that modified ATC reflects a needed refinement over original ATC.


  1. BNSF’s “Disproportionate” Remedy Argument and New Alternative Solution


Finally, we address two related arguments by BNSF on remand. BNSF argues that modified ATC deals with a limited problem in a way that “is vastly disproportionate to the size of the problem and completely unfair.”[13] BNSF therefore maintains that modified ATC was not a rational response to the problem the Board found with original ATC. To the extent there is a problem (a point BNSF contests), BNSF maintains there is a better way of addressing it. This “alternative ATC” methodology would have two steps. The Board would first apply original ATC to all movements. Then, for those movements that received on-SARR revenue allocations below the defendant’s variable costs, as measured by URCS, for the movement over the on-SARR segment, the Board would allocate additional revenues to the on-SARR segment based on the relative on-SARR and off-SARR variable costs up to 100%. While BNSF identifies this new alternative approach, BNSF “does not advocate its use.”[14]


WFA objects to the use of this alternative ATC approach.[15] Further, WFA argues that if the Board were to use this alternative ATC approach, then the Board must give WFA the opportunity to revise its current SAC analysis to address the change from modified ATC to alternative ATC, just as the Board gave WFA the opportunity to revise its SAC analysis to address the change from original to modified ATC.[16]


In remanding the case, the court did not direct the Board to address either of these issues BNSF has raised for the first time on remand: the disproportionate-remedy argument and the proposed alternative ATC method.[17] However, the agency has some discretion to broaden the issues on remand. To guide our decision-making, we will look to the statutory reopening standard. E.g., Market Dominance Determinations, 6 S.T.B. 492, 495-97 (2004). This places BNSF in no better and no worse position than if, instead of appealing, it had filed a petition to reopen the 2009 Decision so that we could consider its new disproportionate-remedy argument or its alternative ATC approach.


Under 49 U.S.C.  722(c), we may reopen a proceeding and reconsider a prior decision because of “material error, new evidence, or substantially changed circumstances.” BNSF has not met any of the criteria for reopening. First, the Board did not commit material error in failing to accept BNSF’s disproportionate-remedy argument or in refusing to adopt the alternative ATC approach. The agency does not commit material error in failing to accept an argument that was not advanced prior to its decision on the merits. BNSF Ry. v. STB, 453 F.3d 473, 479 (D.C. Cir. 2006). In seeking reconsideration, BNSF urged the Board to return to original ATC, but in doing so, BNSF did not argue that modified ATC was a vastly disproportionate response to the problem of revenue allocations below 100% of variable costs. And although BNSF alluded to the alternative that it has now identified on remand,[18] it failed to present that alternative in adequate detail.[19] Second, neither the disproportionate-remedy argument nor the alternative ATC approach qualifies as “new evidence.” New evidence refers to evidence that was not reasonably available to the party when the record was developed, not simply newly raised. Toledo, Peoria & W. Ry. v. STB, 462 F.3d 734, 753 (7th Cir. 2006). BNSF could have raised these matters before the Board’s 2009 decision yet failed to do so. Finally, BNSF points to no materially changed circumstances, nor are we aware of any, that justify BNSF raising these matters for the first time on remand. The Board has similarly denied shipper and carrier requests to reconsider or reopen other rate cases to address issues of substantial magnitude. Cf. PPL Montana, LLC v. The Burlington N. & Santa Fe Ry., 6 S.T.B. 752, 758-61 (2003) (denying request to reopen record to allow shipper to submit different evidence in response to the Board’s first application of its internal cross-subsidy test); Tex. Mun. Power Agency v. Burlington N. & Santa Fe Ry., 7 S.T.B. 803, 805 (2004) (“Nothing in the statute or the Board’s regulations obliges the agency to rethink its decisions whenever a party wishes to try out a new theory or finds new information at a late stage in the process. . . . [I]f a party were free to reshape its case, . . . the administrative process might never end.”).


