In the face of congressional gridlock preventing the passage of comprehensive climate change legislation, the Environmental Protection Agency (EPA) has pursued climate change regulation under existing authority provided by the Clean Air Act1 (CAA). But the Act’s regulatory requirements apply imperfectly to the chemical, physical, and political nature of greenhouse gases.
Last Term, in Utility Air Regulatory Group v. EPA2 (UARG), the Supreme Court held that under the CAA, stationary sources’ emissions of greenhouse gases (GHGs) alone could not trigger either the Act’s Prevention of Significant Deterioration (PSD) program or Title V permitting requirements.3 However, the Court held that the CAA does allow the application of the PSD program’s “best available control technology” (BACT) requirement to the emission of GHGs from those sources that emit sufficient quantities of other pollutants that they would be subject to the PSD program “anyway.”4 The practical effect of the Court’s decision is limited: the EPA maintains regulatory authority over all but three percent of the stationary sources it had found regulable.5 But the Court’s decision is significant for its restrictions on political judgment in agency decisionmaking.
The UARG Court’s holding that emissions of GHGs cannot trigger the PSD program both relies on and reinforces a flawed conclusion at the heart of Massachusetts v. EPA6: that being an “airborne compound of whatever stripe”7 is sufficient to qualify a substance as an “air pollutant.”8 This unnecessarily broad language in Massachusetts, which failed to recognize that the EPA must make a fundamentally political decision when it determines whether an airborne compound is a “pollutant,” created the central problem for the EPA in UARG and may continue to affect the EPA’s regulation of GHGs.
The facts giving rise to UARG began in 2007, when the Massachusetts Court held that the EPA had the authority under Title II of the CAA to regulate GHGs from new motor vehicles if the agency formed a “judgment” that emissions of GHGs contribute to air pollution that may endanger public health or welfare.9 The Court rejected the EPA’s argument that GHGs were not “pollutants” for purposes of the CAA, because the CAA’s act-wide definition of “air pollutant” “embraces all airborne compounds of whatever stripe.”10 Further, the Court held that the EPA was required to make a “reasoned judgment as to whether GHGs contribute to global warming.”11
Two years later, the EPA made that “reasoned judgment” in an Endangerment Finding, determining that GHGs endanger public health and welfare by contributing to global climate change.12 That Endangerment Finding then triggered the EPA’s obligation to regulate emissions of GHGs from new motor vehicles.13 In May 2010, in the “Tailpipe Rule,” the EPA promulgated GHG emission standards for new motor vehicles.14 The EPA previously had explained that, under its understanding of the CAA,15 regulation of GHG emissions from new motor vehicles would trigger broader regulation of GHG emissions: GHGs would become pollutants “subject to regulation” and, as a result, the PSD program and Title V permitting requirements would apply to many other sources emitting GHGs.16
The PSD program imposes emission limits on “major emitting facilities” located in areas in “attainment” for one or more of the National Ambient Air Quality Standards (NAAQS).17 A “major emitting facility” is one that emits (or has the potential to emit) statutorily defined quantities of “any air pollutant”18 — the critical language at issue in UARG. The program is intended to “ensure that the air quality in attainment areas or areas that are already ‘clean’ will not degrade.”19 Thus, in those clean areas, a “major emitting facility” may not be built or modified unless it uses “the best available control technology for each pollutant subject to regulation.”20
Because GHGs are emitted in far greater quantities than other pollutants, straightforward application of the PSD program to sources based on their emissions of GHGs would have produced “overwhelming permitting burdens” for both the EPA and regulated sources.21 The EPA therefore “tailored” its application of the PSD program, first applying the PSD requirements only to “anyway” sources and then to sources with the potential to emit 100,000 tons of carbon dioxide per year.22
