Owner-Operator Independent Drivers Ass'n v. Federal Motor Carrier Safety Administration

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued December 4, 2006 Decided July 24, 2007

No. 06-1035

OWNER-OPERATOR INDEPENDENT DRIVERS ASSOCIATION,
INC.,
PETITIONER

v.

FEDERAL MOTOR CARRIER SAFETY ADMINISTRATION,
RESPONDENT

INTERNATIONAL BROTHERHOOD OF TEAMSTERS, AFL-CIO, ET
AL.,
INTERVENORS

Consolidated with
06-1078

On Petitions for Review of a Final Rule of the
Federal Motor Carrier Safety Administration

Paul D. Cullen, Jr. argued the cause for petitioner
Owner-Operator Independent Drivers Association, Inc. With
him on the briefs was Paul D. Cullen, Sr. Daniel E. Cohen
entered an appearance.
 2

Bonnie I. Robin-Vergeer argued the cause for petitioners
Public Citizen, Inc., et al. With her on the briefs were Brian
Wolfman and Scott L. Nelson. James A. McCall entered an
appearance.

Kenneth E. Siegel was on the briefs for intervenors
California Trucking Association, et al.

Michele M. Fields was on the brief for amicus curiae
Insurance Institute for Highway Safety in support of petitioners.

Matthew M. Collette, Attorney, U.S. Department of Justice,
argued the cause for respondents. With him on the brief were
Peter D. Keisler, Assistant Attorney General, Robert S.
Greenspan, Attorney, and Paul M. Geier, Assistant General
Counsel, Federal Motor Carrier Safety Administration.

Robert Digges, Jr., Erika Z. Jones, Adam C. Sloane, David
M. Gossett, John M. Cutler, Jr., Nicholas J. DiMichael, and C.
Fairley Spillman were on the brief of intervenors American
Trucking Associations, Inc., et al. in support of respondent.
Karyn A. Booth entered an appearance.

Before: GINSBURG, Chief Judge, and HENDERSON and
GARLAND, Circuit Judges.

Opinion for the Court filed by Circuit Judge GARLAND.

GARLAND, Circuit Judge: In order to ensure highway safety
and protect driver health, Congress has charged the Federal
Motor Carrier Safety Administration with regulating the hours
of commercial motor vehicle operators. In 2005, the agency
promulgated a final rule revising its existing regulations in a
number of respects. Two groups -- one led by Public Citizen
and the other by the Owner-Operator Independent Drivers
 3

Association -- now seek review of the portion of the rule that
applies to long-haul truck drivers.

We reject the challenges raised by the Owner-Operators, but
grant the petition filed by Public Citizen. We conclude that the
agency violated the Administrative Procedure Act because it
failed to give interested parties an opportunity to comment on
the methodology of the crash-risk model that the agency used to
justify an increase in the maximum number of daily and weekly
hours that truck drivers may drive and work. We also find that
the agency failed to provide an explanation for critical elements
of that methodology.

I

This is the second time this court has considered a challenge
to the Federal Motor Carrier Safety Administration’s attempt to
modify its hours-of-service regulations. Much of the relevant
background is set forth in our opinion in Public Citizen v.
FMCSA, 374 F.3d 1209 (D.C. Cir. 2004), which vacated a prior
iteration of the rule now before us. We first review that
background and then describe the development of the current
rule.

A

The federal government has regulated the hours of service
(HOS) of commercial motor vehicle operators since the late
1930s, when the Interstate Commerce Commission (ICC)
promulgated the first HOS regulations under the authority of the
Motor Carrier Act of 1935. See 49 U.S.C. § 31502(b)(1)
(authorizing the prescription of “maximum hours of service” for
motor carrier employees). Jurisdiction over HOS regulations
passed from the ICC to the Federal Highway Administration
(FHWA) in 1995, and then to the newly created Federal Motor
 4

Carrier Safety Administration (FMCSA) in 2000. Along the
way, Congress added to the statutory basis for the HOS
regulations. The current rule was promulgated under the
authority of both the Motor Carrier Act of 1935 and the Motor
Carrier Safety Act of 1984, which, as amended, directs the
Secretary of Transportation to “prescribe regulations on
commercial motor vehicle safety,” 49 U.S.C. § 31136(a), and
provides that “[a]t a minimum, the regulations shall ensure”
that:

(1) commercial motor vehicles are maintained,
equipped, loaded, and operated safely; (2) the
responsibilities imposed on operators of commercial
motor vehicles do not impair their ability to operate the
vehicles safely; (3) the physical condition of operators
. . . is adequate to enable them to operate the vehicles
safely . . . ; and (4) the operation of commercial motor
vehicles does not have a deleterious effect on the
physical condition of the operators.

Id. In addition, FMCSA is required to “consider the assignment
and maintenance of safety as the highest priority,” id. § 113(b),
and to consider the “costs and benefits” of its safety regulations,
id. § 31502(d); see id. § 31136(c)(2)(A).

Between 1940 and 2003, the HOS regulations applicable to
long-haul truck drivers1 remained largely unchanged. Five
aspects of the pre-2003 regulations are relevant to the petitions
before us:

1
FMCSA distinguishes long-haul truck drivers from short-haul
drivers. The latter generally “return to their work-reporting location
every night” and operate “within a 150 air-mile radius from their
terminals.” Hours of Service of Drivers, 70 Fed. Reg. 49,978, 50,031
(Aug. 25, 2005).
 5

• The daily driving limit. Drivers were not allowed to
drive for more than a total of 10 hours without taking
a required off-duty period. 49 C.F.R. §395.3(a)(1)
(2002) (superseded).

• The daily on-duty limit. Even if they had not reached
the 10-hour driving limit, drivers could not drive after
they had been on duty for 15 hours. Id. § 395.3(a)(2).
Drivers could, however, “take periodic ‘off-duty’
breaks during the day, thus extending the fifteen-hour
driving-eligible ‘on duty’ period beyond fifteen hours.”
Public Citizen, 374 F.3d at 1212.

• The daily off-duty requirement. In order to restart the
10-hour driving limit and the 15-hour on-duty limit,
drivers were required to take at least 8 consecutive
hours off duty. 49 C.F.R. § 395.3(a) (2002)
(superseded).

• The sleeper-berth exception. The regulations
contained an exception to the 8-hour off-duty
requirement for drivers who took their off-duty time in
a “sleeper berth,” a compartment in the cabin of a truck
with space for a driver to rest. Drivers could
accumulate their required 8 hours of off-duty time in
two separate periods in a sleeper berth as long as each
was at least 2 hours long. Id. § 395.1(g).

• The weekly on-duty limit. Drivers were not allowed to
drive after having been on duty for 60 hours in the past
7 days. Id. § 395.3(b).2

2
Drivers employed by carriers that operated every day of the
week were subject to a slightly different limit, which barred them from
driving after having been on duty for 70 hours in the past 8 days. See
 6

All of these requirements were “limits on the time drivers could
work and still drive; so far as the rules went, drivers who
worked more than the daily or weekly limits could still work as
long as they did not drive.” Public Citizen, 374 F.3d at 1212.

In the ICC Termination Act of 1995, Congress directed the
FHWA to revise these regulations by conducting a rulemaking
“dealing with a variety of fatigue-related issues pertaining to
commercial motor vehicle . . . safety.” 49 U.S.C. § 31136 note.
Congress specifically instructed the agency to address the
following issues:

8 hours of continuous sleep after 10 hours of driving,
loading and unloading operations, automated and
tamper-proof recording devices, rest and recovery
cycles, fatigue and stress in longer combination
vehicles, fitness for duty, and other appropriate
regulatory and enforcement countermeasures for
reducing fatigue-related incidents and increasing driver
alertness.

Id. FHWA issued an advance notice of proposed rulemaking
(ANPRM) in 1996. See 61 Fed. Reg. 57,252 (Nov. 5, 1996).
Jurisdiction over the HOS regulations then passed to FMCSA,
which issued a notice of proposed rulemaking (NPRM) in 2000.
See 65 Fed. Reg. 25,540 (May 2, 2000) (“2000 NPRM”). The
2000 NPRM explained that “[t]here is general consensus that
modifications to current HOS regulations would substantially
improve [commercial motor vehicle (CMV)] safety by reducing
the fatigue factor in CMV-involved crashes.” Id. at 25,540.

