Whitman v. American Trucking Assns., Inc.

The statutory instruction is health-based. For primary standards, EPA must identify the level of pollution that the public health can tolerate, provide an adequate margin of safety, and set a standard no more or less stringent than necessary to protect public health. Costs of attaining that standard are absent from this calculation.

The Act’s structure confirms that omission was deliberate. Congress expressly directed EPA to consider costs, feasibility, or available technology in numerous other Clean Air Act provisions. It would be implausible to infer authority to make the central NAAQS decision turn on costs from the comparatively modest terms “requisite” and “adequate margin of safety.” Congress does not hide an elephant in a mousehole.

Statutory provisions requiring EPA to collect or distribute cost information do not alter the result. States bear primary responsibility for selecting the means of compliance through state implementation plans, and cost data are useful in choosing efficient and feasible control measures at that implementation stage. Their relevance to implementation does not make costs relevant to setting the health-based national standard.

Because the statute has only one reasonable reading, the constitutional-avoidance canon cannot justify reading cost considerations into § 109(b). Courts may select among plausible constructions, but they may not revise an unambiguous statute to avoid a claimed constitutional difficulty.

The constitutional question concerns the limits Congress itself imposed, not whether EPA might voluntarily adopt additional limits. An agency cannot cure a standardless delegation by choosing to exercise only part of the authority Congress gave it, because choosing the limiting standard would itself entail the legislative judgment that Congress allegedly failed to make.

Section 109(b)(1) directs EPA to set uniform national standards for specified pollutants, based on published air-quality criteria reflecting the latest scientific knowledge, at levels requisite to protect public health with an adequate margin of safety. “Requisite” means sufficient, but not more than necessary. Those commands meaningfully confine EPA’s discretion.

The Court’s precedents permit considerably broader delegations. The Court had invalidated only two statutes under the nondelegation doctrine, while sustaining standards involving such concepts as the public interest, fair and equitable prices, and prevention of an imminent hazard to public safety. Congress need not specify exactly how much pollution-related harm is too much.

Although ozone and particulate matter can produce health effects along a continuum rather than at a single threshold, that fact does not invalidate the statute. Some judgment of degree is inherent in executive and judicial administration, and § 109(b)(1) fits comfortably within the discretion allowed by prior cases.

EPA’s interpretation marked the consummation of its decisionmaking process. After proposing an interim policy and receiving comments, EPA announced in the final ozone-rule preamble that it had reconsidered the matter and adopted a new interpretation of Part D. In later rulemakings, EPA treated that interpretation as conclusive rather than open to reconsideration.

The dispute was ripe because it presented a purely legal question of statutory interpretation, further factual development would add nothing, and review would not interfere with ongoing agency deliberation. The affected States also faced an immediate, expensive obligation to prepare implementation plans or risk federal control over implementation.

The implementation question was properly before the D.C. Circuit because it was embedded in the parties’ dispute over the final ozone standard. EPA defended its authority to issue the revised standard by arguing that Subpart 1 governed its implementation, so the court could assess that rationale when reviewing the final rule.

Under Chevron, the Court first asked whether Congress had clearly resolved the relationship between the two subparts. Subpart 1 generally applies to nonattainment areas, but its switching provisions yield when another Part D provision specifically provides for classification of ozone nonattainment areas. Subpart 2 plainly provides a classification table and attainment dates for ozone areas.

Some features of Subpart 2 fit poorly with a revised ozone standard. Its classifications use the old one-hour ozone measurements, its table leaves a potential gap for certain newly nonattaining areas, and its deadlines are measured from 1990. Those difficulties prevent the Court from holding that Subpart 2 is the exclusive and permanent implementation regime for every revised ozone standard.

But the identified gaps could at most support a limited role for Subpart 1; they could not justify EPA’s conclusion that Subpart 2 ceased to apply altogether upon revision of the ozone standard. Subpart 2 deliberately curbed EPA’s discretion by prescribing classifications, deadlines, and implementation requirements that Subpart 1 otherwise left more flexible.

EPA’s reading also conflicted with Subpart 2’s structure and purpose. Congress enacted provisions scheduled to take effect years after 1990, while knowing EPA periodically revises NAAQS. It is implausible that Congress meant EPA to nullify that detailed, long-term ozone program simply by revising the standard. EPA therefore had to develop a different reasonable interpretation on remand.