Union Pacific Railroad v. Surface Transportation Board

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued October 21, 2010 Decided December 28, 2010

No. 10-1019

UNION PACIFIC RAILROAD COMPANY,
PETITIONER

v.

SURFACE TRANSPORTATION BOARD AND UNITED STATES OF
AMERICA,
RESPONDENTS

US MAGNESIUM, LLC,
INTERVENOR

On Petition for Review of an Order
of the Surface Transportation Board

Jonathan Marcus argued the cause for petitioner. With him
on the briefs were J. Michael Hemmer, Louise A. Rinn, and
Michael L. Rosenthal.

James A. Read, Attorney, Surface Transportation Board,
argued the cause for respondent. With him on the brief were
Robert B. Nicholson and John P. Fonte, Attorneys, U.S.
Department of Justice, Craig M. Keats, Deputy General
Counsel, Surface Transportation Board, and J. Frederick Miller,
 2

Jr., Attorney. Ellen D. Hanson, General Counsel, entered an
appearance.

Thomas W. Wilcox and David K. Monroe were on the brief
for intervenor US Magnesium, L.L.C. in support of respondent.

Before: GRIFFITH and KAVANAUGH, Circuit Judges, and
EDWARDS, Senior Circuit Judge.

Opinion for the Court filed by Senior Circuit Judge
EDWARDS.
EDWARDS, Senior Circuit Judge: US Magnesium, L.L.C.
(“USM”) operates a magnesium production facility in Rowley,
Utah, and relies on the Union Pacific Railroad Company (“UP”)
to ship its chlorine co-product from Rowley to two receiving
facilities in Arizona. In May 2009, USM filed a complaint
before the Surface Transportation Board (“STB” or the
“Board”), arguing that UP’s rates for the two chlorine
shipments, also called “movements,” were unreasonably high.
USM opted to bring its challenge under the Three Benchmark
framework set forth in Simplified Standards for Rail Rate Cases,
STB Ex Parte No. 646 (Sub-No. 1), 2007 WL 2493509 (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 other grounds
on reh’g, 584 F.3d 1076 (D.C. Cir. 2009). Under this
framework, both the shipper and the rail carrier submit groups
of comparison rates (or “offers”) for the movements in question,
from which the Board makes an “either/or” choice, with no
modifications allowed once the final submissions have been
made. In January 2010, after oral arguments and the submission
of opening evidence, reply evidence, and rebuttal evidence from
both parties, the Board selected USM’s comparison rate groups.
Using the USM groups, the Board found UP’s rates for the two
chlorine movements to be unreasonable. US Magnesium, L.L.C.
v. Union Pac. R.R., STB Docket No. 42114, 2010 WL 319727
 3

(STB served Jan. 28, 2010) (“STB Decision”), reprinted in Joint
Appendix (“J.A.”) 823-47. UP now petitions for review of the
Board’s decision, arguing that the Board’s selection of USM’s
comparison rate groups was arbitrary and capricious and seeking
vacatur of the judgment and remand for a re-application of the
Three Benchmark framework using UP’s proffered comparison
rate group.
We find no grounds to reverse the Board’s decision,
particularly given the deference owed to the Board in rate-
making disputes. Under the Three Benchmark framework, the
Board was obliged to choose between two positions, neither of
which was ideal. USM’s comparison rate groups were
problematic, because they were primarily comprised of
anhydrous ammonia shipments. UP’s comparison rate group
was flawed, because it included an unduly heavy sample of
rebilled traffic. However, based on the quantitative data and the
residual differences between the comparison rate groups, the
Board determined that USM’s submission was more
representative of the ideal comparison group – namely, single-
line chlorine traffic. Because the Board’s decision “articulated
a rational connection between the facts found and the decision
made,” N. Am. Freight Car Ass’n v. STB, 529 F.3d 1166, 1170-
71 (D.C. Cir. 2008) (quoting PPL Mont., LLC v. STB, 437 F.3d
1240, 1245 (D.C. Cir. 2006) (internal quotation marks omitted)),
we deny the petition for review.
I. BACKGROUND
A. The Three Benchmark Framework for Rate
Reasonableness
By statute, rail carriers are authorized to engage in a certain
amount of demand-based differential pricing in order to earn
“adequate revenues,” 49 U.S.C. § 10701(d)(2), whereby
“captive” shippers who are more dependent on rail service due
to the lack of transportation alternatives may be charged a higher
 4

