Notice: This opinion is subject to formal revision before publication in the
Federal Reporter or U.S.App.D.C. Reports. Users are requested to notify
the Clerk of any formal errors in order that corrections may be made
before the bound volumes go to press.
United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT
Argued May 13, 2004 Decided July 20, 2004
No. 03-1304
PUBLIC CITIZEN, INC. AND
CENTER FOR AUTO SAFETY,
PETITIONERS
v.
NATIONAL HIGHWAY TRAFFIC SAFETY ADMINISTRATION AND
NORMAN Y. MINETA, SECRETARY OF TRANSPORTATION,
RESPONDENTS
AUTOMOTIVE OCCUPANT RESTRAINTS COUNCIL AND
ALLIANCE OF AUTOMOBILE MANUFACTURERS,
INTERVENORS
On Petition for Review of an Order of the
National Highway Traffic Safety Administration
Scott L. Nelson argued the cause for petitioners. With him
on the briefs was David C. Vladeck. Michael E. Tankersley
entered an appearance.
Bills of costs must be filed within 14 days after entry of judgment.
The court looks with disfavor upon motions to file bills of costs out
of time.
2
H. Thomas Byron III, Attorney, U.S. Department of Jus-
tice, argued the cause for respondents. With him on the brief
were Peter D. Keisler, Assistant Attorney General; Douglas
N. Letter, Attorney; Jeffrey A. Rosen, General Counsel, U.S.
Department of Transportation; Paul M. Geier, Assistant
General Counsel; Lloyd S. Guerci, Assistant Chief Counsel,
National Highway Traffic Safety Administration; and Enid
Rubenstein, Attorney.
Erika Z. Jones argued the cause for intervenors in support
of respondents. With her on the brief were Adam Sloane
and David M. Gossett.
Shari T. Kendall, Stephen L. Oesch, and Michele M. Fields
were on the brief for amicus curiae Insurance Institute for
Highway Safety in support of respondents.
Before: EDWARDS, RANDOLPH, and TATEL, Circuit Judges.
Opinion for the Court filed by Circuit Judge TATEL.
TATEL, Circuit Judge: Air bags have saved thousands of
lives since first appearing in passenger vehicles approximate-
ly three decades ago. Their ability to save some occupants,
however, has proven fatal to others, especially children and
small women. Responding to this problem and to a new
congressional directive, the National Highway Traffic Safety
Administration revised one of its auto safety standards to
improve air bags’ life-saving benefits while reducing their
potentially deadly risks. In this case, we consider a challenge
to one aspect of that new standard: the agency’s decision to
set the speed for unbelted vehicle crash testing at twenty-five
rather than thirty miles per hour. Because in doing so the
agency acted consistently with Congress’s directive and nei-
ther arbitrarily nor capriciously, we deny the petition for
review.
I.
In 1993, the National Highway Traffic Safety Administra-
tion (NHTSA) began requiring manufacturers to install air
bags in new cars and light trucks. See Federal Motor
Vehicle Safety Standards; Occupant Crash Protection, 58
3
Fed. Reg. 46,551, 46,553 (Sept. 2, 1993) (codified at 49 C.F.R.
§ 571.208 (2003)). Under NHTSA’s Federal Motor Vehicle
Safety Standard No. 208, manufacturers had to certify that
their air-bag equipped vehicles would protect occupants in the
event of a crash. Specifically, auto makers had to show that
their vehicles satisfied certain injury criteria limits in simulat-
ed rigid barrier crashes at speeds up to and including thirty
miles per hour, using both belted and unbelted fiftieth-
percentile adult male dummies. See id. at 46,552; Federal
Motor Vehicle Safety Standards; Occupant Crash Protection,
65 Fed. Reg. 30,680, 30,741 (May 12, 2000). In plain English,
Standard No. 208 required auto makers to show that a test
dummy representing an average-sized man—both with and
without a seat belt—would avoid serious injury (as defined by
the standard) when his car slams into a fixed barrier at thirty
miles per hour.
Air bags installed in response to this mandate saved thou-
sands of lives. Because air bags are designed to inflate
almost instantly upon impact, however, the force of the
inflation can injure, even kill, smaller occupants sitting too
close to the deploying bag. See Federal Motor Vehicle Safety
Standards; Occupant Crash Protection, 62 Fed. Reg. 12,960,
12,960–61 (Mar. 19, 1997). As of February 1997, NHTSA had
documented thirty-eight crashes in which the force of a
deploying air bag had killed a child. Id. at 12,960. Twenty-
one drivers and two adult passengers had also died from air-
bag induced injuries. Id. at 12,960–61.
