298 F.3d 814, affirmed.
[ Ginsburg ]
[ Kennedy ]
HTML version
PDF version
HTML version
PDF version
HTML version
PDF version

Kennedy, J., dissenting


No. 02—658



[January 21, 2004]

Justice Kennedy, with whom The Chief Justice, Justice Scalia, and Justice Thomas join, dissenting.

The majority, in my respectful view, rests its holding on mistaken premises, for its reasoning conflicts with the express language of the Clean Air Act (CAA or Act), with sound rules of administrative law, and with principles that preserve the integrity of States in our federal system. The State of Alaska had in place procedures that were in full compliance with the governing statute and accompanying regulations promulgated by the Environmental Protection Agency (EPA). As I understand the opinion of the Court and the parties’ submissions, there is no disagreement on this point. Alaska followed these procedures to determine the best available control technology (BACT). EPA, however, sought to overturn the State’s decision, not by the process of judicial review, but by administrative fiat. The Court errs, in my judgment, by failing to hold that EPA, based on nothing more than its substantive disagreement with the State’s discretionary judgment, exceeded its powers in setting aside Alaska’s BACT determination.


As the majority explains, the case begins with §§113(a)(5) and 167 of the Act. 42 U.S.C. § 7413(a)(5), 7477. These provisions give EPA authority to enforce “requirements” of the CAA. The meaning of the word “requiremen[t],” though, is not defined in these provisions. Other provisions of the Act must be consulted. All parties agree that the requirement in this case is the “preconstruction requiremen[t]” that a “major emitting facility” be “subject to the best available technology [BACT] for each pollutant subject to regulation under this chapter emitted from, or which results from, such facility.” §7475(a)(4). BACT, in turn, is defined as

“an emission limitation based on the maximum degree of reduction of each pollutant subject to regulation under this chapter emitted from or which results from any major emitting facility, which the permitting authority, on a case-by-case basis, taking into account energy, environmental, and economic impacts and other costs, determines is achievable for such facility through application of production processes and available methods, systems, and techniques … .” §7479(3).

The majority holds that, under the CAA, state agencies are vested with “initial responsibility for identifying BACT in line with the Act’s definition of that term” and that EPA has a “broad oversight role” to ensure that a State’s BACT determination is “reasonably moored to the Act’s provisions.” Ante, at 18—19. The statute, however, contemplates no such arrangement. It directs the “permitting authority”–here, the Alaska Department of Environmental Conservation (ADEC)–to “determine” what constitutes BACT. To “determine” is not simply to make an initial recommendation that can later be overturned. It is “[t]o decide or settle … conclusively and authoritatively.” American Heritage Dictionary 495 (4th ed. 2000). Cf. 5 U.S.C. § 554 (“to be determined on the record after opportunity for an agency hearing”).

The BACT definition presumes that the permitting authority will exercise discretion. It presumes, in addition, that the BACT decision will accord full consideration to the statutory factors and other relevant and necessary criteria. Contrary to the majority’s holding, the statute does not direct the State to find as BACT the technology that results in the “maximum reduction of a pollutant achievable for [a] facility” in the abstract. Ante, at 19 (internal quotation marks omitted). Indeed, for a State to do so without regard to the other mandatory criteria would be to ignore the words of the statute. The Act requires a more comprehensive judgment. It provides that the permitting authority must “tak[e] into account” a set of contextual considerations–“energy, environmental, and economic impacts and other costs”–to identify the best control technology “on a case-by-case basis.” 42 U.S.C. § 7479(3). The majority reaches its narrow view of the scope of the State’s discretion only by wresting two adjectives, “maximum” and “achievable,” out of context. In doing so, it ignores “the cardinal rule that a statute is to be read as a whole.” King v. St. Vincent’s Hospital, 502 U.S. 215, 221 (1991).