markup as compared to “competitive” shippers who are free to
pursue lower-cost transportation options. See BNSF Ry. v. STB,
526 F.3d 770, 776 (D.C. Cir. 2008) (noting that rail carriers
would lose revenue if they imposed a strict pro rata share of the
joint and common costs to each shipper regardless of the
shipper’s degree of captivity). However, in order to protect
captive shippers from shouldering an unconscionable share of
carriers’ general operations costs, Congress has provided that,
in situations when “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.” 49 U.S.C. § 10701(d)(1). In reviewing rate
challenges, the Board must therefore consider whether carriers
are maximizing revenue from competitive shippers, id.
§ 10701(d)(2)(B), and whether “one commodity is paying an
unreasonable share of the carrier’s overall revenues,” id.
§ 10701(d)(2)(C).
Mindful of the high litigation costs associated with full
stand-alone cost (“SAC”) presentations in cases involving rate
challenges, Congress instructed 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.” Id. § 10701(d)(3); see also Simplified Standards at
5 (commenting that recent full-SAC presentations had incurred
nearly $5 million in litigation expenses and that even simplified-
SAC presentations could run to a cost of $1 million). Pursuant
to this congressional mandate, the Board adopted the Three
Benchmark framework, a relatively straightforward, “final
offer” methodology that measures the reasonableness of a rate
by comparing the revenue-over-variable cost (“R/VC”) ratio of
the challenged rate to the R/VC ratios of a comparison group of
similar movements. The challenged rate is held presumptively
unreasonable if it has a higher R/VC ratio than the calculated
upper boundary of the comparison group. See Simplified
 5

Standards at 21-22 & n.30 (explaining that the upper boundary
is the 90 percent confidence interval from a one-sided
hypothesis test constructed around the mean of the comparison
group).
The Three Benchmark framework allows both the rail
carrier and the challenging shipper to submit a comparison rate
group, with the only requirement being that all movements
included must be “captive movements.” Id. at 17 (explaining
that a captive movement is defined as a movement with a R/VC
ratio greater than 180 percent, meaning that the revenue
produced from shipping the movement was enough to cover 100
percent of the variable costs and still left a surplus in the amount
of 80 percent of the variable costs to be applied towards the rail
carrier’s joint and common costs). The movements are drawn
from an unmasked sample of a carrier’s Waybill data – a
comprehensive database that includes detailed information on all
of the movements shipped by a rail carrier, including origin and
destination points, type of commodity shipped, tonnage shipped,
and revenue. There are multiple rounds of revisions, during
which each party is given an opportunity to adjust its proposed
comparison rate group. After the parties submit their final
offers, the Board then selects the comparison rate group that is
“most similar in the aggregate to the [challenged] movement[].”
Id. at 18.
In making its selection, the Board considers “a variety of
factors, such as length of movement, commodity type, traffic
densities of the likely routes involved, and demand elasticity.”
Id. at 17. While “movements with different cost characteristics
may be included in the comparison group,” the Board “will
favor a comparison group that consists of movements of like
commodities.” Id. To encourage both parties to offer
competitive and reasonable comparison groups, the Board’s
selection is an “either/or” choice between the parties’ final
offers, with no modifications allowed. Id. at 18.
 6