Reacting to these fatalities and to a growing public outcry,
NHTSA amended Standard No. 208 in March 1997 to encour-
age manufacturers to redesign air bags quickly to make them
inflate with less force. Id. at 12,961–62. Under the revised
rule, manufacturers no longer had to test vehicles using the
thirty mile per hour unbelted crash test. Instead, they could
use a thirty mile per hour ‘‘sled test’’—a test roughly equiva-
lent to a twenty-two mile per hour crash test—to measure
vehicle ability to protect fiftieth-percentile male dummies.
See id. at 12,974; 65 Fed. Reg. at 30,689. In a sled test, a
vehicle placed on a sled is accelerated rapidly backward, but
never actually crashed into a barrier. 65 Fed. Reg. at 30,738.
4
As the car moves backward, the test dummies lurch forward,
simulating an actual crash; the cars’ air bags are manually
deployed at a pre-selected time. Id. By permitting auto
makers to certify their vehicles’ crash-protection systems
through this less stringent safety test, NHTSA made it easier
for them to maintain compliance with Standard No. 208 while
‘‘depowering’’—i.e., reducing the force of—air bag inflation.
Although recognizing that depowered air bags could reduce
protection for unbelted adults and teenagers, NHTSA con-
cluded that ‘‘the opportunity to avoid the deaths of a signifi-
cant number of children who would otherwise be fatally
injured by air bags’’ justified its depowering rule. 62 Fed.
Reg. at 12,964. The agency maintained that ‘‘it is not accept-
able that a safety device cause a significant number of
fatalities in circumstances in which fatal or serious injuries
would not otherwise occur.’’ Id. Indeed, NHTSA found it
‘‘particularly unacceptable that the vehicle occupants being
fatally injured are young children[ ] and that the number of
those deaths is steadily growing.’’ Id. Nevertheless, be-
cause of the possible safety trade-offs associated with bag
depowering and because expected technological advances
could reduce risks to children and small women without
diminishing protection for unbelted adults and teens,
NHTSA’s 1997 rule provided that the sled test option would
terminate in September 2001. See id. at 12,967–69. The
agency explained:
[T]here is no need to permanently reduce Standard
No. 208’s performance requirements to enable man-
ufacturers to fully address the adverse effects of air
bags. This is because there are various alternatives,
albeit with longer technological development and
implementation leadtimes than depowering, that are
already allowed by the standard and that appear
likely to result in equal or greater benefits than
depowering without creating adverse safety trade-
offs. Thus, the agency views depowering as an
interim approach, while the vehicle manufacturers
develop and implement better solutions.
5
Id. at 12,968. Responding to the increased flexibility provid-
ed by the sled test option, manufacturers began installing
what are known as ‘‘redesigned air bags’’—air bags designed
to pass the sled test. See 65 Fed. Reg. at 30,738. Although
many redesigned air bags deployed with less force than their
predecessors, they still inflated with more power than needed
to comply with the sled test. Id.; id. at 30,689.
In June 1998, Congress stepped in and directed NHTSA to
require manufacturers to install a new generation of air bags
known as ‘‘advanced air bags.’’ In language central to the
issue before us, the Transportation Equity Act for the 21st
Century (TEA 21) provides:
[T]he Secretary of Transportation shall issue a no-
tice of proposed rulemaking to improve occupant
protection for occupants of different sizes, belted
and unbelted, under Federal Motor Vehicle Safety
Standard No. 208, while minimizing the risk to in-
fants, children, and other occupants from injuries
and deaths caused by air bags, by means that in-
clude advanced air bags.
Pub. L. No. 105–178, § 7103(a)(1), 112 Stat. 107, 466 (1998)
(codified at 49 U.S.C. § 30127 note (2000)). ‘‘Advanced air
bags’’ incorporate new technologies or designs that either
prevent air bags from deploying in inappropriate circum-
stances or ensure that they inflate in low-risk ways. See 65
Fed. Reg. at 30,738. For example, advanced air bags could
include a device that senses the weight of the occupant and
then prevents the air bag from activating if the occupant is a
child. See id. (defining ‘‘occupant weight sensors’’).
Also significant to the challenge before us, TEA 21 super-
seded NHTSA’s 1997 decision to sunset the sled test by
September 2001, providing instead that the ‘‘requirements of
S13 of Standard No. 208 [prescribing the thirty mile per hour
unbelted vehicle sled test] shall remain in effect unless and
until changed by the rule required by this subsection.’’ TEA
21 § 7103(a)(4).
6
Responding to TEA 21, NHTSA issued a Notice of Pro-
posed Rulemaking (NPRM) in September 1998, proposing a
battery of new, complex vehicle-safety performance tests.