49 C.F.R. § 395.3(b) (2002) (superseded). All of the rules and
proposed rules discussed in this Opinion contain(ed) the same set of
7- and 8-day limits. For simplicity, we hereinafter refer only to the 7-
day limit.
 7

In the 2000 NPRM, the agency made a number of specific
findings that identified shortcomings in the pre-2003 HOS
regulations. First, FMCSA found that “people are much more
alert and have better reaction times when they are on regular,
twenty-four-hour circadian schedules.” Public Citizen, 374 F.3d
at 1213; see also 2000 NPRM, 65 Fed. Reg. at 25,554. The pre-
2003 regulations allowed drivers to alternate between 10-hour
driving shifts and 8-hour off-duty periods, creating the potential
for a “backward rotating” schedule in which they started each
“day” six hours earlier than the one before. Second, the agency
concluded that “[e]ach driver should have an opportunity for
eight consecutive hours of uninterrupted sleep every day.” 2000
NPRM, 65 Fed. Reg. at 25,554. In the agency’s view, the old
regulations’ 8-hour off-duty period was too short to allow
drivers to eat, commute, and conduct other personal activities
while still getting enough sleep. See id. Third, FMCSA
concluded that drivers need weekly recovery periods “to negate
the effect of accumulated week-long sleep deprivation and
restore alertness.” Id. at 25,555. The old regulations contained
no weekly off-duty requirement. Finally, FMCSA found that
“performance degrades and crash risk increases markedly after
the 12th hour of any duty time during a work shift.” Id. at
25,556. The pre-2003 regulations permitted only 10 hours of
driving per shift, but allowed an operator to drive anytime up to
his or her 15th hour on duty, and to extend the 15-hour window
by taking off-duty breaks.

Based on these findings, the 2000 NPRM proposed
significant changes to the existing HOS regulations. Instead of
a 10-hour daily driving limit and a 15-hour daily on-duty limit,
the 2000 NPRM proposed allowing up to 12 hours of working
or driving time, and requiring an additional 2 hours of off-duty
time at some point during each shift. Id. at 25,568 tbl.5, 25,581.
The 2000 NPRM also proposed increasing the 8-hour daily off-
duty requirement to 10 consecutive hours and mandating a new
 8

weekly off-duty requirement of between 32 and 56 consecutive
hours, with the required length depending on the amount of time
needed to “include two sleep periods between midnight and 6:00
a.m.” Id. Finally, the NPRM proposed eliminating the sleeper-
berth exception for solo drivers. Team drivers -- who work in
two-person teams and often use the sleeper berth to allow one
driver to obtain the required off-duty time while the other
continues to drive -- could still use the sleeper-berth exception,
but each period of time in the sleeper berth would have to be at
least 5 hours (up from 2 hours under the old regulations). See
id. at 25,586-87.

FMCSA issued a final rule in April 2003. See Hours of
Service of Drivers, 68 Fed. Reg. 22,456 (Apr. 28, 2003) (“2003
Rule”). As we explained in Public Citizen, the 2003 Rule “was
still a significant revision to the old [pre-2003] rules, but
differed markedly from the [2000 NPRM].” 374 F.3d at 1214.
The Rule modified three of the five aspects of the pre-2003
regulations discussed above, and added a sixth:

• The daily driving limit. The 2003 Rule increased (over
the pre-2003 regulations) the daily driving limit from
10 hours to 11 hours. See 2003 Rule, 68 Fed. Reg. at
22,457.

• The daily on-duty limit. The 2003 Rule reduced the
daily on-duty limit from 15 to 14 hours and prohibited
drivers from extending that limit by taking off-duty
breaks during their shifts. See id.

• The daily off-duty requirement. Like the 2000 NPRM,
the 2003 Rule increased the daily off-duty requirement
from 8 hours to 10 hours. See id.
 9

• The sleeper-berth exception. The 2003 Rule
abandoned the 2000 NPRM’s proposed changes to the
sleeper-berth exception. Instead, it preserved the same
kind of exception that existed under the pre-2003
regulations: a driver using a sleeper berth could satisfy
the off-duty requirement in two separate periods as
long as each of them was at least 2 hours long. See id.
at 22,501.

• The weekly on-duty limit. The 2003 Rule preserved the
60-hour weekly on-duty limit, but created a new
exception to this requirement, the “34-hour restart
provision.” See id. at 22,457.

• The 34-hour restart provision. The 2003 Rule allowed
drivers to restart their weekly on-duty clocks whenever
they took 34 consecutive hours off duty. See id. This
was a significant shift from the 2000 NPRM: instead of
requiring a 32- to 56-hour off-duty period each week
(a change that would have decreased average hours
worked), the 2003 Rule permitted drivers to restart
their 60-hour limit anytime they took 34 hours off
duty. This change allowed drivers to work
substantially longer hours per week -- as many as 17
more hours over 7 days. See Public Citizen, 374 F.3d
at 1215.

A group of petitioners led by Public Citizen challenged the
2003 Rule on a variety of grounds. We agreed with them that
“the rule [was] arbitrary and capricious because the agency
failed to consider the impact of the rule[] on the health of
drivers, a factor the agency must consider under its organic
statute.” Id. at 1216; see 49 U.S.C. § 31136(a)(4). “Because the
agency ha[d] wholly failed to comply with this specific statutory
requirement,” we found “this single objection from petitioners
 10

. . . sufficient to establish an arbitrary-and-capricious decision
requiring vacatur of the rule.” Public Citizen, 374 F.3d at 1216.

This disposition made it unnecessary for us to “enter final
judgment” on Public Citizen’s other objections to the 2003 Rule.
Id. We noted, however, that several of those objections “also
raise[d] troubling concerns about [FMCSA’s] decisionmaking
process,” and that the agency could consider the objections on
remand. Id. First, we expressed “very real concerns” about the
increase in the daily driving limit from 10 to 11 hours. Id. at
1217. We noted that the “agency freely concedes that ‘studies
show[] that performance begins to degrade after the 8th hour on
duty and [the degradation] increases geometrically during the
10th and 11th hours.’” Id. at 1218 (quoting 2003 Rule, 68 Fed.
Reg. at 22,471). But “[d]espite this finding, the agency cited
absolutely no studies in support of its notion that the decrease in
[the] daily driving-eligible tour of duty from fifteen to fourteen
hours will compensate for [the] conceded and documented ill
effects from the increase” in driving time. Id.

Second, we also found suspect the agency’s claim that the
increase in the daily driving limit to 11 hours could be justified
by “the cost-benefit analysis it conducted.” Id. The model
employed in that analysis, we noted, “assume[d], dubiously, that
time spent driving is equally fatiguing as time spent resting --
that is, that a driver who drives for ten hours has the same risk
of crashing as a driver who has been resting for ten hours [and]
then begins to drive.” Id. “In other words, the model
disregarded the effects of ‘time on task,’” and thus understated
the risks of driving 11 hours. Id.

Third, “[o]ur doubts extend[ed] as well to the agency’s
justification for retaining the sleeper-berth exception,” which
permitted “solo and team drivers to obtain the necessary ten
hours of off-duty time by splitting their rest in two periods of
 11

time spent in sleeper berths.” Id. at 1219. Public Citizen
“argue[d] persuasively,” we said, “that the agency’s justification
for retaining this exception was not rational in view of the
conceded central premise of the HOS regulations . . . that ‘[e]ach
driver should have an opportunity for eight consecutive hours of
uninterrupted sleep every day.’” Id. (quoting 2003 Rule, 68 Fed.
Reg. at 22,469).

Finally, we regarded as “problematic” the fact that
FMCSA’s justification for the 34-hour restart provision “[did]
not even acknowledge, much less justify, that the rule . . .
dramatically increases the maximum permissible hours drivers
may work each week.” Id. at 1222. That increase, we said, “is
likely an important aspect of the problem[,] [a]nd the agency’s
failure to address it . . . makes this aspect of the [2003] rule’s
rationality questionable.” Id. at 1222-23 (citation and internal
quotation marks omitted).