After a comparison rate group is selected, the R/VC ratio of
each movement in that chosen group is then multiplied by the
carrier’s “revenue need adjustment factor,” a figure that serves
to “reflect demand-based differential pricing principles.” STB
Decision at 12-13, J.A. 834-35. If the challenged rate’s R/VC
ratio is greater than the upper boundary of the adjusted
comparison group, the rate will be presumed unreasonable.
Since the revenue need adjustment factor is derived from static
figures published annually by the Board, the Three Benchmark
framework’s reasonableness determination generally turns on
the Board’s selection of a comparison group. The shipper
bringing the rate challenge will naturally have an economic
incentive to pick movements with low R/VC ratios while the rail
carrier defending the reasonableness of the rate will have an
equal incentive to pick comparison movements with high R/VC
ratios. Although the Board recognizes that the Three
Benchmark framework has some inherent limitations, neither
party challenges the validity of the framework itself in this case.
B. Factual History and Proceedings Before the Board
USM operates the nation’s only primary magnesium
production facility in Rowley, Utah, and generates chlorine as
a co-product of the magnesium production process. Due to
chlorine’s status as a toxic inhalation hazard (“TIH”), USM
relies on UP to transport the chlorine as liquefied compressed
gas by rail service to two receiving facilities in Eloy and
Sahuarita, Arizona. In March 2009, USM began shipping its
chlorine under UP’s common carrier tariff rates, having been
unable to secure an extension of the previous contract rates. UP
charged USM $13,396 per carload for shipments from Rowley
to Eloy, and $10,410 per carload for shipments from Rowley to
Sahuarita, a marked increase from the previous years’
contractually negotiated rates. Based on variable costs of
$2,549 and $2,485, respectively, the resulting R/VC ratios were
526 percent for Eloy and 419 percent for Sahuarita. STB
 7

Decision at 3, J.A. 825. USM filed a rate challenge with the
Board in May 2009, opting to use the Three Benchmark
framework established in Simplified Standards. Since UP
acknowledged its market dominance over the chlorine
movements at issue, the Board immediately moved to the
comparison group selection process.
The Eloy and Sahuarita movements at issue are chlorine
shipments traveling in privately owned tanker cars with
capacities of less than 22,000 gallons for 1,290 and 1,250 load
miles, respectively, on single-line car service. USM’s opening
evidence proposed two separate comparison groups, one for the
Eloy movement and one for the Sahuarita movement, including
movements of all TIH commodities but limiting the groups to
movements on single-line service within a range of plus or
minus 200 miles of the actual rail distances. STB Decision at 6,
J.A. 828. UP submitted a single 24-movement comparison
group for both the Eloy and Sahuarita movements, including
both single-line and rebilled traffic and increasing the mileage
band to 400 miles but limiting the movements to those of
chlorine-only traffic and those traveling in cars that are smaller
than 22,000 gallons in capacity. Id. at 6-7, J.A. 828-29.
Both USM and UP challenged the other’s proposed groups.
Among other things, UP argued that USM did not select “like
commodities” because its groups were comprised, in large part,
of anhydrous ammonia movements and wrongly included
movements traveling in cars larger than 22,000 gallons. USM
countered that UP’s group had too few movements, wrongly
included rebilled traffic, and was over-inclusive due to the larger
400-mile range. The Board agreed that the comparison rate
groups submitted by both parties were flawed.
All else being equal, local single-line chlorine movements
would be the preferable comparison group for the issue
movements. However, neither party presented that
comparison group. Rather, both parties chose relatively
 8

extreme comparison groups in their initial tenders. And
while the deficiencies in each of the comparison groups
were evident at opening, and either party could have
strengthened its selection by adopting movements from the
other’s group(s), both parties stood rigidly behind their
initial, deficient comparison group selections.
Id. at 9, J.A. 831.
The Board’s analysis focused on the two most “pivotal”
differences: USM’s inclusion of non-chlorine movements and
UP’s inclusion of rebilled traffic. Id. at 7-9, J.A. 829-31. It
acknowledged UP’s argument that anhydrous ammonia had
significantly different demand characteristics and lower
transportation risks as compared to chlorine, thus distorting the
comparability of USM’s proposed groups. Id. at 7 & nn.9-10.
However, it found UP’s inclusion of rebilled traffic similarly
problematic, noting that the R/VC ratios for the rebilled
movements in the UP comparison group were much higher than
those of the single-line movements. Id. at 8, J.A. 830. Faced
with two imperfect offers, the Board attempted to compare the
distortions. The Board found that single-line anhydrous
ammonia had an average R/VC ratio 19 percent less than single-
line chlorine and that rebilled chlorine had an average R/VC
ratio 31 percent greater than single-line chlorine. Id. at 9-10,
J.A. 831-32. When these distortions were weighted based on
their proportional representation in the parties’ comparison
groups, USM’s Eloy and Sahuarita comparison groups had a
distortion effect of 16 percent and 17 percent, respectively,
while UP’s comparison group had a distortion effect of 18
percent. Id. The Board conceded that this difference in
weighted impacts “may not be ‘statistically significant,’” but
concluded that the analysis “confirms what USM has advocated,
which is [that] significant bias [was] introduced when [UP]
included rebilled traffic in its comparison group.” Id. at 10, J.A.
832.
 9