See Federal Motor Vehicle Safety Standards: Occupant
Crash Protection, 63 Fed. Reg. 49,958 (Sept. 18, 1998). To
satisfy TEA 21’s mandate to ‘‘improve occupant protection,’’
NHTSA proposed phasing out the prevailing sled test and
reinstating the thirty mile per hour rigid barrier crash test
for the fiftieth-percentile male dummy. See id. at 49,970–71.
NHTSA also proposed (among other things) requiring manu-
facturers to satisfy the unbelted and belted rigid barrier
crash test with a fifth-percentile female dummy—a dummy
representing a small woman. Id. at 49,972. With respect to
TEA 21’s requirement that the rule minimize risk to infants
and children, NHTSA proposed various crash test alterna-
tives. See id. at 49,973–75. For example, to protect infants
in rear-facing child seats, the NPRM included two alternative
series of tests from which manufacturers could choose: tests
proving that air bags could deactivate themselves or tests
demonstrating that air bags would deploy in a low-risk way.
Id. at 49,973. Although vehicle manufacturers initially re-
sponded to the proposed rule by supporting retention of the
sled test, they later expressed their consensus view that the
revised Standard No. 208 should include an unbelted crash
test with a maximum speed of twenty-five miles per hour, not
thirty. See Federal Motor Vehicle Safety Standards; Occu-
pant Crash Protection, 64 Fed. Reg. 60,556, 60,560 (Nov. 5,
1999).
In November 1999, NHTSA issued a Supplemental Notice
of Proposed Rulemaking (SNPRM), explaining that ‘‘[t]he
public comments and the agency research and analysis since
our 1998 NPRM have enabled us to refine and in some cases
simplify the proposed amendments that we are considering.’’
Id. at 60,557. Of particular significance, NHTSA proposed
two alternative unbelted crash tests instead of the single
thirty mile per hour rigid barrier test included in its initial
NPRM. Only one of these alternatives is relevant here: an
unbelted rigid barrier test with a maximum speed ‘‘to be
established in the final rule within the range of’’ twenty-five
7
to thirty miles per hour. Id. at 60,558–59. As to this
alternative, the agency explained that ‘‘the potential exists for
a phase-in sequence in which the maximum speed would
initially be set at TTT 25 mph[ ] to provide vehicle manufactur-
ers additional flexibility when they are introducing advanced
air bags,’’ and then increased to thirty miles per hour ‘‘after a
reasonable period of time.’’ Id. at 60,563. Following issu-
ance of the SNPRM, several transportation safety organiza-
tions, including the National Transportation Safety Board, the
Insurance Institute for Highway Safety, the National Safety
Council, the American Trauma Society, and the National
Association of Governors’ Highway Safety Representatives
expressed to NHTSA their ‘‘strong[ ]’’ opposition to returning
to the thirty mile per hour standard. See Letter from James
E. Hall, Chairman, National Transportation Safety Board, et
al., to the Honorable Rodney E. Slater, Secretary, United
States Department of Transportation 2 (Feb. 16, 2000). That
standard, they explained, ‘‘would result in some vehicles’
airbag inflator power returning to levels that have caused
occupant deaths and injuries without increasing the benefits
to unbelted occupants in crashes at higher speeds.’’ Id. at 1.
NHTSA then submitted a draft final regulation to the
Office of Management and Budget for review. See Memoran-
dum from Stephen P. Wood, Assistant Chief Counsel for
Rulemaking, National Highway Traffic Safety Administration,
to Docket No. NHTSA 00–7013; Notice 1 at 1 (May 10, 2000)
[hereinafter ‘‘Wood Memorandum’’]; see also Exec. Order
No. 12,866 § 6(a)(3)(B)(i), 58 Fed. Reg. 51,735 (Sept. 30, 1993)
(requiring agencies to submit certain draft regulations to
OMB’s Office of Information and Regulatory Affairs). The
draft proposed an unbelted rigid barrier test to be imple-
mented in two stages. During the first stage, from Septem-
ber 2003 through August 2006, manufacturers would have to
test vehicles using a maximum test speed of twenty-five miles
per hour, while in the second stage, from September 2007
through August 2010, the maximum test speed would increase
to thirty miles per hour. See Wood Memorandum at 1.
Two months later, NHTSA published an ‘‘interim final
rule,’’ establishing an unbelted rigid barrier crash test with a
8
maximum test speed of twenty-five miles per hour for the
period extending from September 2003 through August 2006.
See 49 C.F.R. § 571.208 S5.1.2; 65 Fed. Reg. at 30,685.
