JUSTICE WHITE delivered the opinion of the Court.
The development of the automobile gave Americans unprecedented freedom to travel, but exacted a high price for
The Act also authorizes judicial review under the provisions of the Administrative Procedure Act (APA), 5 U. S. C. § 706, of all "orders establishing, amending, or revoking a Federal motor vehicle safety standard," 15 U. S. C. § 1392(b). Under this authority, we review today whether NHTSA acted arbitrarily and capriciously in revoking the requirement in Motor Vehicle Safety Standard 208 that new motor vehicles produced after September 1982 be equipped with passive restraints to protect the safety of the occupants of the vehicle in the event of a collision. Briefly summarized, we hold that the agency failed to present an adequate basis and explanation for rescinding the passive restraint requirement and that the agency must either consider the matter further or adhere to or amend Standard 208 along lines which its analysis supports.
The regulation whose rescission is at issue bears a complex and convoluted history. Over the course of approximately 60 rulemaking notices, the requirement has been imposed, amended, rescinded, reimposed, and now rescinded again.
As originally issued by the Department of Transportation in 1967, Standard 208 simply required the installation of seatbelts in all automobiles. 32 Fed. Reg. 2415. It soon became apparent that the level of seatbelt use was too low to reduce traffic injuries to an acceptable level. The Department therefore began consideration of "passive occupant restraint systems" — devices that do not depend for their effectiveness
In 1969, the Department formally proposed a standard requiring the installation of passive restraints, 34 Fed. Reg. 11148, thereby commencing a lengthy series of proceedings. In 1970, the agency revised Standard 208 to include passive protection requirements, 35 Fed. Reg. 16927, and in 1972, the agency amended the Standard to require full passive protection for all front seat occupants of vehicles manufactured after August 15, 1975. 37 Fed. Reg. 3911. In the interim, vehicles built between August 1973 and August 1975 were to carry either passive restraints or lap and shoulder belts coupled with an "ignition interlock" that would prevent starting the vehicle if the belts were not connected.
In preparing for the upcoming model year, most car makers chose the "ignition interlock" option, a decision which was highly unpopular, and led Congress to amend the Act to prohibit a motor vehicle safety standard from requiring or permitting compliance by means of an ignition interlock or a continuous buzzer designed to indicate that safety belts were not in use. Motor Vehicle and Schoolbus Safety Amendments of 1974, Pub. L. 93-492, § 109, 88 Stat. 1482, 15 U. S. C. § 1410b(b). The 1974 Amendments also provided that any safety standard that could be satisfied by a system other than seatbelts would have to be submitted to Congress where it could be vetoed by concurrent resolution of both Houses. 15 U. S. C. § 1410b(b)(2).
The effective date for mandatory passive restraint systems was extended for a year until August 31, 1976. 40 Fed. Reg. 16217 (1975); id., at 33977. But in June 1976, Secretary of Transportation William T. Coleman, Jr., initiated a new rulemaking on the issue, 41 Fed. Reg. 24070. After hearing testimony and reviewing written comments, Coleman extended the optional alternatives indefinitely and suspended the passive restraint requirement. Although he found passive
Coleman's successor as Secretary of Transportation disagreed. Within months of assuming office, Secretary Brock Adams decided that the demonstration project was unnecessary. He issued a new mandatory passive restraint regulation, known as Modified Standard 208. 42 Fed. Reg. 34289 (1977); 49 CFR § 571.208 (1978). The Modified Standard mandated the phasing in of passive restraints beginning with large cars in model year 1982 and extending to all cars by model year 1984. The two principal systems that would satisfy the Standard were airbags and passive belts; the choice of which system to install was left to the manufacturers. In Pacific Legal Foundation v. Department of Transportation, 193 U. S. App. D. C. 184, 593 F.2d 1338, cert. denied, 444 U.S. 830 (1979), the Court of Appeals upheld Modified Standard 208 as a rational, nonarbitrary regulation consistent with the agency's mandate under the Act. The Standard also survived scrutiny by Congress, which did not exercise its authority under the legislative veto provision of the 1974 Amendments.
Over the next several years, the automobile industry geared up to comply with Modified Standard 208. As late as July 1980, NHTSA reported:
In February 1981, however, Secretary of Transportation Andrew Lewis reopened the rulemaking due to changed economic circumstances and, in particular, the difficulties of the automobile industry. 46 Fed. Reg. 12033. Two months later, the agency ordered a one-year delay in the application of the Standard to large cars, extending the deadline to September 1982, id., at 21172, and at the same time, proposed the possible rescission of the entire Standard. Id., at 21205. After receiving written comments and holding public hearings, NHTSA issued a final rule (Notice 25) that rescinded the passive restraint requirement contained in Modified Standard 208.