Having decided that single-line, chlorine traffic would be
the “preferable comparison group,” id. at 9, J.A. 831, the Board
noted that the R/VC ratios of USM’s comparison groups were
“nearly identical to the markups of the single-line, chlorine
traffic in UP’s group.” Id. at 11, J.A. 833 (finding adjusted
R/VC ratio for USM’s groups to be 304 percent and 298 percent
as compared to 301 percent for UP’s single-line chlorine
movements while UP’s rebilled movements had an R/VC ratio
of 475 percent). The Board referenced the residual differences
between the groups, finding that USM’s groups were
strengthened by a “larger sample size and narrower mileage
bands, which are more reasonable with respect to the length of
haul than UP’s mileage bands because they are closer to the two
issue movements’ actual miles.” Id. at 10, J.A. 832.
The Board then offered this telling explanation:
In sum, based on our quantitative analysis and the
residual differences between the groups, we find that USM’s
comparison groups provide the best evidence of a reasonable
level of contribution to joint and common costs for the issue
movements. . . .
[B]ased on this record, USM’s comparison groups
provide a better gauge of the proper comparison group. The
traffic at issue constitutes single-line, chlorine movements.
Yet UP submitted a comparison group containing 42%
single-line traffic and 58% rebilled (i.e., interchanged)
traffic. Single-line movements are different from rebilled
movements. Single-line movements travel over a single
carrier, and in this case an average distance of just over
1,000 miles. Rebilled traffic, on the other hand, is carried by
multiple carriers with multiple interchanges and handoffs,
and the total distance traveled by the commodity typically
will be much longer. As such, the costs of transporting the
commodity are different. Further, the markup over variable
cost UP will charge for its portion of the total trip will not
 10

only be a function of the shipper’s elasticity of demand for
transportation, but also a function of UP’s relative
bargaining power compared with other carriers involved in
the movement.
When we scrutinize UP’s comparison group, the
difference in the average adjusted R/VC ratios between
single-line and rebilled movements is significant: 301% for
single-line chlorine traffic in UP’s comparison group
compared to 475% for rebilled traffic in UP’s comparison
group. UP was unable to justify allowing rebilled traffic to
skew its comparison group. Moreover, as USM stressed at
our oral argument, its comparison groups actually provide
R/VC levels nearly identical to the markups of the single-
line, chlorine traffic in UP’s group. The adjusted R/VC
ratios submitted by USM for the Eloy and Sahuarita
movements were 304% and 298%, compared to 301% for
single-line, chlorine movements in UP’s group. . . .
To address differences in both demand elasticities and
transportation risks, nothing prevented UP from providing
a comparison group containing just single-line, chlorine
movements, and excluding rebilled movements. We
appreciate why it chose not to: it would have won the battle
(the selection of the comparison group) and lost the war (the
result would be the same as in this case). If UP had
provided a comparison group of single-line, chlorine
movements, the issue of how our costing model treats
movements of chlorine would be largely irrelevant, because
this is a comparison approach, and the issue movements and
comparison movements would be treated similarly by our
costing model. See Simplified Standards, at 17, 84.
Id. at 11, J.A. 833 (emphases in original). The Board ultimately
chose USM’s comparison rate groups, stating that “USM’s
submission (notwithstanding the relative lack of chlorine traffic)
 11