B

After our July 16, 2004 decision in Public Citizen vacated
the 2003 Rule, FMCSA sought and received temporary relief
from the vacatur in Congress. The Surface Transportation
Extension Act of 2004, signed by the President on September
30, 2004, provided that the 2003 Rule “shall be in effect until
the earlier of -- (1) the effective date of a new final rule
addressing the issues raised by [Public Citizen]; or (2)
September 30, 2005.” Pub. L. No. 108-310, § 7(f), 118 Stat.
1144, 1154 (2004).

FMCSA issued a new NPRM in January 2005. See 70 Fed.
Reg. 3339 (Jan. 24, 2005) (“2005 NPRM”). The 2005 NPRM
used the 2003 Rule as its proposal and sought “comment on
what changes to that rule, if any, [were] necessary to respond to
the concerns raised by the court” in Public Citizen. Id. at 3339.
 12

In August 2005, FMCSA promulgated the rule now under
review. See Hours of Service of Drivers, 70 Fed. Reg. 49,978
(Aug. 25, 2005) (“2005 Rule”). With a single exception, the
2005 Rule is identical to the 2003 Rule. The 2005 Rule
preserves the 11-hour daily driving limit, the 14-hour daily on-
duty limit, the 10-hour daily off-duty requirement, the 60-hour
weekly on-duty limit, and the 34-hour restart provision. See id.
at 49,980. The only difference between the two rules is the
sleeper-berth exception. The 2003 Rule -- like the pre-2003
regulations -- permitted drivers using a sleeper berth to split
their daily off-duty requirement into two periods as long as each
period was at least 2 hours long. The 2005 Rule, by contrast,
requires that one period consist of at least 8 hours in the sleeper
berth. The other period can be spent either in the sleeper berth
or elsewhere and must be at least two hours long. See id. at
50,030.

Although the 2005 Rule was largely unchanged from the
2003 Rule, FMCSA said that it had considered and addressed
the concerns identified by this court in Public Citizen. As to
driver health, the agency explained that it had conducted an
extensive literature review to determine the effect of the rule on
a variety of health issues, and concluded that the 2005 Rule
would either have no effect or yield a net improvement over the
pre-2003 regulations. See id. at 49,991-92.

FMCSA also determined that the 2005 Rule would improve
highway safety by reducing fatigue-related accidents. First, it
explained that the 2005 Rule’s change to the sleeper-berth
exception would reduce driver fatigue. It cited evidence that
“sleep accumulated in short time blocks is less refreshing than
sleep accumulated in one long time period,” and studies
indicating that drivers using the split-sleeper-berth provision of
the pre-2003 regulations were more likely to be involved in fatal
accidents. Id. at 49,994 (internal quotation marks omitted).
 13

Second, FMCSA contended that the longer hours permitted by
the 34-hour restart provision would not be detrimental, pointing
to evidence that “a recovery period of 34 consecutive hours is
sufficient for recovery from moderate cumulative fatigue” due
to several days of limited sleep. Id. at 49,995.

Finally, FMCSA relied on a new cost-benefit analysis that
it described in a Regulatory Impact Analysis (RIA) released
along with the 2005 Rule. See FMCSA, Regulatory Impact
Analysis and Small Business Impact Analysis for Hours of
Service Options (2005) (J.A. 1627) (“2005 RIA”). Based on this
analysis, FMCSA concluded that the economic costs to industry
of rescinding the two provisions of the rule that this court had
criticized in Public Citizen -- the increase (over the pre-2003
regulations) in the daily driving limit from 10 hours to 11 hours,
and the addition of the 34-hour restart provision -- outweighed
the safety benefits that rescission would bring. See 2005 Rule,
70 Fed. Reg. at 49,981. As explained below, a key component
of the cost-benefit analysis was an operator-fatigue model that
the agency used to analyze crash risks under different HOS
regimes.

Public Citizen petitioned for review of the 2005 Rule,
challenging the 11-hour daily driving limit and the 34-hour
restart provision. The Owner-Operator Independent Drivers
Association (OOIDA) also petitioned for review, challenging the
provision that renders the 14-hour daily on-duty limit
nonextendable and the provision that modifies the sleeper-berth
exception. A group led by the California Trucking Association
intervened in support of OOIDA. We consider Public Citizen’s
challenge in Part II and OOIDA’s challenge in Part III.
 14

II

We have jurisdiction to review the 2005 Rule pursuant to 28
U.S.C. § 2342(3)(A). Our standard of review is governed by the
Administrative Procedure Act (APA), 5 U.S.C. § 706. See
Advocates for Highway & Auto Safety v. FMCSA, 429 F.3d
1136, 1144-45 (D.C. Cir. 2005); Public Citizen, 374 F.3d at
1216. Under the APA, we must set the rule aside if it is
“arbitrary, capricious, an abuse of discretion, or otherwise not in
accordance with law,” 5 U.S.C. § 706(2)(A), or if it was
promulgated “without observance of procedure required by
law,” id. § 706(2)(D).

Public Citizen challenges the 2005 Rule, and specifically its
11-hour daily driving limit and 34-hour restart provision, on four
grounds. The petitioner contends that: (1) FMCSA violated the
APA’s requirements for notice-and-comment rulemaking by
failing to disclose in time for comment the methodology of a
model that was central to the agency’s justification for the rule;
(2) when the methodology finally was disclosed, FMCSA failed
to provide a reasoned explanation for some of its critical
elements, thus rendering it (and the rule) arbitrary and
capricious; (3) FMCSA’s treatment of a number of other safety
considerations was also arbitrary and capricious; and (4) the rule
is contrary to law and arbitrary and capricious because it fails to
protect driver health. As explained below, we reach only the
first two arguments.

A

Public Citizen’s first contention is that FMCSA violated the
APA’s requirements for notice-and-comment rulemaking by
failing to disclose the methodology of the agency’s operator-
fatigue model, a crash-risk analysis that was a central
component of the justification for the 2005 Rule.
 15

1. The APA requires that an agency publish notice of
proposed rulemaking, including “either the terms or substance
of the proposed rule or a description of the subjects and issues
involved,” 5 U.S.C. § 553(b)(3), and that it “give interested
persons an opportunity to participate in the rule making through
submission of written data, views, or arguments,” id. § 553(c).
As we have explained, “[i]ntegral” to these requirements “is the
agency’s duty ‘to identify and make available technical studies
and data that it has employed in reaching the decisions to
propose particular rules. . . . An agency commits serious
procedural error when it fails to reveal portions of the technical
basis for a proposed rule in time to allow for meaningful
commentary.’” Solite Corp. v. EPA, 952 F.2d 473, 484 (D.C.
Cir. 1991) (quoting Connecticut Light & Power Co. v. NRC, 673
F.2d 525, 530-31 (D.C. Cir. 1982)); see Chamber of Commerce
v. SEC, 443 F.3d 890, 899 (D.C. Cir. 2006); see also Air Transp.
Ass’n of Am. v. FAA, 169 F.3d 1, 7 (D.C. Cir. 1999) (“‘[T]he
most critical factual material that is used to support the agency’s
position on review must have been made public in the
proceeding and exposed to refutation.’” (quoting Association of
Data Processing Serv. Orgs. v. Bd. of Governors of the Fed.
Reserve Sys., 745 F.2d 677, 684 (D.C. Cir. 1984))).

2. In order to assess the costs and benefits of alternative
changes to the HOS rules, FMCSA created a carrier-operations
model for estimating the costs to industry of each option, as well
as an operator-fatigue model for calculating the crash risks
under each option. (The benefits of avoiding crashes were then
monetized and incorporated into the cost-benefit analysis.) The
agency explained the models in the RIA that it released along
with the 2005 Rule. See 2005 RIA (J.A. 1627); see also 2005
Rule, 70 Fed. Reg. at 50,045-55. FMCSA summarized its
analysis as follows:
 16

To produce a realistic measurement of the impacts of
each option, we divided the industry into broad
segments, collected information on operations within
these segments, and then created a model of carrier
operations . . . .

