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.