Several states and industry groups petitioned for review of the Endangerment Finding and Tailpipe, Triggering, and Tailoring Rules in the D.C. Circuit. A unanimous panel dismissed all of these challenges to the EPA’s actions.23
On the question ultimately presented to the Supreme Court, the panel upheld the EPA’s interpretation of the scope of the PSD program.24 The EPA argued that “any air pollutant” in the definition of a “major emitting facility” could be interpreted only as “any air pollutant regulated under the CAA.”25 The court agreed: it found that “any regulated air pollutant”26 was “the only logical reading.”27 Moreover, it reasoned that the EPA’s interpretation “harmonize[d]”28 the scope of the PSD program with the requirements of BACT, which apply to “each pollutant subject to regulation under [the CAA].”29 The court rejected all three alternative interpretations of “any air pollutant” offered by petitioners to limit the PSD program’s scope.30
The D.C. Circuit denied petitions for rehearing en banc,31 with Judges Brown and Kavanaugh each dissenting. Judge Brown, refusing to “go quietly,” called on the “Court or Congress” to “restore order to the CAA” by overturning Massachusetts.32 Judge Kavanaugh disagreed with the panel’s conclusion that the statutory text unambiguously compelled the EPA’s interpretation. Rather, he thought “any air pollutant” could bear “two plausible interpretations,”33 including a NAAQS-only interpretation that avoided the “absurd results” caused by the EPA’s interpretation.34 Judge Kavanaugh nevertheless agreed that the PSD’s BACT program could apply to non-NAAQS pollutants “subject to regulation” because BACT, “by its terms,” includes non-NAAQS pollutants.35
The Supreme Court reversed in part and affirmed in part.36 Writing for the Court, Justice Scalia37 found that the CAA neither compelled nor permitted the EPA’s interpretation of the scope of the PSD program and Title V.38 Nevertheless, the Court found that the EPA reasonably interpreted the Act to cover sources subject to the PSA program “anyway” (because of their emission of NAAQS pollutants), and thus the Court approved of applying the BACT requirement to these “anyway” sources.39
Justice Scalia began by confronting the EPA’s core contention: that the CAA mandates regulation of GHG emitters under the PSD program. The EPA’s interpretation — and the D.C. Circuit’s holding — were wrong, Justice Scalia explained, because they relied on a “flawed syllogism.”40 The reasoning of the syllogism was straightforward: “air pollutant” in the Act-wide definition includes GHGs, as Massachusetts held; the PSD program requires permits for emitters of “any air pollutant”; thus, the PSD program must require permits for emitters of GHGs.41 The error arose, Justice Scalia reasoned, because the “conclusion follows from the premises only if the air pollutants referred to” in the CAA’s definitional section and the PSD program’s trigger “are the same.”42 Such an interpretation is “obviously untenable”43 because the Act-wide definition, “embrac[ing] all airborne compounds of whatever stripe,” cannot be applied in the Act’s operative provisions.44 Those operative provisions could not apply in each instance to any airborne compound of any kind because — as the EPA itself recognized45 — “air pollutant” must be read in context.46
Justice Scalia then moved on to explain that the EPA’s interpretation could not be upheld as reasonable under Chevron. The EPA’s interpretation was an impermissible reading of the statutory text47 because it was inconsistent with the statute’s structure and purpose.48 And this conflict with the statute’s purpose could not be resolved by “tailoring” the EPA’s application of the PSD program: the Tailoring Rule, because it was itself invalid, was unable to “validate the [EPA’s] interpretation of the triggering provisions.”49 Justice Scalia characterized the Tailoring Rule as the “EPA’s rewriting of the statutory thresholds,” an exercise of agency discretion that went far beyond “the interstices created by statutory silence or ambiguity.”50
Nevertheless, the Court held that the EPA could apply the PSD’s BACT requirements to the emission of GHGs from “anyway” sources.51 BACT’s mandate to reduce emissions of “each pollutant subject to regulation,”52 Justice Scalia reasoned, is more explicit and “far less open-ended” than the phrase “any air pollutant” in the text of the PSD trigger.53 The language of the BACT requirement cannot “bear a narrowing construction.”54