The model was . . . used to simulate carrier operations
under different conditions and HOS rules. . . .
[W]eighted changes in productivity from this
procedure were then used to estimate the cost increases
imposed on the industry by each option . . . .

Safety impacts were measured by feeding the on duty
and driving schedules from the carrier simulation
model into an operator fatigue model to project driver
effectiveness levels, and then using the fatigue model
results to estimate the resulting changes in crash risks
under each HOS option . . . . Changes in fatigue-related
crash risks . . . were then multiplied by the value of all
affected crashes to yield estimates of total benefits.

2005 RIA at ES-2 (J.A. 1629).

Many of the details of the models were unchanged from
those used in the cost-benefit analysis of the 2003 Rule. See
FMCSA, Regulatory Impact Analysis and Small Business
Analysis for Hours of Service Options (2002) (J.A. 787) (“2003
RIA”). In several important respects, however, FMCSA
modified the operator-fatigue model. In part, those
modifications represented an effort to respond to criticisms this
court leveled in our opinion vacating the 2003 Rule. One of
those criticisms was that the agency had “excluded time-on-task
effects from the cost-benefit analysis.” Public Citizen, 374 F.3d
at 1219. The agency’s 2003 operator-fatigue model had
estimated a driver’s crash risk solely by reference to his or her
 17

sleep patterns, and thus assumed that time spent driving was no
more fatiguing than time spent resting. Id. We found the
decision to ignore time-on-task effects puzzling, because “the
agency admits that studies show that crash risk increases, in
[FMCSA’s] words, ‘geometrically,’ after the eighth hour on
duty, and the agency does not deny that this geometric risk
increase results at least in substantial part from time-on-task
effects.” Id. (quoting 2003 Rule, 68 Fed. Reg. at 22,471).

In its 2005 analysis, FMCSA modified its 2003 operator-
fatigue model to account for time-on-task effects. To do so, the
agency commissioned a study of crash data from a national
database known as “Trucks Involved in Fatal Accidents”
(TIFA). See Kenneth L. Campbell, Estimates of the Prevalence
and Risk of Fatigue in Fatal Crashes Involving Medium/Heavy
Trucks from the 1991-2002 TIFA Files (2005) (J.A. 1712)
(“TIFA Study”). The database includes over 50,000 truck-
involved accidents over the years 1991-2002. 2005 RIA at 44
(J.A. 1664). The TIFA Study generated what FMCSA referred
to as the “fatigue-related crash risk” for each successive hour of
driving. Id. at 59 (J.A. 1679). For each driving hour, the study
calculated the percentage of all fatal truck crashes in which it
was determined that the driver was fatigued at the time of the
crash. This calculation yielded the risk that a crash will be
fatigue-related for each of the first twelve hours of driving time,
plus an aggregated figure for all driving in Hour 13 and beyond.
The figures ranged from less than 1% for Hour 1, to 4.4% for
Hour 10, to 9.6% for Hour 11, to 25% for Hour 13 and beyond.
Id. at 45 (J.A. 1665). As FMCSA observed, the “risk of . . . a
fatigue-related crash in the 11th hour of driving or later is
notably higher than in the 10th hour of driving.” 2005 Rule, 70
Fed. Reg. at 49,997.

But in the RIA that it released with the 2005 Rule, FMCSA
did not use the crash risk figures contained in the TIFA Study.
 18

Instead, the agency plotted the aggregate figure for Hour 13 and
beyond at Hour 17, and then “fit[] a cubic curve” (derived a
regression equation) for that and the other hour-by-hour figures
from the TIFA Study. 2005 RIA at 58-59 (J.A. 1678-79). (The
curve is reproduced in Part II.B.1 below.) Like the TIFA Study,
the curve estimated the actual percentage of crashes related to
fatigue for each hour of driving. The curve’s figures, however,
were different from those in the TIFA Study. See id. In
particular, the percentage difference between the figures for the
10th and 11th hours was substantially smaller than in the TIFA
Study. Finally, FMCSA divided each of the hourly figures from
the curve by the average risk for Hours 1 through 11, creating a
risk increase “relative to average driving hours.” Id. at 61 (J.A.
1681). The resulting “TOT [time-on-task] crash risk
multipliers” were then used in the operator-fatigue model to
determine the safety impact of different HOS rules. Id.

FMCSA used the operator-fatigue model, along with the
carrier-operations model, to determine the benefits and costs of
four regulatory options. Option 1 was the 2003 Rule. Option 2
was the 2005 Rule. Option 3 was the same as Option 2, but with
a 10-hour rather than 11-hour daily driving limit, with a 58-hour
rather than 34-hour restart provision, and without a sleeper-berth
exception. Finally, Option 4 was the same as Option 3, but with
a 44-hour restart provision. See 2005 Rule, 70 Fed. Reg. at
50,045; 2005 RIA at ES-1 (J.A. 1628). FMCSA concluded that
Option 2 -- the 2005 Rule -- was the most cost-effective of the
options. See 2005 Rule, 70 Fed. Reg. at 50,046; 2005 RIA at
ES-7 (J.A. 1634).

In addition to examining these four options, FMCSA
considered a variant of Option 2, which differed only in that it
imposed a 10-hour (as in the pre-2003 regulations) rather than
11-hour daily driving limit. Applying the same models, the
agency concluded that reducing the driving limit to 10 hours was
 19

“considerably less cost-effective than the basic version of
Option 2.” 2005 RIA at ES-7 (J.A. 1634); see 2005 Rule, 70
Fed. Reg. at 50,046-47.

3. Public Citizen objects to FMCSA’s reliance on the
operator-fatigue model because FMCSA did not disclose (inter
alia) the methodology by which it would derive time-on-task
multipliers until it published the 2005 Rule -- too late for
interested parties to comment. Because the time-on-task
multipliers were an integral part of the operator-fatigue model,
and because the output of that model was central to FMCSA’s
decision to adopt the 2005 Rule (and particularly the 11-hour
daily driving limit and 34-hour restart provision), the model and
its methodology were unquestionably among “the most critical
factual material that [was] used to support the agency’s
position.” Air Transp. Ass’n of Am., 169 F.3d at 7 (citation and
internal quotation mark omitted). The failure to provide an
opportunity for comment on the model’s methodology therefore
constitutes a violation of the APA’s notice-and-comment
requirements. See Chamber of Commerce, 443 F.3d at 902
(finding a violation where “extra-record materials suppl[ied] the
basic assumptions used by the [agency] to establish the range of
costs [of] complying with [a rule’s] conditions”).

FMCSA contends that “petitioners can hardly express
surprise at the use of the . . . model for assessing safety benefits,
since that model is an update of the [one] used in the 2003 RIA.”
FMCSA Br. 44. It is true that an agency does not violate the
APA if its “methodology remain[s] constant” and new data is
merely “used to check or confirm prior assessments.” Solite
Corp., 952 F.2d at 485; see Chamber of Commerce, 443 F.3d at
900 (stating that “further notice and comment are not required
when additional fact gathering merely supplements information
in the rulemaking record . . . without changing methodology”).
But FMCSA’s methodology did not remain constant. The
 20

operator-fatigue model may have employed an “update” of the
methodology disclosed in the 2003 RIA, but the nature of the
update -- the derivation of the time-on-task multipliers, and even
the use of time-on-task multipliers -- was entirely new.
Moreover, the addition of the time-on-task element to the model
was not a minor modification used to check or confirm prior
analyses: it constituted the agency’s response to an important
defect in its previous methodology identified by this court in
Public Citizen. See 374 F.3d at 1218-19.