In a statement explaining the rescission, NHTSA maintained that it was no longer able to find, as it had in 1977, that the automatic restraint requirement would produce significant safety benefits. Notice 25, id., at 53419. This judgment reflected not a change of opinion on the effectiveness of the technology, but a change in plans by the automobile industry. In 1977, the agency had assumed that airbags would be installed in 60% of all new cars and automatic seatbelts in 40%. By 1981 it became apparent that automobile manufacturers planned to install the automatic seatbelts in approximately 99% of the new cars. For this reason, the lifesaving potential of airbags would not be realized. Moreover, it now appeared that the overwhelming majority of passive belts
In view of the possibly minimal safety benefits, the automatic restraint requirement no longer was reasonable or practicable in the agency's view. The requirement would require approximately $1 billion to implement and the agency did not believe it would be reasonable to impose such substantial costs on manufacturers and consumers without more adequate assurance that sufficient safety benefits would accrue. In addition, NHTSA concluded that automatic restraints might have an adverse effect on the public's attitude toward safety. Given the high expense and limited benefits of detachable belts, NHTSA feared that many consumers would regard the Standard as an instance of ineffective regulation, adversely affecting the public's view of safety regulation and, in particular, "poisoning . . . popular sentiment toward efforts to improve occupant restraint systems in the future." Id., at 53424.
State Farm Mutual Automobile Insurance Co. and the National Association of Independent Insurers filed petitions for review of NHTSA's rescission of the passive restraint Standard. The United States Court of Appeals for the District of Columbia Circuit held that the agency's rescission of the passive restraint requirement was arbitrary and capricious. 220 U. S. App. D. C. 170, 680 F.2d 206 (1982). While observing that rescission is not unrelated to an agency's refusal to take action in the first instance, the court concluded that, in this case, NHTSA's discretion to rescind the passive restraint requirement had been restricted by various forms of congressional "reaction" to the passive restraint issue. It then
The court allowed NHTSA 30 days in which to submit a schedule for "resolving the questions raised in th[e] opinion." Id., at 206, 680 F. 2d, at 242. Subsequently, the agency filed a Notice of Proposed Supplemental Rulemaking setting forth a schedule for complying with the court's mandate. On August 4, 1982, the Court of Appeals issued an order staying the compliance date for the passive restraint requirement until September 1, 1983, and requested NHTSA to inform the court whether that compliance date was achievable. NHTSA informed the court on October 1, 1982, that based on representations by manufacturers, it did not appear that practicable compliance could be achieved before September 1985. On November 8, 1982, we granted certiorari, 459 U.S. 987, and on November 18, the Court of Appeals entered an order recalling its mandate.
Unlike the Court of Appeals, we do not find the appropriate scope of judicial review to be the "most troublesome
Petitioner Motor Vehicle Manufacturers Association (MVMA) disagrees, contending that the rescission of an agency rule should be judged by the same standard a court would use to judge an agency's refusal to promulgate a rule in the first place — a standard petitioner believes considerably narrower than the traditional arbitrary-and-capricious test. We reject this view. The Act expressly equates orders "revoking" and "establishing" safety standards; neither that Act nor the APA suggests that revocations are to be treated as refusals to promulgate standards. Petitioner's view would render meaningless Congress' authorization for judicial review of orders revoking safety rules. Moreover, the revocation of an extant regulation is substantially different than a failure to act. Revocation constitutes a reversal of the agency's former views as to the proper course. A "settled course of behavior embodies the agency's informed judgment that, by pursuing that course, it will carry out the policies
In so holding, we fully recognize that "[r]egulatory agencies do not establish rules of conduct to last forever," American Trucking Assns., Inc. v. Atchison, T. & S. F. R. Co., 387 U.S. 397, 416 (1967), and that an agency must be given ample latitude to "adapt their rules and policies to the demands of changing circumstances." Permian Basin Area Rate Cases, 390 U.S. 747, 784 (1968). But the forces of change do not always or necessarily point in the direction of deregulation. In the abstract, there is no more reason to presume that changing circumstances require the rescission of prior action, instead of a revision in or even the extension of current regulation. If Congress established a presumption from which judicial review should start, that presumption — contrary to petitioners' views — is not against safety regulation, but against changes in current policy that are not justified by the rulemaking record. While the removal of a regulation may not entail the monetary expenditures and other costs of enacting a new standard, and, accordingly, it may be easier for an agency to justify a deregulatory action, the direction in which an agency chooses to move does not alter the standard of judicial review established by law.