NHTSA explained:
The provisions of this rule, particularly the maxi-
mum test speed for the unbelted rigid barrier test,
reflect the uncertainty associated with simultaneous-
ly achieving the twin goals of TEA 21. This uncer-
tainty leads us to take an approach that best assures
improved air bag protection for occupants of all
sizes, without compromising efforts to reduce the
risks of injury to vulnerable occupants, including
children and short women seated very close to air
bags and out-of-position occupants. Such an ap-
proach is one that involves the least uncertainty for
the occupants who have been most at risk. As long
as the manufacturers improve the already substan-
tial overall level of real world protection provided by
current redesigned air bags, the uncertainty associ-
ated with the challenge of simultaneously achieving
the twin goals of TEA 21 is best resolved at this
point in favor of minimizing risk. This is especially
true in the early stages of the introduction of ad-
vanced air bag technologies.
65 Fed. Reg. at 30,680.
NHTSA’s final rule also added a wide range of new safety
performance tests. Among other things, the revised Stan-
dard No. 208 requires auto makers to use an entire family of
test dummies—not just the fiftieth-percentile adult male dum-
my, but also new dummies representing fifth-percentile adult
females, six-year-old children, three-year-old children, and
one-year-old infants. Id. at 30,685. In addition, the rule
includes new and more stringent injury criteria. Id. at
30,691.
Eight petitions for reconsideration were filed, including one
by Public Citizen and several other consumer groups request-
ing that the agency amend the unbelted rigid barrier test to
require a thirty mile per hour maximum test speed for
9
passenger cars while retaining a twenty-five mile per hour
top speed for light trucks, vans, and sport utility vehicles.
NHTSA denied the reconsideration petition, reiterating the
rationales offered in the final rule. Federal Motor Vehicle
Safety Standards; Occupant Crash Protection, 66 Fed. Reg.
65,376, 65,379–82 (Dec. 18, 2001).
Public Citizen and others filed petitions for review in the
Ninth Circuit. After addressing various jurisdictional issues,
the Ninth Circuit transferred the petition filed by Public
Citizen and the Center for Auto Safety to this court. Pub.
Citizen Inc. v. Mineta, 343 F.3d 1159, 1171 (9th Cir. 2003).
(Throughout this opinion, we shall refer to petitioners Public
Citizen and the Center for Auto Safety collectively as ‘‘Public
Citizen.’’) In its petition, Public Citizen challenges only one
element of the revised Standard No. 208: NHTSA’s decision
to set the maximum unbelted crash test speed at twenty-five
rather than thirty miles per hour. Public Citizen argues that
the twenty-five mile per hour test speed (1) violates TEA 21’s
requirement that Standard No. 208 ‘‘improve occupant pro-
tection for occupants of different sizes, belted and unbelted,’’
and (2) amounts to arbitrary and capricious agency action.
We consider each claim in turn.
II.
Because Public Citizen challenges NHTSA’s interpretation
of a statute the agency is charged with implementing, we
apply the two-part test of Chevron U.S.A. Inc. v. Natural
Resources Defense Council, Inc., 467 U.S. 837 (1984). We
thus begin by asking ‘‘whether Congress has directly spoken
to the precise question at issue,’’ for if ‘‘the intent of Congress
is clear, that is the end of the matterTTTT [T]he court, as
well as the agency, must give effect to the unambiguously
expressed intent of Congress.’’ Id. at 842–43. If the statute
is ‘‘silent or ambiguous with respect to the specific issue,’’ we
then determine whether NHTSA’s interpretation ‘‘is based on
a permissible construction of the statute.’’ Id. at 843.
Public Citizen does not claim that its statutory challenge
can be resolved at Chevron step one, and for good reason.
10
TEA 21’s plain language—requiring NHTSA ‘‘to improve
occupant protection for occupants of different sizes, belted
and unbelted, under Federal Motor Vehicle Safety Standard
No. 208, while minimizing the risk to infants, children, and
other occupants from injuries and deaths caused by air bags,
by means that include advanced air bags’’—unambiguously
requires only three things: Standard No. 208 must (1) ‘‘im-
prove occupant protection,’’ (2) ‘‘minimiz[e] the risk to infants,
children, and other occupants’’ from air-bag induced injuries,
and (3) accomplish these two goals at least partly through the
use of ‘‘advanced air bags.’’ Nowhere does TEA 21 say
anything about the particular vehicle-testing requirements
NHTSA must adopt, much less the speed the agency must
use in its unbelted rigid barrier crash test.