Although FMCSA concedes that the time-on-task
multipliers were nowhere to be found in the 2003 RIA, it
contends that our opinion in Public Citizen should have put
interested parties on notice “that the agency would have to
adjust the model to account for [time-on-task] effects.” FMCSA
Br. 45. This, too, misses the point. Although interested parties
may have known that FMCSA would incorporate time-on-task
effects into its crash-risk model, they had no way of knowing
that the agency would calculate the impact of time on task in the
way that it did. In particular, there was no way for Public
Citizen to foresee the following elements of the agency’s
methodology, the import of which we discuss in Part II.B:

• Instead of using the figures from the TIFA Study to
determine crash risk as a function of time on task
(hours driving), FMCSA fit the figures to a cubic
curve. None of the studies in the rulemaking record
derived such a curve. An earlier version of the TIFA
Study, which was in the record, simply presented its
risk figures for each hour -- derived directly from the
crash data -- in bar-chart form. See Kenneth L.
Campbell & Michael H. Belzer, Hours of Service
Regulatory Evaluation Analytical Support 51 (2000)
(“2000 TIFA Study”). The same is true of the final
 21

TIFA Study, which was not released in time for
comment. See TIFA Study at 12 (J.A. 1727).

• The agency drew its curve by plotting all of the data
regarding fatigue-related crashes at Hour 13 and
beyond at Hour 17. The TIFA studies, by contrast,
aggregated all such data at “13+” hours. See (Final)
TIFA Study at 12 (J.A. 1727); 2000 TIFA Study at 51.

• After deriving hourly fatigue-related crash-risk
estimates using its curve, FMCSA divided each of
those figures by the average risk for the first 11 hours.
The TIFA studies did not do so. See (Final) TIFA
Study at 12 (J.A. 1727); 2000 TIFA Study at 51.

• The operator-fatigue model did not take into account
cumulative fatigue caused by the increased weekly
driving and working hours permitted under the 34-hour
restart provision.

In light of these undisclosed elements, we cannot say that the
agency’s operator-fatigue model was “made public in the
proceeding and exposed to refutation” as required by the APA.
Air Transp. Ass’n of Am., 169 F.3d at 7 (citation, internal
quotation mark, and emphasis omitted).

4. Finally, before we may vacate an agency action for
failure to disclose supporting documents (like the methodology
of the operator-fatigue model) during the notice-and-comment
period, we must take “due account . . . of the rule of prejudicial
error.” 5 U.S.C. § 706. “To show that error was prejudicial, a
[petitioner] must indicate with reasonable specificity what
portions of the documents it objects to and how it might have
responded if given the opportunity.” Gerber v. Norton, 294 F.3d
173, 182 (D.C. Cir. 2002) (internal quotation marks omitted).
 22

Moreover, a petitioner must “‘show that on remand [it] can
mount a credible challenge . . . and [was] thus prejudiced by the
absence of an opportunity to do so before’” the agency. Id. at
184 (quoting Utility Solid Waste Activities Group v. EPA, 236
F.3d 749, 755 (D.C. Cir. 2001)).

We have no difficulty in concluding that the agency’s
failure to disclose the methodology of the operator-fatigue
model in time for comment was prejudicial. The arguments that
the petitioner has raised before this court amply demonstrate that
it would have mounted a “credible challenge” had it been
afforded an opportunity to do so. See Chamber of Commerce,
443 F.3d at 904 (finding a sufficient showing of prejudice where
the petitioner’s objections had “creat[ed] enough” uncertainty as
to whether its “comments would have had some effect if they
had been considered” (citation and internal quotation marks
omitted)). We discuss those arguments in the following subpart.
Indeed, as we explain below, Public Citizen’s critique of the
model persuades us not only that it was prejudiced by FMCSA’s
failure to provide an opportunity for comment, but also that
FMCSA failed to provide an adequate explanation for its
decision to adopt the 11-hour daily driving limit and the 34-hour
restart provision.

B

Public Citizen charges that when FMCSA finally disclosed
its time-on-task methodology in the RIA that accompanied the
2005 Rule, it failed to provide a reasoned explanation for a
number of the methodology’s critical elements. To satisfy the
APA’s “arbitrary and capricious” standard, an agency must
“articulate a satisfactory explanation for its action including a
‘rational connection between the facts found and the choice
made.’” Motor Vehicle Mfrs. Ass’n v. State Farm Mut. Auto.
Ins. Co., 463 U.S. 29, 43 (1983) (quoting Burlington Truck
 23

Lines, Inc. v. United States, 371 U.S. 156, 168 (1962)). The
“agency must cogently explain why it has exercised its
discretion in a given manner,” id. at 48, and that explanation
must be “sufficient to enable us to conclude that the [agency’s
action] was the product of reasoned decisionmaking,” id. at 52.

1. Public Citizen notes that the TIFA data, upon which
FMCSA’s time-on-task multipliers were ultimately based,
indicates that “the risk of fatal-crash involvement more than
doubled from the 10th hour to the 11th.” Public Citizen Br. 48-
49 (citing 2005 RIA at 45 (J.A. 1665)). The actual time-on-task
multiplier for the eleventh hour used in FMCSA’s model,
however, was “only 30% higher than the . . . multiplier for the
10th hour.” Id. at 49 (citing 2005 RIA at 61 (J.A. 1681)).
Public Citizen contends that the two steps FMCSA used to
transform the TIFA data into the time-on-task multipliers were
unexplained, and that they had the effect of improperly
minimizing the crash risk associated with the 11th hour of
driving.

First, as explained above, instead of using the crash risk
figures for each hour of driving that the TIFA Study had
calculated directly from the actual crash data, FMCSA derived
a cubic curve of crash risk as a function of time on task. To
derive the curve, FMCSA first plotted the TIFA figures for
Hours 1 through 12, and then used an aggregate measure for
Hour 13 and beyond. It did not, however, plot the 13+ figure at
Hour 13, but rather at Hour 17. See 2005 RIA at 59 (J.A. 1679).
As shown in the accompanying graph, the curve that fit those 13
points yielded a crash risk at Hour 11 that was substantially
below the figure that the TIFA Study had calculated directly
from the actual crash data:
 24

2005 RIA at 59 (J.A. 1679). Moreover, as Public Citizen points
out, if the agency had plotted the figure for 13+ hours at Hour 13
rather than Hour 17, the resulting curve would have produced a
significantly higher estimate of the risk of a fatigue-related crash
at Hour 11 -- a figure close to that which the TIFA Study had
calculated directly. See Public Citizen Br. 49-50, A-3
(displaying alternative curve).

FMCSA’s decision to plot the data point for Hour 13 and
beyond at Hour 17 -- instead of at Hour 13 (or some other point)
-- was entirely unexplained in the RIA and final rule. This
complete lack of explanation for an important step in the
agency’s analysis was arbitrary and capricious. “When an
agency uses a computer model, it must ‘explain the assumptions
and methodology used in preparing the model.’” U.S. Air Tour
Ass’n v. FAA, 298 F.3d 997, 1008 (D.C. Cir. 2002) (quoting
Small Refiner Lead Phase-Down Task Force v. EPA, 705 F.2d
506, 535 (D.C. Cir. 1983)). Although the agency’s brief defends
the use of a cubic curve on the ground that the margin of error
 25

in the underlying TIFA data was relatively large,3 this again
misses the point: the issue is not whether a curve should have
been used, but why the agency chose to draw the curve by
plotting the Hour 13+ data at Hour 17. On that point, the brief
is entirely silent.4

Second, after deriving an estimate of crash risk for each
hour of driving using a cubic curve, FMCSA divided those
figures by the average risk for Hours 1 through 11, creating an
estimate of risk “relative to average driving hours.” 2005 RIA
at 61 (J.A. 1681); see id. at 46 (J.A. 1666). Public Citizen
objects that FMCSA gave no explanation for its decision to
divide each hourly risk figure by the average for the first 11