provides more reasonable comparison groups than one so
sharply skewed by rebilled traffic.” Id. at 12, J.A. 834.
Using USM’s groups for the remainder of the Three
Benchmark analysis, the Board found the maximum reasonable
R/VC ratio to be 356 percent for the Eloy movement and 346
percent for the Sahuarita movement. Id. at 13-19, J.A. 835-41.
As both challenged rates had R/VC ratios that exceeded these
upper boundaries, the Board ordered UP to establish and
maintain new rates within the allowable range and to reimburse
USM for amounts previously collected above the prescribed
levels. Id. at 19, J.A. 841.
Commissioner Nottingham dissented from the decision,
focusing on the lack of chlorine movements in USM’s
comparison groups and questioning the majority’s reliance on
rebilled traffic’s distortive effect without a clear “understanding
of the reasons for the apparent difference between the R/VC
ratios of single-line vs. rebilled chlorine movements,”
particularly when USM had the burden as the challenging party
of showing unreasonableness. Id. at 24-25, J.A. 846-47. The
dissent criticized the majority’s use of a narrow and self-created
quantitative analysis to “break the tie.” Id. at 24, J.A. 846. UP
timely petitioned for review of the Board’s decision, and USM
intervened.
II. ANALYSIS
A. Standard of Review
This court reviews decisions of the Board with deference.
A Board decision will not be reversed unless it is “unsupported
by substantial evidence,” 5 U.S.C. § 706(2)(E), “arbitrary,
capricious, an abuse of discretion, or otherwise not in
accordance with law,” AEP Tex. N. Co. v. STB, 609 F.3d 432,
438 (D.C. Cir. 2010); 5 U.S.C. § 706(2)(A). Deference is
particularly high in rate disputes, where “the Board acts at the
zenith of its powers.” Id. (quoting Burlington N. R.R. v. STB,
 12

114 F.3d 206, 210 (D.C. Cir. 1997) (internal quotation marks
omitted). “Where the Board’s findings rest on such relevant
evidence as a reasonable mind might accept as adequate to
support a conclusion, and where the Board has articulated a
rational connection between the facts found and the decision
made, [this court] will not disturb its judgment.” N. Am. Freight
Car Ass’n, 529 F.3d at 1170-71 (quoting PPL Mont., 437 F.3d
at 1245 (internal quotation marks omitted)).
B. UP’s Challenge Fails
UP’s challenge to the Board’s decision is premised on its
belief that its proposed comparison rate group, even with its
heavy inclusion of rebilled traffic, was not skewed. UP
maintains that the Board’s decision is arbitrary and capricious,
because it considers rebilled traffic on par with non-chlorine
traffic without explaining why rebilled traffic would have a
distortive effect. UP’s argument is unpersuasive, for it is
grounded on a serious misunderstanding of the Three
Benchmark framework.
The Three Benchmark, final-offer process does not
facilitate a “search for the truth.” The Board’s final judgment in
any case involving this final-offer process necessarily will be
constrained by what the parties present as their final offers. The
Board merely selects between two final offers, neither of which
may be ideal. There is no room for compromise, and no
procedure to adjust two flawed final offers.
It is true that USM, as the party bringing the rate challenge,
had the ultimate burden of persuasion to show that UP’s rates
were unreasonable. See BNSF Ry. v. STB, 453 F.3d 473, 485
(D.C. Cir. 2006). However, in the comparison group selection
stage of the Three Benchmark analysis, USM and UP each had
an obligation to defend its own proposal against the opposing
side’s attacks. See id. (citing “Complex” Consol. Edison Co. v.
FERC, 165 F.3d 992, 1008 (D.C. Cir. 1999)) (explaining that the
 13