To resolve Public Citizen’s claim that NHTSA’s twenty-five
mile per hour unbelted test speed violates TEA 21’s require-
ment to ‘‘improve occupant protection for occupants of differ-
ent sizes, belted and unbelted’’—a Chevron step two issue—
we must first answer an antecedent question: improve pro-
tection compared to what? Specifically, does TEA 21 require
improvement measured from the sled test, as NHTSA con-
tends, or from the amount of protection provided by rede-
signed air bags in use at the time of NHTSA’s rulemaking, as
Public Citizen argues? If we can reasonably read TEA 21 to
permit NHTSA to use the sled test as the baseline of
comparison, then this becomes an easy case, for as Public
Citizen concedes, the twenty-five mile per hour crash test
improves protection over the sled test. In particular, the
twenty-five mile per hour standard tests vehicles at a faster
speed than the thirty mile per hour sled test, which has an
effective crash test speed of twenty-two miles per hour. In
addition, unlike the sled test, the twenty-five mile per hour
rigid barrier test measures vehicle performance in actual
collisions. See 65 Fed. Reg. at 30,698 (stating that because
the sled test does not actually crash the vehicle, ‘‘it cannot
measure the performance provided by the vehicle structure in
combination with the air bags or even the full air bag system
by itself’’). Given these more rigorous testing features,
NHTSA estimates that air bags designed to pass the twenty-
11
five mile per hour crash test would save at least sixty-four to
144 more lives than air bags designed to pass the generic sled
test. Id. at 30,735.
According to Public Citizen, however, the sled test cannot
serve as the standard from which to measure the improve-
ment TEA 21 requires because NHTSA, in its 1997 rulemak-
ing, made the sled test temporary and because Congress, in
TEA 21, preserved the sled test only until the agency promul-
gates a new crash test standard. Public Citizen argues that
in order to satisfy Congress’s directive to improve occupant
protection, NHTSA must issue a standard that forces manu-
facturers to design air bags to provide better protection than
that provided by the redesigned air bags in use at the time of
NHTSA’s rulemaking—a showing the agency cannot make,
Public Citizen contends, because those air bags already satis-
fy the more stringent thirty mile per hour crash test. More-
over, Public Citizen claimed at oral argument that NHTSA
itself had interpreted TEA 21 as requiring the agency to
measure improvement by reference to the level of ‘‘real world
protection,’’ not the sled test’s regulatory standard. Counsel
stated:
[O]n the very first page of the final rule[,] [the
agency says] that the occupant protection criterion is
satisfied ‘‘as long as the manufacturers improve the
already substantial overall level of real-world protec-
tion provided by current redesigned air bags.’’
That’s where they on the very first page of their
final rule defined the baseline, and that’s the base-
line that they don’t satisfy.
Tr. of Oral Argument at 55 (emphasis added).
Public Citizen’s arguments are unconvincing. Not only
does TEA 21 section 7103(a)(1) require NHTSA to improve
protection ‘‘under Federal Motor Vehicle Safety Standard
No. 208,’’ TEA 21 § 7103(a)(1) (emphasis added), which in-
cluded the sled test at the time the statute was enacted, but
section 7103(a)(4) expressly authorizes NHTSA to retain the
sled test: the ‘‘requirements of S13 of Standard No. 208
12
[prescribing the vehicle sled test] shall remain in effect unless
and until changed by the rule required by this subsection,’’ id.
§ 7103(a)(4) (emphasis added). As the congressional confer-
ees explained, this provision ensures that ‘‘[t]he availability of
the current sled test certification option available under [Fed-
eral Motor Vehicle Safety Standard] 208 (S13) remains in
effect unless and until phased out according to the schedule in
the final rule.’’ H.R. CONF. REP. NO. 105–550, at 521 (1998),
reprinted in 1998 U.S.C.C.A.N. 70, 196. Because Congress
directed NHTSA to keep the sled test ‘‘unless’’ otherwise
phased out, NHTSA’s use of that test as the baseline of
comparison can hardly be unreasonable. See Chevron, 467
U.S. at 843.
We also disagree with Public Citizen that NHTSA acted
inconsistently with its own interpretation of TEA 21.
NHTSA did not, as counsel claimed at oral argument, inter-
pret TEA 21 as requiring improvement to be measured by
reference to redesigned air bags actually installed in vehicles.