3
The Federal Register notice announcing the 2005 Rule stated
that the “TIFA data must be treated with caution,” because the number
of fatigue-related crashes in the database at the longer hours was small
and because the 2003 Rule (not yet in effect when the data was
collected) had changed the regulatory environment. 2005 Rule, 70
Fed. Reg. at 49,997. Nonetheless, the agency determined that the
TIFA data “represent[s] the only recently-published data available for
considering” time-on-task effects, and chose to rely on it as the
“prudent” and conservative course. Id. at 50,052.
4
FMCSA’s counsel did attempt to explain the derivation of the
Hour 17 figure at oral argument and in a post-argument letter to the
court. Counsel represented that “the agency averaged the numbers
concerning fatigue-related crashes after 12 hours,” and that the
average number of hours driven at the time of the crash for all crashes
after 12 hours was approximately 17. FMCSA Rule 28(j) Letter, Dec.
12, 2006; see Oral Arg. Recording at 44:15. But counsel also
conceded that FMCSA neither explained this calculation in the RIA
nor put the data on which it was based in the record. See FMCSA
Rule 28(j) Letter, Dec. 12, 2006. Whatever the merits of the agency’s
averaging methodology, we cannot affirm on the basis of a post-hoc
explanation by agency counsel. See SEC v. Chenery Corp., 332 U.S.
194, 196 (1947); Public Citizen, 374 F.3d at 1218.
 26

hours, and points out that the effect of this step was “to diminish
the increase[d] [risk of driving in Hour 11] by dividing the
heightened risk at the 11th hour by an average that includes that
heightened 11th-hour risk.” Public Citizen Br. 50. We express
no view on the validity of FMCSA’s statistical method, but we
agree with Public Citizen that -- once again -- the agency offered
no explanation for its decision during the rulemaking and failed
even to respond to the petitioner’s argument in its brief.
Although we apply a deferential standard of review to an
agency’s use of a statistical model, we cannot uphold a rule
based on such a model when an important aspect of its
methodology was wholly unexplained. See U.S. Air Tour Ass’n,
298 F.3d at 1008 (holding that, when a model’s methodology is
challenged, the agency must “provide a complete analytic
defense” (quoting Small Refiner Lead Phase-Down Task Force,
705 F.2d at 535) (internal quotation mark omitted)).5

2. Public Citizen further contends that the RIA’s operator-
fatigue model “also ignored cumulative fatigue from increased
weekly driving and working hours allowed by the 34-hour
restart.” Public Citizen Br. 51. In Public Citizen, we said --
with respect to the identical restart provision of the 2003 Rule --
that this increase in weekly hours was likely “‘an important
aspect of the problem,’” and that the “agency’s failure to
address” it made “the rule’s rationality questionable.” 374 F.3d
at 1222-23 (quoting State Farm, 463 U.S. at 43). Indeed, in the

5
FMCSA conducted a sensitivity analysis that concluded that an
11-hour daily driving limit remained more cost-effective than a 10-
hour limit, even if the agency assumed a higher risk of fatigue-related
crash in the 11th hour than it employed in its operator-fatigue model.
See 2005 Rule, 70 Fed. Reg. at 50,046-47; 2005 RIA at ES-7 to -8
(J.A. 1634-35). FMCSA does not argue, however, that this sensitivity
assumption was high enough to offset the two errors asserted by
Public Citizen.
 27

2000 NPRM, FMCSA expressed concern about evidence that
many drivers were exceeding the weekly limit of 60 hours in 7
days. See 2000 NPRM, 65 Fed. Reg. at 25,558. In particular,
the agency was troubled by a study showing, inter alia, that
“[t]wenty-five percent of drivers reported working at least 75
hours in the last 7 days.” Id. Yet, as FMCSA acknowledges, the
34-hour restart provision of the 2005 Rule could “allow another
17 hours of driving time . . . in a 7-day work week, compared to
the limit of 60 hours of driving time without the [restart]
provision.” 2005 Rule, 70 Fed. Reg. at 49,984. In light of these
statements by the agency, Public Citizen argues that the
operator-fatigue model should have taken into account the
increased crash risk caused by “cumulative fatigue” associated
with the increased driving and working hours that it would
permit.

FMCSA’s counsel responds with a single conclusory
sentence denying that the agency “ignored cumulative fatigue”
and citing three pages of the RIA. See FMCSA Br. 45 (citing
2005 RIA at 41, 44, C-20 (J.A. 1661, 1664, 1709)). The cited
pages, however, address a different kind of “cumulative fatigue”
-- the “sleep deficit[]” that “accumulates with successive sleep-
deprived days.” 2005 RIA at 41 (J.A. 1661). Public Citizen’s
argument, which it pressed in the rulemaking, is that longer
hours spent driving over the course of a few days have a fatigue-
inducing effect that is independent of that caused by insufficient
sleep. There is no indication that the operator-fatigue model
considered this kind of “cumulative fatigue,” and FMCSA offers
no explanation for the omission.

Of course, it could be that “cumulative fatigue” due to
longer weekly service hours will not constitute a significant
problem because, for example, the number of such hours will be
minimal. Indeed, in the final rule, FMCSA declared that it
“believes the average driver is not, and cannot realistically, drive
 28

and work the longer weekly hours, on a regular basis, as
described by some of the commenters.” 2005 Rule, 70 Fed.
Reg. at 50,022. But whatever the “average driver” will do on a
“regular basis,” it is clear that FMCSA contemplates that many
drivers will work those longer hours -- as those hours are the
basis for the agency’s conclusion that the 34-hour restart
provision will have economic benefits. See id. at 50,049; 2005
RIA at ES-3 to -4, 68-69 (J.A. 1630-31, 1688-89). FMCSA
concedes as much. See 2005 RIA at 18 (J.A. 1654) (“[M]any
drivers work and drive longer hours than the averages.”); id. at
67 (J.A. 1687) (“[M]ore than half of for-hire operations, and
somewhat less than half of private fleet operations, are intensive
enough to press the HOS limits, and should therefore be affected
by those limits.”).

In any event, FMCSA gave no explanation for the failure of
its operator-fatigue model to account for cumulative fatigue due
to the increased weekly driving and working hours permitted by
the 34-hour restart provision. This court may “not attempt itself
to make up for such deficiencies: We may not supply a reasoned
basis for the agency’s action that the agency itself has not
given.” State Farm, 463 U.S. at 43 (internal quotation marks
omitted); see SEC v. Chenery Corp., 332 U.S. 194, 196 (1947).
Accordingly, the agency’s failure of explanation renders the
restart provision arbitrary and capricious. Cf. U.S. Air Tour
Ass’n, 298 F.3d at 1019 (“[I]n the absence of any reasonable
justification for excluding non-tour aircraft from its noise model,
we must conclude that this aspect of the FAA’s methodology is
arbitrary and capricious . . . .”).

C

For the foregoing reasons, we conclude both that FMCSA
failed to provide an opportunity for comment on the
methodology of its operator-fatigue model, and that it failed to
 29

provide an explanation for the elements of that methodology that
Public Citizen disputes. Because the model is the basis for the
cost-benefit analysis that led FMCSA to adopt the two
provisions of the 2005 Rule that Public Citizen challenges -- the
increase in the daily driving limit from 10 to 11 hours, and the
34-hour restart provision -- we must vacate those provisions.
And because those are the only provisions of the 2005 Rule that
Public Citizen challenges, we have no occasion to address the
balance of its arguments.

III

We turn next to OOIDA’s petition, which challenges three
aspects of the 2005 Rule. First, OOIDA contends that FMCSA
failed to deal with loading and unloading issues, as required by
Congress. Second, OOIDA asserts that the 14-hour daily on-
duty limit is arbitrary and capricious, because FMCSA failed to
consider its negative effects on driver health and safety. Finally,
the petitioner objects to the modification of the sleeper-berth
exception on several grounds. The following subparts address
each of these challenges.

A

OOIDA’s first contention is that the 2005 Rule is arbitrary,
capricious, or otherwise contrary to law because “FMCSA did
not deal with the effects of loading and unloading operations on
driver fatigue as required by Congress.” OOIDA Br. 23. This
argument refers to Congress’ 1995 instruction to the agency to
conduct a rulemaking and issue a rule “dealing with a variety of
fatigue-related issues pertaining to commercial motor vehicle .
. . safety (including . . . loading and unloading operations . . .).”
49 U.S.C. § 31136 note. Loading and unloading operations
contribute to driver fatigue primarily by prolonging drivers’
hours. Because most drivers are paid by the mile or by the load
 30

rather than by the hour, it costs a shipper or receiver nothing to
keep drivers waiting to load or unload their vehicles. The
resulting delays associated with loading and unloading can
reduce the time available for sleep and require drivers to stay
awake longer in order to finish their trips.

