CSX Transportation, Inc. v. Surface Transportation Board ( 2014 )

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.