Under the heading, ‘‘Rationales for Protection Improvement
Requirements TTT Selection of TTT 25 mph[ ] as [the] Top
Speed for [the] Unbelted Rigid Barrier Test,’’ 65 Fed. Reg. at
30,686, NHTSA maintained that its ‘‘decision to replace the
TTT 30 mph[ ] generic sled test with the TTT 25 mph[ ]
unbelted rigid barrier test requires a significantly higher level
of safety [because] the sled test is roughly equivalent to a TTT
22 mph[ ] rigid barrier TTT crash,’’ id. at 30,689. Moreover,
responding to comments that a twenty-five mile per hour
standard would violate TEA 21’s protection-improvement
mandate, NHTSA explained:
We also note that the suggestion that TEA 21
somehow requires an unbelted barrier test with a
test speed not lower than TTT 30 mph[ ] is inconsis-
tent with the language of that statute. In fact, TEA
21 expressly left open the possibility of our retaining
the sled test. That test has a severity level signifi-
cantly below that of a TTT 30 mph[ ] barrier test and
a TTT 25 mph[ ] barrier test.
13
Id. at 30,705 (emphasis added). And despite Public Citizen’s
claim to the contrary, NHTSA never said that ‘‘the occupant
protection criterion is satisfied ‘as long as the manufacturers
improve the already substantial overall level of real-world
protection provided by current redesigned air bags.’’’ Tr. of
Oral Argument at 55. NHTSA actually said this: ‘‘As long as
the manufacturers improve the already substantial overall
level of real world protection provided by current redesigned
air bags, the uncertainty associated with the challenge of
simultaneously achieving the twin goals of TEA 21 is best
resolved at this point in favor of minimizing risk.’’ 65 Fed.
Reg. at 30,680. Read in its entirety and in the context of
NHTSA’s clear explanation of its view of TEA 21, this
sentence cannot plausibly be read to mean that the agency
viewed the ‘‘level of real world protection’’ as the statutorily
required baseline of comparison.
III.
We turn next to Public Citizen’s arbitrary-and-capricious
challenge. Under the Administrative Procedure Act, we will
‘‘hold unlawful and set aside agency action, findings, and
conclusions found to be TTT arbitrary, capricious, an abuse of
discretion, or otherwise not in accordance with law.’’ 5
U.S.C. § 706(2)(A) (2000). As the Supreme Court explained
in its seminal decision in Motor Vehicle Manufacturers Ass’n
of the United States, Inc. v. State Farm Mutual Automobile
Insurance Co., 463 U.S. 29 (1983), a case that also involved
air bags and Standard No. 208:
The scope of review under the ‘‘arbitrary and capri-
cious’’ standard is narrow and a court is not to
substitute its judgment for that of the agency. Nev-
ertheless, the agency must examine the relevant
data and articulate a satisfactory explanation for its
action including a rational connection between the
facts found and the choice made. In reviewing that
explanation, we must consider whether the decision
was based on a consideration of the relevant factors
and whether there has been a clear error of judg-
14
ment. Normally, an agency rule would be arbitrary
and capricious if the agency has relied on factors
which Congress has not intended it to consider,
entirely failed to consider an important aspect of the
problem, offered an explanation for its decision that
runs counter to the evidence before the agency, or is
so implausible that it could not be ascribed to a
difference in view or the product of agency exper-
tise.
Id. at 43 (citations and internal quotation marks omitted).
With this highly deferential standard in mind, we consider
each of Public Citizen’s specific claims.
Public Citizen contends first that nothing in the record
supports NHTSA’s conclusion that the twenty-five mile per
hour unbelted test speed is ‘‘in the best overall interest of
safety.’’ 65 Fed. Reg. at 30,687. According to Public Citizen,
unless NHTSA found that ‘‘retaining the 30 mph test speed
could result in a new risk of substantial fatalities or serious
injuries that would outweigh the hundreds of fatalities among
teenagers and adults TTT that might flow from reducing the
30 mph test to [a] 25 mph standard,’’ NHTSA could not
rationally conclude that a twenty-five mile per hour standard
best serves the overall interests of safety. Pet’rs’ Br. at 35–
36. Because NHTSA failed to make such a finding, Public
Citizen concludes, the agency violated the APA standard of
reasoned decisionmaking. We disagree.
To begin with, the factual premise underlying Public Citi-
zen’s claim—that establishing a twenty-five mile per hour
unbelted test speed will result in hundreds of new fatalities
among unbelted teens and adults—is unsupported by the
record, for it rests on the assumption, rejected by NHTSA,
that manufacturers will depower air bags to the minimum
level permitted by the twenty-five mile per hour standard.