An agency acts arbitrarily if it ignores an issue that
Congress directs it to address. See Public Citizen, 374 F.3d at
1216. In this case, however, the record confirms that FMCSA
complied with the congressional mandate by considering the
implications of loading and unloading operations for the
problem of driver fatigue. In fact, FMCSA adopted the 14-hour
daily on-duty limit, and eliminated the provision in the pre-2003
regulations that extended the daily limit when the driver took
breaks, in part because of those concerns. As the agency
explained:

Under the pre-2003 rules, drivers were allowed a 15-
cumulative-hour duty period but could extend their
maximum duty period indefinitely by taking off-duty
time during their workday. This perpetuated the
problem of excessive waiting time for pick up and
delivery of freight at shippers and receivers, because
the drivers were expected to place themselves in off-
duty status while waiting.

2005 Rule, 70 Fed. Reg. at 49,986. Under the 2005 Rule,
drivers can no longer extend their driving-eligible period by
taking off-duty breaks while waiting to load or unload. FMCSA
concluded that, as a consequence, the 2005 Rule “prevents
shippers, receivers, and companies from abusing the off-duty
hours and forcing drivers to use them as unpaid time.” Id. at
50,013.
 31

In support of this conclusion, the agency reviewed
comments from drivers and industry associations regarding how
the limit was working in practice. (It was instituted as part of
the 2003 Rule.) FMCSA explained that “[t]he general
consensus among drivers was that their workday, on average, is
shorter under the new rules. They no longer work 20-hour days
due to the 14-hour consecutive requirement.” Id.; see also id.
(“The [National Industrial Transportation League] commented
that . . . operations at loading docks have been reconfigured to
decrease dwell time and to expedite loading and unloading in
order to minimize driver on-duty time not devoted to driving .
. . .” (citation and internal quotation marks omitted)).

We conclude that, by expressly considering fatigue-related
issues pertaining to loading and unloading operations, and by
reasonably identifying the nonextendable 14-hour limit as
responsive to those issues, the 2005 Rule satisfied Congress’
mandate to “deal with” this problem. Although OOIDA may
have preferred that FMCSA deal with the problem in a different
manner, the statute does not mandate that the agency reach any
particular substantive result.

B

OOIDA’s next contention is that FMCSA acted arbitrarily
and capriciously by ignoring adverse health and safety effects of
its decision to make the 14-hour daily on-duty limit
nonextendable through the use of breaks. Specifically, OOIDA
contends that this provision “discourages drivers from taking
short rest breaks and naps,” because they can no longer exclude
such breaks from their daily on-duty limit. OOIDA Br. 43. In
support of this argument, OOIDA cites a study of its members
indicating that 60% of drivers “reported foregoing short breaks,
naps, and meals under the 2003 Rule” -- which was the first rule
to make the daily on-duty limit nonextendable. Id. at 44-45
 32

(citing Comments of OOIDA 6 (Mar. 10, 2005), Docket No.
FMCSA-2004-19608-1790 (J.A. 1193)).

FMCSA did not, however, “fail[] to consider” this aspect of
the problem. State Farm, 463 U.S. at 43. To the contrary, the
agency agreed that short naps and breaks “are an important tool
in combating fatigue.” 2005 Rule, 70 Fed. Reg. at 50,030. It
also acknowledged driver comments that “the consecutive duty
time requirement caused them to skip meals or naps.” Id. at
50,013. FMCSA concluded, however, that several other
considerations minimized or outweighed this disadvantage.

First, a FMCSA survey of operations under the 2003 Rule
indicated that the “vast majority of drivers are not using the full
14-consecutive hour duty tour. The data suggest that drivers
represented in the survey have time available within the current
14-hour duty tour to take breaks.” Id. Second, FMCSA found
that the nonextendable 14-hour daily on-duty limit has
significant offsetting benefits. Most important, the agency noted
that there is a consensus that “longer wakeful hours result in
alertness and performance degradation,” and that by barring
drivers from driving past 14 hours after the start of their shift,
the rule keeps the most fatigued drivers off the road. Id. at
50,014. In light of these explanations, we cannot say either that
the agency failed to consider this aspect of the problem, or that
its weighing of the relevant considerations was arbitrary and
capricious.

OOIDA also specifically objects to FMCSA’s decision that
the 2-hour break used to qualify (in part) for the sleeper-birth
exception to the daily off-duty requirement may not be used to
extend the 14-hour daily on-duty limit. See OOIDA Br. 45-46.
OOIDA contends that this aspect of the rule discourages drivers
from taking breaks during their remaining 12 hours of driving-
eligible time, and also “violates one of the central tenets” of the
 33

rulemaking because it “puts the driver on a backward rotating
schedule by two hours per day.” OOIDA Br. 46. (A driver who
alternates between 8-hour sleeper-berth shifts and 14-hour on-
duty periods will work 22-hour days.)

In responding to this point during the rulemaking, FMCSA
explained that the mandatory 2-hour break itself “provide[s] the
driver with the opportunity to nap, if and when needed.” 2005
Rule, 70 Fed. Reg. at 50,030. And while it is true that a driver
maximizing working time under the sleeper-berth exception
could theoretically end up on a backward-rotating, 22-hour
schedule, FMCSA never sought to impose a rigid 24-hour clock.
Instead, the agency simply concluded that the 2005 Rule
“promotes movement toward a 24-hour clock.” Id. at 50,014.
The possibility that some drivers will have 22-hour schedules
does not contradict this.

We therefore conclude that FMCSA neither ignored the
health and safety effects of making the 14-hour daily on-duty
limit nonextendable, nor otherwise acted arbitrarily in making
that determination.

C

Finally, OOIDA challenges the 2005 Rule’s modification of
the sleeper-berth exception. Under the rule, a driver using a
sleeper berth may divide the daily off-duty requirement -- which
otherwise consists of 10 consecutive hours -- into two periods.
One of the periods must consist of at least 8 continuous hours in
a sleeper berth; the other must be at least 2 hours long and can
be spent in a sleeper berth or elsewhere. This reflects a change
from the 2003 Rule, which had allowed drivers to divide the
required 10 hours into two sleeper-berth periods of any length
as long as each of them was at least 2 hours long. OOIDA
challenges this modification on two grounds. A group of
 34

intervenors led by the California Trucking Association adds a
third.

1. OOIDA’s first contention is that FMCSA violated the
APA because the 2005 NPRM did not give interested parties
adequate notice that the agency was considering such a
modification of the 2003 Rule’s sleeper-berth provision.

In order to comply with the APA’s notice requirement,
“[a]n agency’s final rule need only be a ‘logical outgrowth’ of
its notice.” Covad Commc’ns Co. v. FCC, 450 F.3d 528, 548
(D.C. Cir. 2006). The “logical outgrowth” test is satisfied if
interested parties “‘should have anticipated’ the agency’s final
course in light of the initial notice.” Id. (quoting Small Refiner
Lead Phase-Down Task Force, 705 F.2d at 548-49); see
Crawford v. FCC, 417 F.3d 1289, 1295-96 (D.C. Cir. 2005). As
the Supreme Court recently explained, the object of the logical
outgrowth test “is one of fair notice.” Long Island Care at
Home, Ltd. v. Coke, 127 S. Ct. 2339, 2351 (2007).