shipper bringing the challenge who holds the ultimate burden of
persuasion does not necessarily have the burden of production
“[s]o long as the record supports [the Board’s ultimate]
conclusion”). There is nothing in this record to indicate that
UP’s proposed comparison group, and the resulting maximum
rate level, would have been judged objectively reasonable under
a full-SAC rate challenge. Indeed, even UP does not make this
claim. Thus, the question before the Board was which of the
proposed comparison rate groups was least undesirable. This is
the nature of the beast in the “final offer” process upon which
the Three Benchmark framework is founded. UP’s protests
ignore this reality.
UP challenged the comparability of anhydrous ammonia to
chlorine traffic, arguing that USM’s heavy inclusion of
anhydrous ammonia skewed the demand characteristics of its
comparison rate group. USM tried to defend its selections by
referencing the Board’s stated preference for comparison groups
with a mix of TIH commodities in E.I. DuPont de Nemours and
Co. v. CSX Transportation, Inc., STB Docket No. 42100, 2008
WL 2588609 (STB served June 30, 2008), the only previous
application of the Three Benchmark framework. STB Decision
at 7-8, J.A. 829-30. However, the Board pointed out that
DuPont is easily distinguishable. Id. (explaining that DuPont’s
carrier had confessed to overt chlorine pricing manipulation).
USM, in turn, challenged the comparability of rebilled traffic to
single-line traffic, arguing that UP’s heavy inclusion of rebilled
traffic skewed the R/VC ratios of its own comparison group.
Like USM, UP then had the opportunity to explain or refute the
alleged distortion. Like USM, UP failed to make a convincing
defense.
UP acknowledged that rebilled traffic has lower variable
costs as compared to single-line traffic. While a difference in
variable costs, by itself, does not affect the comparability of two
movements, the presumption that comparability is not affected
 14

depends on the existence of similar R/VC ratios. See Simplified
Standards at 17 (noting that movements with different costs are
acceptable because “there is no reason, a priori, to presume that
the R/VC ratios (or their share of joint and common costs)
should be different”). Here, there was undisputed evidence that
the rebilled traffic in UP’s sample had much higher R/VC ratios
as compared to single-line traffic. The difference in variable
costs was, therefore, not matched by similarly lowered prices,
but rather by a greater profit margin, suggesting differences in
demand elasticity under a differential pricing system.
Building upon USM’s allegation that rebilled movements
had much higher R/VC ratios, the Board compared the relative
distortive impact of anhydrous ammonia and rebilled
movements, and found that rebilled chlorine movements had an
average R/VC ratio 31 percent higher than single-line chlorine
movements, but that anhydrous ammonia movements had an
average R/VC ratio only 19 percent lower than chlorine
movements. STB Decision at 9-10, J.A. 831-32. Though it
acknowledged that the weighted impact between the two
distortions might “not be ‘statistically significant,’” the analysis
“confirm[ed] what USM ha[d] advocated, which is [that]
significant bias [was] introduced when [UP] included rebilled
traffic in its comparison group.” Id. at 10, J.A. 832.
At oral argument before the Board, counsel for UP was
specifically asked to explain the difference caused by the
inclusion of rebilled traffic, see Oral Arg. Tr. 28, Nov. 23, 2009,
J.A. 743 (“How do you respond to [USM’s] argument regarding
the re-bill issue in his exhibit?”). UP offered no explanation,
either in its rebuttal evidence or in its response to the Board’s
questioning. UP’s only response was to maintain that even if the
use of rebilled traffic was distortive, USM’s inclusion of non-
chlorine traffic was even more damaging:
[If] it’s a choice between re-billed movements with some
cost characteristics that might be slightly different and a
 15

comparison group that’s comprised 99 percent or 96 percent
of non-chlorine traffic, where should you be more concerned
that you’re not reflecting the demand characteristics? I
submit it’s the comparison group that has hardly any
chlorine, not the fact that you’ve got some re-billed
movements in there.
Id. at 35, J.A. 750. Since “UP was unable to justify allowing
rebilled traffic to skew its comparison group,” STB Decision at
11, J.A. 833, the Board reasonably concluded based on the
evidence presented that rebilled traffic, like anhydrous
ammonia, had a comparably improper distortive effect.
Since both UP and USM’s groups were flawed, the Board
proceeded to compare both groups to the ideal comparison
group – namely, single-line chlorine movements. The Board
explained why USM came out ahead in the final-offer process:
In the end, it is our responsibility to select the better
comparison group between those submitted by the parties in
order to judge the reasonableness of the challenged
rates. . . . Here, we agree with UP and our dissenting
colleague that there are important differences between
movements of chlorine and anhydrous ammonia, as
illustrated by UP’s evidence of the differences between the
two groups. While that difference, as measured by the R/VC
ratios in the broader Waybill sample data, is significant
(19%), a more pronounced difference is present between
single-line and rebilled, chlorine traffic (31%). These
simplified proceedings do not offer us a proper platform to
explore why either discrepancy exists. But all of the
quantitative evidence points toward accepting USM’s groups
over that proffered by UP. Given that: (1) UP’s own
comparison group shows average R/VC levels for single-
line, chlorine movements that are practically identical to
those from USM’s groups; (2) the supplemental quantitative
analysis we undertook of the relevant data confirms that the
 16