In the end, modified ATC was the Board’s solution to accommodate the two competing principles described earlier in this decision. We do not suggest that this is the only solution or that there may not be other approaches that could better accommodate the two competing principles. Indeed, the Board is planning to begin a rulemaking proceeding to consider whether a methodology similar to BNSF’s alternative ATC might be just such an approach.


We considered placing this case in abeyance (again) while that rulemaking runs its course, but decided against doing so for three reasons. First, having concluded that the alternative ATC approach was raised too late, we will not hold this case in abeyance to await the outcome of the rulemaking because that would encourage future litigants to also try out new theories at late stages in the process. Second, we have not yet finalized our proposal, made it public, or received comments on it; thus, we cannot be sure of the precise nature of our planned rule. Given this uncertainty, and with the case approaching the eight-year mark, considerations of administrative finality come increasingly to the fore. Third, if the Board were to adopt something like BNSF’s alternative ATC as a result of the forthcoming rulemaking, that could lead to still more litigation. WFA has argued that, if the Board applies BNSF’s alternative ATC to its revised SARR, WFA would be entitled to revise its SARR once again under Hatch v. FERC, 654 F.2d 825 (D.C. Cir. 1981). Litigation must come to an end at some point, and, for the reasons discussed, that point has been reached here.


It is ordered:


1. This proceeding is discontinued.


2. This decision is effective on its service date July 15, 2012.


By the Board, Chairman Elliott, Vice Chairman Mulvey, and Commissioner Begeman. Commissioner Begeman dissented with a separate expression.






It is regrettable that the WFA case has been such a time-consuming process for all parties involved. While I would like this case resolved once and for all as soon as possible, particularly given that the court remanded this matter to the Board more than two years ago, I cannot ignore that valid concerns have been raised over what can be a critical factor used in determining the outcome of a rate challenge. And those concerns were raised throughout the course of this case, not merely on remand.


I do not believe that modified ATC, which was developed after the conclusion of Major Issues, and without an opportunity for public comment, provides for the unbiased revenue allocation approach that was intended. I also do not believe that because the parties in this case have endured such a lengthy process, in part due to actions by the Board, that we can justify taking a pass here. Furthermore, I do not see why if it was appropriate to hold the case in abeyance when the Board was creating original ATC in Major Issues, that it is not also appropriate to do so now for a proceeding to address the very problems posed as a result of ATC and modified ATC.


As suggested in the majority’s decision, where the Board identifies a problem that it can address directly, it is not appropriate instead to put the burden on the complainant to remedy the problem. I agree. By the same token, it would not be appropriate to put the burden on the defendant to remedy a problem with the Board’s rate processes. Yet that is just what the majority’s decision does when it argues that BNSF’s objections to the flawed modified ATC methodology were not accompanied by a detailed alternative proposal. Indeed, it is the agency’s ultimate responsibility and duty to develop a superior alternative. That is why I believe the Board should initiate a fast-track proceeding to take public comment from interested parties in an effort to determine the best methodology based on economic principles for allocating cross-over traffic revenues, to the extent that such traffic is appropriate in rate cases. The methodology that results should then be applied to this case.


I am committed to ensuring the Board provides a fair and accessible venue for resolving rate disputes in a timely manner. In that effort, the Board must remain committed to ensuring the integrity of the rate case processes, as they can be the only available option for a shipper to obtain appropriate rate relief. I cannot support maintaining a questionable allocation methodology for this case, while at the same time announcing plans to begin a rulemaking proceeding to develop a superior alternative (based on BNSF’s proposal) that would only be applied to future cases.


The time to find the best revenue allocation approach is long overdue. It can and should be resolved for this and future cases.



[1] The digest constitutes no part of the decision of the Board but has been prepared for the convenience of the reader. It may not be cited to or relied upon as precedent. Policy Statement on Plain Language Digests in Decisions, EP 696 (STB served Sept. 2, 2010).