Justice Breyer concurred in part and dissented in part. He agreed with the majority that “any air pollutant” could “admit of unwritten limitations and exceptions.”55 But Justice Breyer disagreed with the majority as to the nature of those exceptions. The Court, he argued, read “any air pollutant” to include an implicit exception for a type of pollutant — GHGs.56 Instead, Justice Breyer would have read an exception for a type of source — those so small that regulation at the statutory threshold would be contrary to Congress’s intent.57 His exception would serve the CAA’s purpose “while going no further,” Justice Breyer argued.58
Justice Alito also concurred in part and dissented in part.59 Concurring in the Court’s holding that the PSD program could not apply to sources based on their emission of GHGs alone, Justice Alito reiterated his view that Massachusetts was wrongly decided and explained why “these cases further expose the flaws with that decision.”60 Justice Alito also emphasized the ways in which regulation of GHGs under BACT is in tension with the PSD program’s emphasis on local harms and BACT’s required case-by-case balancing of expected harms and benefits from control technology.61
The practical importance of the Court’s decision is limited — far more limited than might have been expected given the initial scope of the litigation when it began in the D.C. Circuit. The case is significant, however, for the way in which it interprets and relies on Massachusetts’s discussion of what constitutes a pollutant. The Massachusetts Court misguidedly attempted to define “air pollutant” in purely scientific and textual terms, in keeping with an opinion distrustful of the political motivations at the heart of the dispute.62 But the Massachusetts Court’s aversion to politics produced a definition of “air pollutant” that failed to recognize that political judgment is inescapable when labeling something a “pollutant.” And this error in Massachusetts lies at the heart of the majority’s reasoning in UARG: the “flaw” in the D.C. Circuit’s “syllogism” is based entirely on the mistaken language in Massachusetts.
The CAA defines “air pollutant” as “any air pollution agent or combination of such agents, including any physical, chemical, biological, radioactive . . . substance or matter which is emitted into or otherwise enters the ambient air.”63 In Massachusetts, the Court held that this definition “embraces airborne compounds of whatever stripe,” a category that unambiguously includes GHGs.64 But the Act-wide definition cannot be as broad as Justice Stevens’s majority opinion suggested. As Justice Scalia argued in dissent in Massachusetts, the all-encompassing phrase “any physical or chemical . . . substance” ought to be read in conjunction with the limiting phrase “air pollution agent.”65 Thus, only those “airborne compounds” that are also “pollutants” satisfy the Act-wide definition. Justice Scalia’s dissent contended that the majority’s interpretation “pretend[ed] . . . half of the definition does not exist.”66 He would have upheld as reasonable (at Chevron Step Two67) the EPA’s GHG-excluding definition of “air pollutant.”68 Justice Scalia was right to fault the majority for ignoring “air pollutant.” But his reasoning — like the majority’s — failed to account for the inherently political judgment necessary in defining a “pollutant” and thus failed to recognize that such a definition was unreasonable.
Because the CAA’s definition of “air pollutant” leaves “air pollution agent” undefined, the latter phrase should be given “its ordinary meaning.”69 The ordinary meaning of “pollution” is “a substance that makes land, water, air, etc., dirty and not safe or suitable to use.”70 Thus, an “air pollutant” is an “agent or combination of . . . agents” that “makes . . . air . . . dirty and not safe or suitable to use.” Whether a substance makes the air unsuitable for use requires a judgment similar, but not identical, to the § 7521 “endangerment judgment.”71 This “air pollution judgment” compels the EPA to consider an antecedent question: whether the substance makes the air dirty, unsafe, or unsuitable for use.72 Thus the EPA is instructed to exercise judgment as to its own jurisdiction73: the agency must decide whether an “airborne compound” is a pollutant at all, before considering whether it ought to be regulated under particular operative provisions.