The Department of Transportation accepts the applicability of the "arbitrary and capricious" standard. It argues that under this standard, a reviewing court may not set aside an agency rule that is rational, based on consideration of the relevant factors, and within the scope of the authority delegated to the agency by the statute. We do not disagree with
The Court of Appeals correctly found that the arbitrary-and-capricious test applied to rescissions of prior agency regulations, but then erred in intensifying the scope of its review based upon its reading of legislative events. It held that congressional reaction to various versions of Standard 208 "raise[d] doubts" that NHTSA's rescission "necessarily demonstrates an effort to fulfill its statutory mandate," and therefore the agency was obligated to provide "increasingly clear and convincing reasons" for its action. 220 U. S. App. D. C., at 186, 193, 680 F. 2d, at 222, 229. Specifically, the Court of Appeals found significance in three legislative occurrences:
From these legislative acts and nonacts the Court of Appeals derived a "congressional commitment to the concept of automatic crash protection devices for vehicle occupants." Ibid.
This path of analysis was misguided and the inferences it produced are questionable. It is noteworthy that in this Court respondent State Farm expressly agrees that the post-enactment legislative history of the Act does not heighten the
That we should not be so quick to infer a congressional mandate for passive restraints is confirmed by examining the postenactment legislative events cited by the Court of Appeals. Even were we inclined to rely on inchoate legislative action, the inferences to be drawn fail to suggest that NHTSA acted improperly in rescinding Standard 208. First, in 1974 a mandatory passive restraint standard was technically not in effect, see n. 6, supra; Congress had no reason to foreclose that course. Moreover, one can hardly infer support for a mandatory standard from Congress' decision to provide that such a regulation would be subject to disapproval by resolutions of disapproval in both Houses. Similarly, no mandate can be divined from the tabling of resolutions of disapproval which were introduced in 1977. The failure of Congress to exercise its veto might reflect legislative deference to the agency's expertise and does not indicate that Congress would disapprove of the agency's action in 1981. And even if Congress favored the Standard in 1977, it — like NHTSA — may well reach a different judgment, given changed circumstances four years later. Finally, the Court of Appeals read too much into floor action on the 1980 authorization bill, a bill which was not enacted into law. Other
The ultimate question before us is whether NHTSA's rescission of the passive restraint requirement of Standard 208 was arbitrary and capricious. We conclude, as did the Court of Appeals, that it was. We also conclude, but for somewhat different reasons, that further consideration of the issue by the agency is therefore required. We deal separately with the rescission as it applies to airbags and as it applies to seatbelts.
The first and most obvious reason for finding the rescission arbitrary and capricious is that NHTSA apparently gave no consideration whatever to modifying the Standard to require that airbag technology be utilized. Standard 208 sought to achieve automatic crash protection by requiring automobile manufacturers to install either of two passive restraint devices: airbags or automatic seatbelts. There was no suggestion in the long rulemaking process that led to Standard 208 that if only one of these options were feasible, no passive restraint standard should be promulgated. Indeed, the agency's original proposed Standard contemplated the installation of inflatable restraints in all cars.
