Motor Vehicle Mfrs. Assn. of United States, Inc. v. State Farm Mut. Automobile Ins. Co.

The Safety Act expressly makes APA procedures and judicial review applicable to orders "establishing, amending, or revoking" a federal motor-vehicle-safety standard. Nothing in either the Act or the APA supports treating repeal of an existing rule as though the agency had merely declined to issue a new rule.

Rescinding a rule differs materially from initial inaction. An existing rule reflects the agency's informed judgment about how best to carry out Congress's policies. Although agencies may adapt to new circumstances and are not required to preserve regulations forever, they must explain why abandoning their prior judgment is justified.

The familiar APA test remains narrow: a court may not substitute its policy judgment for the agency's. But the agency must examine the relevant data, articulate a satisfactory explanation, and show a rational connection between the facts it found and the choice it made. A rule is arbitrary and capricious when, among other defects, the agency entirely fails to consider an important aspect of the problem or offers an explanation contrary to the evidence before it.

The D.C. Circuit erred by requiring NHTSA to give increasingly clear and convincing reasons based on Congress's prior reactions to Standard 208. The Supreme Court has not held that subsequent legislative events enlarge or diminish the APA standard governing review of agency action.

The cited events were, in any event, too ambiguous to establish a binding congressional commitment to mandatory passive restraints. Congress's ban on ignition interlocks did not affirmatively require passive restraints; Congress's failure to veto the 1977 rule could reflect deference to agency expertise; and unsuccessful legislative proposals do not reliably show Congress's settled intent.

Even an unequivocal congressional endorsement of the 1977 rule, short of incorporating it into statute, would not settle whether rescission was reasonable in 1981. That later agency decision still had to be judged under the ordinary arbitrary-and-capricious standard on the rulemaking record.

NHTSA had continued to recognize that airbag technology was sound, demonstrated, and effective. Its new concerns centered on detachable automatic belts, which drivers could disconnect, not on airbags. Thus, the perceived weakness of one compliance option did not itself justify eliminating the entire standard.

An airbags-only rule was not a remote or newly invented policy alternative. Airbags were an express compliance method under the existing Standard 208 and had originally been the principal form of passive protection contemplated by the agency. At minimum, the agency had to address this direct means of fulfilling the statute's safety purpose and give adequate reasons if it chose to reject it.

The automobile industry's preference for detachable belts did not excuse NHTSA's failure to consider airbags. The statute was designed in part to require safety measures beyond the industry's voluntary choices and could permissibly be technology-forcing. Courts also could not accept lawyers' post hoc explanations for why an airbag mandate might have been undesirable; the agency itself had supplied no explanation at all.

This conclusion did not impose additional rulemaking procedures forbidden by Vermont Yankee. The Court did not require NHTSA to consider every imaginable alternative. It held only that the agency could not entirely disregard an effective technology already within the existing standard's framework before repealing that standard.

NHTSA was entitled to make policy judgments under uncertainty. It reasonably could decline to impose, or could repeal, a safety measure if serious uncertainty about its effectiveness was supported by the record and rationally explained. The absence of direct evidence supporting a prediction would not automatically invalidate the agency's action.

But NHTSA's explanation did not adequately connect its evidence to its conclusion that detachable belts could not reliably produce even a five-percentage-point increase in usage. The agency could discount field studies of automatic belts because those studies involved ignition interlocks and self-selected purchasers. Yet discounting those studies established only that the precise nationwide effect was uncertain; it did not itself demonstrate that benefits would be negligible.

The agency failed to grapple with a crucial feature of detachable automatic belts. Once a driver reattached a disconnected belt, it would continue to operate automatically until detached again. Because many people failed to use manual belts due to inconvenience or forgetfulness rather than opposition to belts, this feature gave reason to think that occasional belt users might use restraints more often. NHTSA had to bring its expertise to bear on that point rather than ignore it.

The agency's conclusions about the rule's costs and its feared adverse public reaction depended on its unsupported premise that detachable belts would produce negligible benefits. On remand, NHTSA therefore had to reconsider the cost-benefit judgment in light of a reasoned assessment of likely belt usage. The Court also stressed that, under the Safety Act, safety was Congress's paramount consideration, although practicability, technology, costs, and lead time remained relevant.

NHTSA also failed to give a reasoned explanation for rejecting nondetachable continuous belts. It grouped them with ignition interlocks as "use-compelling" features, even though a continuous belt does not prevent a vehicle from operating and the record indicated that its emergency-release mechanism permitted safe exit. The agency could change its earlier favorable view of continuous belts, but it had to explain why.