Whether a substance ultimately renders the air unsuitable for use, such that it falls under the Act-wide definition of air pollutant, depends on choices among uses and thus requires balancing costs and benefits — a fundamentally political exercise.74 Science will get the EPA only so far in, say, determining whether the benefits of increased agricultural productivity in certain areas outweigh the loss of coastline in others.75 Rather, the EPA must reconcile these incommensurable values through the exercise of its own reasoned discretion, within the bounds set by Congress. The CAA’s bar on consideration of economic costs in determining whether a pollutant is harmful simplifies the EPA’s analysis somewhat,76 but it also limits the EPA’s recourse to cost-benefit analysis as a tool for reconciling competing values.77 Reconciling incommensurable values to define a “pollutant” necessarily requires political judgment.78
Once the EPA has determined that a substance is an air pollutant under the Act-wide definition, the agency must further determine whether it is an air pollutant for the purposes of a particular operative provision. This determination requires a more specific judgment: the precise way in which the pollutant renders the air unsuitable for a particular use. In some areas, the CAA defines for what uses the air is to be rendered suitable: the Act’s visibility section, for example, makes clear that the relevant “use” is the enjoyment of visibility and that “suitable” air is that which does not impair visibility.79 But elsewhere, the Act simply refers to “health and welfare,”80 and explicitly recognizes that “welfare” is a capacious term encompassing a slew of public interests.81 In those contexts, what a particular “use” ought to be and what air quality conditions are “suitable” for that use cannot be decided in the abstract or “objectively.” Rather, those decisions require context-specific judgments, sometimes weighing incommensurable values.
Accounting for the judgment inherent in defining a pollutant, the Massachusetts Court should have instructed the EPA to determine whether GHGs are air pollutants at all under the Act-wide definition — that is, whether GHGs make the air dirty, unsafe, or unsuitable for any use.82 If the EPA found that GHGs do make the air dirty, unsafe, or unsuitable, then GHGs would, by definition, be “air pollution agents” and, by the terms of the statute, therefore be “air pollutants.” The Court, by attempting to define “air pollutant” itself, and thereby circumvent the political judgment of the EPA, established a precedent that failed to recognize the discretionary judgment inherent in defining a “pollutant.”
Recognizing the political nature of defining a “pollutant” undermines the Court’s core holding in UARG: that the Act-wide definition and its context-specific application cannot be reconciled.83 Rather, the two definitions can be reconciled — as different stages in the exercise of judgment. The Act-wide definition should be understood as “any airborne compound which, in the expert judgment of the agency conferred with the discretion to decide, renders the air unsuitable for any use.”84 Such a definition makes explicit the role of political judgment inherent in the concept of “pollutant,” which necessarily includes judgment as to what constitutes “harm.”85
The operative-provision definition should then be understood as “an airborne compound that, in the expert judgment of the agency conferred with the discretion to decide, renders the air unsuitable for the particular statutorily relevant use.” Reading these two definitions together thus makes clear that the Act-wide definition of “air pollutant” and the PSD-specific definition do not need to refer to “the same” pollutants for the PSD program to apply to GHGs, as Justice Scalia argued.86 Rather, the two definitions refer to different stages in the exercise of expert judgment: first, in the Act-wide definition, determining whether GHGs render the air unsuitable for any uses, and second, in the PSD-specific definition, whether GHGs render the air unsuitable for the uses protected by the PSD program.
The EPA effectively adopted a heuristic for this second expert judgment, establishing in its 1978 Triggering Rule that any regulated pollutant would qualify as an airborne compound that renders the air unsuitable for the purposes of the PSD program.87 Moreover, the EPA specifically determined that GHGs render the air unsuitable for the particular purposes of the PSD program.88 It was this judgment that should have been the central issue in UARG, and there were legitimate reasons to doubt the EPA’s judgment on this question.89 Focus on this judgment would have mirrored the approach the Court should have taken in Massachusetts: interrogating the basis for the EPA’s judgment.
The Court, however, did not address the EPA’s judgment that GHGs render the air unsuitable for the PSD program, focusing instead on the practical impact of regulating sources based on their GHG emissions under the PSD program.90 There may be good reasons to be distrustful of politically motivated decisionmaking in federal agencies,91 but not all “political” decisions are alike. In some contexts, as here, the terms of the decision entrusted to an agency are inherently political. Courts seeking to police such political judgments are engaged in an ultimately quixotic pursuit.