NHTSA explained that auto makers (1) have no economic
incentive to depower air bags because doing so would achieve
no significant cost savings, (2) did not depower their bags to
the bare minimum permitted by the sled test, see supra p. 5,
and (3) have indicated that, due to other regulatory require-
15
ments, they will not do so now. 65 Fed. Reg. at 30,689; see
also id. at 30,704 (stating that NHTSA’s Final Economic
Analysis shows that air bags certified under a twenty-five
mile per hour standard cost essentially the same as those
certified to a thirty mile per hour standard). Given these
findings, NHTSA’s prediction that manufacturers will not
likely reduce the protectiveness of current air bags is reason-
able. ‘‘Predictions regarding the actions of regulated entities
are precisely the type of policy judgments that courts routine-
ly and quite correctly leave to administrative agencies.’’ Pub.
Utils. Comm’n v. FERC, 24 F.3d 275, 281 (D.C. Cir. 1994).
Moreover, NHTSA did not, as Public Citizen claims, rely on
manufacturers’ ‘‘voluntary action’’ to satisfy TEA 21’s protec-
tion-improvement requirement. Pet’rs’ Br. at 41. As we
have just explained, NHTSA’s revised Standard No. 208
improves occupant protection by requiring auto makers to
satisfy a performance test (the twenty-five mile per hour
crash test) that is more stringent than the agency’s preexist-
ing test (the sled test). Public Citizen cites Public Citizen v.
Nuclear Regulatory Commission, 901 F.2d 147 (D.C. Cir.
1990), but there, interpreting a very different statute, we held
that an agency ordered by Congress to promulgate binding
regulatory requirements may not issue a non-binding policy
statement that encourages but does not compel action. Id. at
157. Here, by contrast, NHTSA’s revised Standard No. 208
imposes mandatory testing standards. In other words, man-
ufacturers must meet NHTSA’s twenty-five mile per hour
unbelted crash test, and NHTSA’s conclusion that manufac-
turers are unlikely to depower to the minimum permitted by
the twenty-five mile per hour standard does not transform
that mandatory standard into an optional one.
Public Citizen’s challenge to NHTSA’s conclusion that the
revised Standard No. 208 serves the best overall interest of
safety fails for a second reason: NHTSA explained, reason-
ably in our view, why a twenty-five mile per hour unbelted
test speed, considered in the context of the entire rule, serves
the agency’s overall safety goals. Given the complex array of
new requirements imposed by the rule—including additional
crash test dummies, revised injury criteria, and new risk-
16
minimization tests—NHTSA was ‘‘concerned about the diffi-
culties of trying to meet the unbelted rigid barrier test at TTT
30 mph[ ] with both adult dummies while simultaneously
trying to reduce the risks of air bag-induced injuries and
deaths.’’ 65 Fed. Reg. at 30,687. And because air bags,
unlike other safety devices, can themselves injure or even kill
vehicle occupants, NHTSA thought it should ‘‘be cautious in
how far and how fast vehicle manufacturers are required to
advance the state of advanced air bag technologies in their
vehicles.’’ Id. NHTSA continued:
Since a significant percentage of current vehicles can
already satisfy the new unbelted barrier crash test
at TTT 25 mph[ ] with both the 5th percentile adult
female dummy and the 50th percentile adult male
dummy, we conclude that setting the maximum
speed at that level will help vehicle manufacturers to
focus their resources and compliance efforts TTT on
meeting the risk reduction requirements.
Id. Giving manufacturers the ability to so focus their efforts,
NHTSA maintained, would help ensure that ‘‘the installation
of advanced air bag technologies by the vehicle manufacturers
across the full spectrum of their fleets [is] done correctly—
the first timeTTTT Compared with a TTT 30 mph[ ] unbelted
rigid barrier test, a TTT 25 mph[ ] unbelted rigid barrier test
presents less chance of inadvertently increasing risks to out-
of-position occupants.’’ Id. at 30,688. In view of NHTSA’s
twin statutory obligations—to improve occupant protection
while minimizing the risks of air-bag induced injuries—and
our highly deferential standard of review, we have no basis
for second-guessing the agency’s safety assessment.
Public Citizen next argues that ‘‘NHTSA TTT acted unrea-
sonably by reducing the maximum test speed based on ‘uncer-
tainties’ about how air bag improvements will be implemented
even though it has no evidence that these uncertainties
warrant such action.’’ Pet’rs’ Br. at 25. Again, we disagree.