The agency has now determined that the detachable automatic belts will not attain anticipated safety benefits because so many individuals will detach the mechanism. Even if this conclusion were acceptable in its entirety, see infra, at 51-54, standing alone it would not justify any more than an amendment of Standard 208 to disallow compliance by means of the one technology which will not provide effective passenger protection. It does not cast doubt on the need for a passive restraint standard or upon the efficacy of airbag technology. In its most recent rulemaking, the agency again acknowledged the lifesaving potential of the airbag:
Given the effectiveness ascribed to airbag technology by the agency, the mandate of the Act to achieve traffic safety would suggest that the logical response to the faults of detachable seatbelts would be to require the installation of airbags. At the very least this alternative way of achieving the objectives of the Act should have been addressed and adequate reasons given for its abandonment. But the agency not only did not require compliance through airbags, it also did not even consider the possibility in its 1981 rulemaking. Not one sentence of its rulemaking statement discusses the airbags-only option. Because, as the Court of Appeals stated, "NHTSA's . . . analysis of airbags was nonexistent," 220 U. S. App. D. C., at 200, 680 F. 2d, at 236, what we said in Burlington Truck Lines, Inc. v. United States, 371 U. S., at 167, is apropos here:
We have frequently reiterated that an agency must cogently explain why it has exercised its discretion in a given manner,
The automobile industry has opted for the passive belt over the airbag, but surely it is not enough that the regulated industry has eschewed a given safety device. For nearly a decade, the automobile industry waged the regulatory equivalent of war against the airbag
Although the agency did not address the mandatory airbag option and the Court of Appeals noted that "airbags seem to have none of the problems that NHTSA identified in passive seatbelts," 220 U. S. App. D. C., at 201, 680 F. 2d, at 237, petitioners recite a number of difficulties that they
Petitioners also invoke our decision in Vermont Yankee Nuclear Power Corp. v. Natural Resources Defense Council, Inc., 435 U.S. 519 (1978), as though it were a talisman under which any agency decision is by definition unimpeachable. Specifically, it is submitted that to require an agency to consider an airbags-only alternative is, in essence, to dictate to the agency the procedures it is to follow. Petitioners both misread Vermont Yankee and misconstrue the nature of the remand that is in order. In Vermont Yankee, we held that a court may not impose additional procedural requirements upon an agency. We do not require today any specific procedures
Although the issue is closer, we also find that the agency was too quick to dismiss the safety benefits of automatic seatbelts. NHTSA's critical finding was that, in light of the industry's plans to install readily detachable passive belts, it could not reliably predict "even a 5 percentage point increase as the minimum level of expected usage increase." 46 Fed. Reg. 53423 (1981). The Court of Appeals rejected this finding because there is "not one iota" of evidence that Modified Standard 208 will fail to increase nationwide seatbelt use by at least 13 percentage points, the level of increased usage necessary for the Standard to justify its cost. Given the lack of probative evidence, the court held that "only a well justified refusal to seek more evidence could render rescission non-arbitrary." 220 U. S. App. D. C., at 196, 680 F. 2d, at 232.
Petitioners object to this conclusion. In their view, "substantial uncertainty" that a regulation will accomplish its intended purpose is sufficient reason, without more, to rescind a regulation. We agree with petitioners that just as an agency reasonably may decline to issue a safety standard if it is uncertain about its efficacy, an agency may also revoke a
In these cases, the agency's explanation for rescission of the passive restraint requirement is not sufficient to enable us to conclude that the rescission was the product of reasoned decisionmaking. To reach this conclusion, we do not upset the agency's view of the facts, but we do appreciate the limitations of this record in supporting the agency's decision. We start with the accepted ground that if used, seatbelts unquestionably would save many thousands of lives and would prevent tens of thousands of crippling injuries. Unlike recent regulatory decisions we have reviewed, Industrial Union Dept. v. American Petroleum Institute, 448 U.S. 607 (1980); American Textile Mfrs. Institute, Inc. v. Donovan, 452 U.S. 490 (1981), the safety benefits of wearing seatbelts are not in doubt, and it is not challenged that were those benefits to accrue, the monetary costs of implementing the Standard would be easily justified. We move next to the fact that there is no direct evidence in support of the agency's finding that detachable automatic belts cannot be predicted
But accepting the agency's view of the field tests on passive restraints indicates only that there is no reliable real-world experience that usage rates will substantially increase. To be sure, NHTSA opines that "it cannot reliably predict even a 5 percentage point increase as the minimum level of
The agency is correct to look at the costs as well as the benefits of Standard 208. The agency's conclusion that the incremental costs of the requirements were no longer reasonable was predicated on its prediction that the safety benefits of the regulation might be minimal. Specifically, the
The agency also failed to articulate a basis for not requiring nondetachable belts under Standard 208. It is argued that the concern of the agency with the easy detachability of the currently favored design would be readily solved by a continuous passive belt, which allows the occupant to "spool out" the belt and create the necessary slack for easy extrication from the vehicle. The agency did not separately consider the continuous belt option, but treated it together with the ignition interlock device in a category it titled "Option of Adopting Use-Compelling Features." 46 Fed. Reg. 53424
By failing to analyze the continuous seatbelts option in its own right, the agency has failed to offer the rational connection between facts and judgment required to pass muster under the arbitrary-and-capricious standard. We agree with the Court of Appeals that NHTSA did not suggest that the emergency release mechanisms used in nondetachable belts are any less effective for emergency egress than the buckle release system used in detachable belts. In 1978, when General Motors obtained the agency's approval to install a continuous passive belt, it assured the agency that nondetachable belts with spool releases were as safe as detachable belts with buckle releases. 43 Fed. Reg. 21912, 21913-21914 (1978). NHTSA was satisfied that this belt design assured easy extricability: "[t]he agency does not believe that the use of [such] release mechanisms will cause serious occupant egress problems . . . ." Id., at 52493, 52494. While the agency is entitled to change its view on the acceptability of continuous passive belts, it is obligated to explain its reasons for doing so.