distortion from rebilled movements appears more
significant; and (3) the residual differences between the
parties’ groups – such as the number of observations and the
mileage bands – also favor USM’s groups, it is our judgment
that USM’s submission (notwithstanding the relative lack of
chlorine traffic) provides more reasonable comparison
groups than one so sharply skewed by rebilled traffic.
Id. at 12, J.A. 834.
It is important to say again that the Board’s judgment need
not reflect mathematical certainty in order to survive judicial
scrutiny. Indeed, there is good reason to believe that judgments
rendered pursuant to the Three Benchmark framework more
often than not will be the antithesis of mathematical certainty.
The final-offer process does not offer or promise certainty or
precision in ratemaking. It would be absurd for anyone to think
otherwise. Therefore, because, as we have already noted, “the
Board acts at the zenith of its powers when it sets rail rates, we
recognize that the Board is entitled to particular deference. As
long as the Board has rationally set forth the grounds on which
it acted, this court may not substitute its judgment for that of the
agency.” AEP Tex. N. Co., 609 F.3d at 438 (citations, ellipsis,
and internal quotation marks omitted). The Board’s decision
easily meets this standard.
UP disregards the reality of the situation here and criticizes
the Board’s emphasis on single-line chlorine traffic as the ideal
comparison group. UP argues that the Board should have
considered whether rebilled movements may actually be more
comparable. This argument is illogical: why would the Board
have found rebilled traffic more comparable than single-line
traffic when the two movements in question were single-line
movements? UP also criticizes the Board’s emphasis on its
“supplemental quantitative analysis,” arguing that a test
resulting in statistically insignificant weighted differences
cannot be dispositive. However, the Board did not say that the
 17

quantitative analysis was dispositive. The Board’s comparison
of R/VC distortion was merely one portion of its overall
analysis. Finally, UP criticizes the Board’s reliance on the
residual differences, but it cannot dispute the fact that USM’s
mileage bands were more representative of the two movements
at issue. Where, as here, the Board’s selection was based on a
rational application of evidence presented, this court has no
power to “substitute its [own] judgment for that of the agency.”
Id. (quoting BNSF Ry., 453 F.3d at 480).
The Three Benchmark framework was created to give
shippers a less costly vehicle through which to bring their rate
disputes. The process was designed to give parties an incentive
to advance proposals that would minimize their differences, in
search of reasonable accommodations. Unfortunately, both
USM and UP chose to take extreme positions in their
comparison rate group offers. The final-offer process thus failed
to work as intended in this case.
The Board has explained how the process should work:
Any final tender that is skewed too far in one direction
might well result in the selection of a more reasonable final
tender presented by the opposing party. By having two
rounds of simultaneous tenders and a technical conference,
both sides will participate in the winnowing
process. . . . This approach will work as intended only if the
parties know that the agency will not attempt to find a
compromise position somewhere in the middle. To create
the proper incentives for the litigants not to take extreme
positions, we commit to selecting the more reasonable of the
two groups as tendered.
Simplified Standards at 18. Here, the anticipated “winnowing
process” did not occur. Both sides stood firmly behind their
initial, clearly skewed offers, apparently hoping that the other
side’s proposal would be rejected as even more extreme.
 18

Indeed, “[t]he parties seem[ed] to be intent on testing the outer
boundaries of what might qualify as an acceptable comparison
group, rather than adhering to the spirit of the Three-Benchmark
process and seeking middle ground in an effort at reaching a
reasonable and expedited outcome.” STB Decision at 22
(Commissioner Nottingham, dissenting), J.A. 844. UP lost this
high-stakes gamble and now protests that the result is unjust.
But UP cannot blame the Board for this result. The Board’s
decision, being rationally based on the facts presented, was not
in error.
III. CONCLUSION
For the foregoing reasons, we deny the petition for review.