NHTSA did not, as Public Citizen suggests, ‘‘merely recite
the terms ‘substantial uncertainty’ as a justification for its
actions.’’ State Farm, 463 U.S. at 52. Instead, just as the
17
Supreme Court required in State Farm, ‘‘[t]he agency TTT
explain[ed] the evidence which is available, and TTT offer[ed]
a rational connection between the facts found and the choice
made.’’ Id. (quoting Burlington Truck Lines, Inc. v. United
States, 371 U.S. 156, 168 (1962)) (internal quotation marks
omitted). Specifically, NHTSA stated that although it be-
lieved that vehicles could eventually pass a thirty mile per
hour unbelted rigid barrier crash test without compromising
safety for children and small women, ‘‘[o]ur laboratory tests
and knowledge of advanced technologies do not tell us how or
when developments might reach that point. They also do not
provide us with a full picture of the real world consequences
of adopting that test speed.’’ 65 Fed. Reg. at 30,688. In
addition, with respect to certain vehicles such as light trucks
and vans, NHTSA stated that at the time it issued the final
rule, its experience testing such vehicles with fifth-percentile
female dummies was limited. Id. at 30,689. The record also
shows that NHTSA had cause for concern over how manufac-
turers could simultaneously achieve TEA 21’s protection-
improvement and risk-minimization goals: in the agency’s
thirty mile per hour unbelted crash tests, over half the
vehicles failed to meet the new injury criteria for small female
occupants, one of the groups most at risk of air-bag induced
injuries. See National Highway Traffic Safety Administra-
tion, Final Economic Assessment, FMVSS No. 208 Advanced
Air Bags at E–3, IV–32 to IV–33 (May 2000). And despite
Public Citizen’s claim to the contrary, NHTSA did consider
alternatives—including the thirty mile per hour crash test
standard that Public Citizen favors—and coherently explained
why the twenty-five mile per hour standard better accommo-
dated its uncertainty about future technological develop-
ments. See 65 Fed. Reg. at 30,686–89.
Public Citizen calls our attention to Public Citizen v. Steed,
733 F.2d 93 (D.C. Cir. 1984), in which we granted a petition
for review challenging NHTSA’s decision to suspend a tire-
testing regulation. In that case, however, NHTSA was ‘‘un-
certain[ ]’’ whether the evidence on which it had relied to
suspend the regulation truly supported its ultimate decision.
Id. at 101 (‘‘As for the evidence indicating flaws in the test
18
procedures themselves, NHTSA’s statement explaining the
suspension [of the rule] reflects considerable uncertainty con-
cerning the extent to which those flaws lead to significant
inaccuracies in the test results.’’). NHTSA had no such
uncertainty here. It merely explained, based on entirely
rational reasons, its inability to forecast the pace or nature of
technological change; NHTSA had no uncertainty about that
uncertainty.
For its final APA argument, Public Citizen contends that
NHTSA acted arbitrarily and capriciously by failing to ex-
plain cogently why it declined to increase the unbelted test
speed to thirty miles per hour after the year 2006. This
claim fails because NHTSA offered rational reasons for
adopting an ‘‘interim final rule’’ establishing the unbelted
crash test speed through August 2006 only. The agency
explained that because it could not ‘‘assess whether the
uncertainty about the manufacturers’ ability to improve pro-
tection further and minimize risk simultaneously will persist’’
into the future, it would leave open the question of the post-
2006 unbelted crash test speed, using that time instead to
undertake ‘‘a multi-year effort to obtain additional data.’’ 65
Fed. Reg. at 30,685. ‘‘Based on the results of those informa-
tion gathering and analysis efforts’’ and public input, NHTSA
said that it would then make a final decision regarding the
maximum test speed for unbelted dummy testing in the long
run. Id. We see no defect in this explanation, for nothing in
the APA precludes an agency from collecting data and moni-
toring real-world experience with regulatory standards before
adopting new standards governing periods of time far into the
future—especially in cases, as here, that involve unpredict-
able technological change. Indeed, gathering evidence before
making a long-term decision is eminently sensible. See Nat’l
Ass’n of Broadcasters v. FCC, 740 F.2d 1190, 1211 (D.C. Cir.
1984) (holding that the FCC acted reasonably in postponing a
decision on certain details of a spectrum allocation rule
because ‘‘the Commission acted against an evolving back-
ground,’’ certain relevant factors could not ‘‘be known at the
19
present,’’ and ‘‘[a]s a result, too fine a calibration of the
relocation [of spectrum space] would have been premature’’).
At bottom, Public Citizen’s arbitrary-and-capricious chal-
lenge boils down to a policy disagreement with NHTSA.
Public Citizen believes that NHTSA should have set the
unbelted test speed at thirty miles per hour. Perhaps the
record could have supported such a standard. But because
NHTSA’s selection of twenty-five miles per hour is both
supported by the record and rationally explained, we have no
basis for substituting Public Citizen’s views for the agency’s,
particularly given NHTSA’s judgment that doing so would
increase the risk of harm to children and small women. See
State Farm, 463 U.S. at 43.
IV.
The petition for review is denied.
So ordered.