The agency also failed to offer any explanation why a continuous passive belt would engender the same adverse public reaction as the ignition interlock, and, as the Court of Appeals concluded, "every indication in the record points the other way." 220 U. S. App. D. C., at 198, 680 F. 2d, at 234.
"An agency's view of what is in the public interest may change, either with or without a change in circumstances. But an agency changing its course must supply a reasoned analysis . . . ." Greater Boston Television Corp. v. FCC, 143 U. S. App. D. C. 383, 394, 444 F.2d 841, 852 (1970) (footnote omitted), cert. denied, 403 U.S. 923 (1971). We do not accept all of the reasoning of the Court of Appeals but we do conclude that the agency has failed to supply the requisite "reasoned analysis" in this case. Accordingly, we vacate the judgment of the Court of Appeals and remand the cases to that court with directions to remand the matter to the NHTSA for further consideration consistent with this opinion.
JUSTICE REHNQUIST, with whom THE CHIEF JUSTICE, JUSTICE POWELL, and JUSTICE O'CONNOR join, concurring in part and dissenting in part.
I join Parts I, II, III, IV, and V-A of the Court's opinion. In particular, I agree that, since the airbag and continuous
I do not believe, however, that NHTSA's view of detachable automatic seatbelts was arbitrary and capricious. The agency adequately explained its decision to rescind the Standard insofar as it was satisfied by detachable belts.
The statute that requires the Secretary of Transportation to issue motor vehicle safety standards also requires that "[e]ach such . . . standard shall be practicable [and] shall meet the need for motor vehicle safety." 15 U. S. C. § 1392(a) (1976 ed., Supp. V). The Court rejects the agency's explanation for its conclusion that there is substantial uncertainty whether requiring installation of detachable automatic belts would substantially increase seatbelt usage. The agency chose not to rely on a study showing a substantial increase in seatbelt usage in cars equipped with automatic seatbelts and an ignition interlock to prevent the car from being operated when the belts were not in place and which were voluntarily purchased with this equipment by consumers. See ante, at 53, n. 16. It is reasonable for the agency to decide that this study does not support any conclusion concerning the effect of automatic seatbelts that are installed in all cars whether the consumer wants them or not and are not linked to an ignition interlock system.
The Court rejects this explanation because "there would seem to be grounds to believe that seatbelt use by occasional users will be substantially increased by the detachable passive belts," ante, at 54, and the agency did not adequately explain its rejection of these grounds. It seems to me that the agency's explanation, while by no means a model, is adequate. The agency acknowledged that there would probably be some increase in belt usage, but concluded that the increase would be small and not worth the cost of mandatory
The agency explicitly stated that it will increase its educational efforts in an attempt to promote public understanding, acceptance, and use of passenger restraint systems. 46 Fed. Reg. 53425 (1981). It also stated that it will "initiate efforts with automobile manufacturers to ensure that the public will have [automatic crash protection] technology available. If this does not succeed, the agency will consider regulatory action to assure that the last decade's enormous advances in crash protection technology will not be lost." Id., at 53426.
The agency's changed view of the standard seems to be related to the election of a new President of a different political party. It is readily apparent that the responsible members of one administration may consider public resistance and uncertainties to be more important than do their counterparts in a previous administration. A change in administration brought about by the people casting their votes is a perfectly reasonable basis for an executive agency's reappraisal of the costs and benefits of its programs and regulations. As long as the agency remains within the bounds established by Congress,
"The promotion of motor vehicle safety through voluntary standards has largely failed. The unconditional imposition of mandatory standards at the earliest practicable date is the only course commensurate with the highway death and injury toll." S. Rep. No. 1301, 89th Cong., 2d Sess., 4 (1966).