OOIDA asserts that the 2005 NPRM “was too broad and
unfocused to provide adequate notice[,] thus precluding
meaningful comment.” OOIDA Br. 30. It is certainly true that
a notice can be “too general to be adequate.” Small Refiner
Lead Phase-Down Task Force, 705 F.2d at 549. “Agency notice
must describe the range of alternatives being considered with
reasonable specificity[;] [o]therwise, interested parties will not
know what to comment on.” Id. The 2005 NPRM, however,
was sufficiently specific to satisfy this requirement.6

6
Simultaneously taking a somewhat opposite tack, OOIDA also
argues that the NPRM was too specific because it listed “virtually
every option for changing the sleeper berth exception” and thus “really
proposed nothing.” OOIDA Br. 32. But as explained below, there
was no “needle in a haystack” quality to the NPRM.
 35

The 2005 NPRM set out the 2003 Rule “as the ‘proposal’
on which public comments [were] sought.” 70 Fed. Reg. at
3339. FMCSA then asked “the public to comment on what
changes to that rule, if any, [were] necessary to respond to the
concerns raised by the court” in Public Citizen. Id. With
respect to the sleeper-berth exception, the notice outlined
specific options as follows:

FMCSA will consider a variety of possible changes to
the sleeper-berth provisions, including but not limited
to: (1) Not permitting any split sleeper-berth use to
count toward the minimum 10-hours off duty, (2)
allowing one continuous sleeper-berth period of less
than 10 hours, such as 8 hours, to substitute for the
otherwise minimum 10 hours, (3) eliminating split-
sleeper-berth periods or establishing a minimum time
for one of the two “splits,” such as 5 hours, 8 hours, or
some other appropriate level, (4) revising the manner
in which sleeper-berth periods affect the calculation of
the 14-consecutive-hour period, and (5) restricting
variations on permissible sleeper-berth use to team
drivers only.

Id. at 3349-50 (emphasis added).

The third option, italicized above, forecast the terms that
would ultimately appear in the 2005 Rule -- or very nearly so.
Indeed, the prospect that a new sleeper-berth exception would
contain a minimum requirement of 8 hours for one of the two
periods (or “splits”) could hardly have been a surprise to
anyone. As noted in Part I.A above, our opinion in Public
Citizen had sharply criticized the 2003 Rule for containing an
exception directly at odds with “the conceded central premise of
the HOS regulations, . . . that ‘[e]ach driver should have an
opportunity for eight consecutive hours of uninterrupted sleep
 36

every day.’” 374 F.3d at 1219 (quoting 2003 Rule, 68 Fed. Reg.
at 22,469).

OOIDA acknowledges that the NPRM’s option (3) was
close to the 8-hour minimum of the final rule. See OOIDA Br.
32. Nonetheless, the petitioner objects that even that option did
not indicate that the second sleeper-berth period would have to
be at least 2 hours long. But FMCSA actually directed the
attention of interested parties to this issue as well. See 2005
NPRM, 70 Fed. Reg. at 3350 (“If one [sleeper-berth] period is
[required to be] 7 or more hours in length, . . . would a second
sleeper-berth period still be required?”). We conclude that, in
light of the options discussed in the NPRM, the modification of
the sleeper-berth exception in the final rule was “reasonably
foreseeable” and hence satisfied the logical outgrowth test.
Long Island Care at Home, 127 S. Ct. at 2351.7

2. OOIDA’s second contention is that FMCSA “failed to
consider important issues” related to the modification of the
sleeper-berth exception. OOIDA Br. 40. As we noted above, an
agency’s action is “arbitrary and capricious if the agency has .
. . entirely failed to consider an important aspect of the
problem.” State Farm, 463 U.S. at 43. OOIDA claims that
FMCSA failed to consider two issues relating to the impact of
the modification on team drivers.

7
OOIDA further notes that FMCSA received a large volume of
unsolicited comments criticizing the sleeper-berth exception after it
issued the final rule, and contends that this proves that interested
parties did not anticipate the provision. In fact, the post-promulgation
outpouring may merely indicate that the commenters strenuously
opposed the final rule. In any event, it tells us little about what was
“reasonably foreseeable,” which is the crux of the logical outgrowth
test. Long Island Care at Home, 127 S. Ct. at 2351 (emphasis added);
see Small Refiner Lead Phase-Down Task Force, 705 F.2d at 548-49.
 37

First, OOIDA contends that FMCSA failed to consider the
adverse economic impact that the 8-hour requirement will have
on team drivers. OOIDA explains that the 8-hour requirement
will make it impossible for team drivers to synchronize their on-
and off-duty schedules so that when one driver reaches the end
of his or her shift, that driver’s partner is eligible to begin
driving. The net result is “[l]ess total available driving time”
and therefore “fewer miles” and “less money” for team drivers.
OOIDA Br. 42.

Contrary to OOIDA’s claim, FMCSA did not “fail[] to
consider” this aspect of the problem. In fact, FMCSA
acknowledged comments from trucking companies and others
contending that the burden of changes to the sleeper-berth
exception would fall primarily on team drivers. See 2005 Rule,
70 Fed. Reg. at 50,028-29. But the agency concluded that this
marginal loss of productivity and flexibility was justified by
gains in other areas, and particularly by the reduction in fatigue
and fatigue-related accidents. See id. at 50,031. OOIDA may
disagree with this policy balance, but it does not reflect a failure
to consider relevant factors.

Second, OOIDA contends that the modified rule causes
particular problems for those team drivers who carry hazardous
materials. OOIDA concedes, however, that these issues were
not raised with FMCSA during the comment period. See
OOIDA Br. 40; Oral Arg. Recording at 11:15.8 We therefore
reject this argument without addressing its merits. See Nat’l
Wildlife Fed’n v. EPA, 286 F.3d 554, 562 (D.C. Cir. 2002) (“It
is well established that issues not raised in comments before the

8
While conceding that these issues were not raised, OOIDA
maintains that this was because of inadequate notice that the final
sleeper-berth exception might include an 8-hour break requirement.
We have rejected that argument in Part III.C.1.
 38

agency are waived and this Court will not consider them.”);
accord National Ass’n of Clean Air Agencies v. EPA, ___ F.3d
___, slip op. at 14-15 (D.C. Cir. June 1, 2007).

3. Finally, intervenor California Trucking Association
(CTA) challenges the 8-hour requirement on an additional
ground not raised by OOIDA: that the record does not support
FMCSA’s finding that drivers need 7 to 8 consecutive hours of
sleep each day. See CTA Br. 18-40. We reject this challenge as
well.

First, CTA contends that there is no evidence that drivers
who split their sleep into two shorter periods are more likely to
be involved in fatigue-related accidents. A review of the record,
however, reveals ample support for FMCSA’s view.9 The
record also contradicts CTA’s assertion that FMCSA conflated
“split sleep” -- sleep obtained in two or more periods -- with
“fragmented sleep” -- sleep interrupted every few minutes. See
2005 Rule, 70 Fed. Reg. at 50,026.

Second, CTA argues that FMCSA misinterpreted the only
study it cited in stating that “[t]he split-sleeper berth exception
is also problematic from a driver health standpoint.” Id. at
50,027. But FMCSA based the modification of the sleeper-berth
exception not on concerns about driver health but rather on its
well-supported finding that the 8-hour requirement would
reduce fatigue-related accidents. See id. at 50,026-31.
Therefore, even if the agency’s passing mention of health effects

9
See, e.g., National Transp. Safety Bd., Factors that Affect
Fatigue in Heavy Truck Accidents (1995), Docket No.
FMCSA-2004-19608-2013 (J.A. 1517); Robin P. Hertz, Tractor
Trailer Driver Fatality: The Role of Nonconsecutive Rest in a Sleeper
Berth (1987), Docket No. FMCSA-2004-19608-2011 (J.A. 1493); see
also 2005 Rule, 70 Fed. Reg. at 49,994.
 39

were unsupported, the error was harmless because the context
makes clear that it did not affect the agency’s conclusion. Cf.
National Ass’n of Home Builders v. Defenders of Wildlife, 127
S. Ct. 2518, 2530 (2007) (“[W]e do not believe that this stray
statement, which could have had no effect on the underlying
agency action being challenged, requires that we . . . remand[]
to the agency for clarification.”). We therefore reject the last of
the challenges to the sleeper-berth exception.

IV

For the foregoing reasons, we deny OOIDA’s petition. At
the same time, we grant Public Citizen’s petition and vacate
those portions of the 2005 Rule that increase the daily driving
limit from 10 to 11 hours, and that permit an off-duty period of
34 hours to restart the weekly on-duty limits.

So ordered.