[2] The Board has sought to make its rail rate dispute resolution procedures more affordable and accessible through the adoption of a simplified version of the SAC test for shippers of medium and smaller size disputes. Simplified Standards for Rail Rate Cases (Simplified Standards), EP 646 (Sub-No. 1) (STB served Sept. 5, 2007), aff’d sub nom. CSX Transp., Inc. v. STB, 568 F.3d 236 (D.C. Cir. 2009), vacated in part on reh’g, CSX Transp., Inc. v. STB, 584 F.3d 1076 (D.C. Cir. 2009). For medium size disputes, the Board applied a Simplified-SAC test, which relies on various simplifying assumptions, one of which is that the revenue from cross-over traffic would be apportioned based on whatever methodology is used in SAC cases. Simplified Standards, slip op. at 15.

[3] Because we do not rely on the reasoning employed in the portions of WFA’s comments that BNSF challenges, we deny its motion to strike as moot.

[4] In BNSF’s example, the on-SARR and off-SARR segments are of equal length and the average variable costs of the on-SARR and the off-SARR segments are both $50, whereas the average fixed costs for those segments are $20 and $40, respectively. Under original ATC, the on-SARR segment would be allocated $44—less than the variable costs of providing service over that segment—whereas the off-SARR segment would be allocated $56—all of the variable costs plus some contribution to joint and common costs.

[5] BNSF’s Comments 23 (quoting Simplified Standards for Rail Rate Cases, EP 646 (Sub-No. 1), slip op. at 21 (STB served July 28, 2006)).

[6] Id.

[7] Moreover, as a general purpose model built on system average figures, it is mathematically inevitable that for some movements that make up the average URCS will appear to overstate the “actual” carrier data variable costs, while for other movements it may understate those costs.

[8] See, e.g., BNSF’s Comments 25 (“It makes no sense . . . to deal with a problem that the complaining shipper can avoid in the traffic selection process.”).

[9] SAC and Simplified-SAC analyses necessarily are not identical, given that the goal in Simplified-SAC cases is to streamline the analysis. For example, we adopted a longer time period for the analysis in SAC cases (10 years) than would be used in a Simplified-SAC analysis (5 years). But the unifying principle between the differences is an attempt to simplify without losing accuracy. That principle cannot justify using original ATC in SAC cases and modified ATC in Simplified-SAC cases. The Board adopted a rule that the revenue allocation method used in Simplified-SAC cases would follow the approach used in SAC cases. Simplified Standards for Rail Rate Cases, EP 646 (Sub-No. 1), slip op. at 15 (STB served Sept. 5, 2007). In doing so, we sought to provide that the Simplified-SAC approach would automatically use the latest and most improved revenue allocation methodology.

[10] BNSF’s Comments 11-13.

[11] Id. at 19.

[12] Id. at 17.

[13] BNSF’s Comments 31-33.

[14] Id. at 30.

[15] See WFA’s Reply Comments, V.S. Crowley 46-47.

[16] Id. at 39-40.

[17] In its appeal of the original decision, BNSF contended that our modification of ATC was arbitrary and capricious “because modified ATC fails appropriately to consider economies of density and artificially inflates the revenues attributable to the SARR.” BNSF Ry. v. STB, 604 F.3d at 612. Because we failed to address this claim, the Court granted the petition for review “so that the Board on remand can address BNSF’s double-counting objection to modified ATC.” Id. at 613. We have addressed that objection in the section above.

[18] See BNSF’s Pet. for Recons. 19 (Oct. 22 2007).

[19] See BNSF’s Comments 30 (conceding that “BNSF did not propose a specific methodology” for addressing revenue allocations below variable costs). BNSF vaguely suggested in its petition for reconsideration that the Board not apply modified ATC to “high-rated traffic.” BNSF’s Pet. for Recons. 19. This passing comment was insufficient to alert the Board to BNSF’s alternate ATC methodology.