To be sure, §§113(a)(5) and 167 authorize EPA to enforce requirements of the Act. These provisions, however, do not limit the States’ latitude and responsibility to balance all the statutory factors in making their discretionary judgments. If a State has complied with the Act’s requirements, §§113(a)(5) and 167 are not implicated and can supply no separate basis for EPA to exercise a supervisory role over a State’s discretionary decision. The Court of Appeals for the Ninth Circuit had it altogether backwards when it reasoned that, “because neither Section 113(a)(5) nor Section 167 contains any exemption for requirements that involve the state’s exercise of discretion,” EPA had the authority to issue orders countermanding the State’s BACT determination. 298 F.3d 814, 820 (2002). The question is not whether the two sections contain any exemption. Rather, it is about the nature of the Act’s requirements and whether EPA has the authority to set aside a BACT determination when no requirement of the Act was violated in the first place. In affirming the judgment of the Court of Appeals, the majority repeats the same analytical error. See ante, at 24 (“We fail to see why Congress, having expressly endorsed an expansive surveillance role for EPA in two independent CAA provisions, would then implicitly preclude the Agency from verifying substantive compliance with [BACT] …”). When the statute is read as a whole, it is clear that the CAA commits BACT determinations to the discretion of the relevant permitting authorities. Unless an objecting party, including EPA, prevails on judicial review, the determinations are conclusive.

Here the state agency, ADEC, recognized it was required to make a BACT determination. It issued two detailed reports in response to comments by interested parties and concluded that Low Nitrogen Oxide (NOx) was BACT. The requirement that the agency weigh the list of statutory factors, study all other relevant considerations, and decide the technology that can best reduce pollution within practical constraints was met in full. As even EPA acknowledged, ADEC “provid[ed] a detailed accounting of the process.” App. 286. This is not a case, then, where the state agency failed to have a BACT review procedure in place or altogether refused to apply the statute’s formal requirements. EPA’s only quarrel is with ADEC’s substantive conclusion. In disagreeing with ADEC, EPA’s sole contention, in the section of its order titled “Findings of Fact,” is that “SCR is BACT.” App. to Pet. for Cert. 30a, 34a (emphasis added). In addition, EPA does not allege that using Low NOx would violate other CAA requirements, such as the National Ambient Air Quality Standards, Alaska’s Prevention of Significant Deterioration (PSD) increments, or other applicable emission standards, see 42 U.S.C. § 7475(a)(3). On this state of the record there is no deviation from any statutory “requirement.” As a result, EPA has no statutory basis to invoke the enforcement authority of §§113(a)(5) and 167.

When Congress intends to give EPA general supervisory authority, it says so in clear terms. In addition to requiring EPA’s advance approval of BACT determinations in some instances, 42 U.S.C. § 7475(a)(8), the statute grants EPA powers to block the construction or operation of polluting sources in circumstances not at issue here, §§7426(b), (c)(1), 7410(a)(2)(D)(i). Outside the context of the CAA, Congress likewise knows how to establish federal oversight in unambiguous language. See, e.g., 42 U.S.C. § 1396a(a)(13)(A) (1994 ed.) (requiring, under the Medicaid Act, reimbursement according to rates that a “State finds, and makes assurances satisfactory to the Secretary [of Health and Human Services], are reasonable and adequate to meet the costs which must be incurred by efficiently and economically operated facilities”); Wilder v. Virginia Hospital Assn., 496 U.S. 498 (1990). No analogous language is used in the statutory definition of BACT.

EPA insists it needs oversight authority to prevent a “race to the bottom,” where jurisdictions compete with each other to lower environmental standards to attract new industries and keep existing businesses within their borders. Whatever the merits of these arguments as a general matter, EPA’s distrust of state agencies is inconsistent with the Act’s clear mandate that States bear the primary role in controlling pollution and, here, the exclusive role in making BACT determinations. In “cho[osing] not to dictate a Federal response to balancing sometimes conflicting goals” at the expense of “[m]aximum flexibility and State discretion,” H. R. Rep. No. 95—294, p. 146 (1977), Congress made the overriding judgment that States are more responsive to local conditions and can strike the right balance between preserving environmental quality and advancing competing objectives. By assigning certain functions to the States, Congress assumed they would have a stake in implementing the environmental objectives of the Act. At the same time, Congress charged EPA with setting ambient standards and enforcing emission limits, 42 U.S.C. § 7475(a)(3), to ensure that the Nation takes the necessary steps to reduce air pollution.

The presumption that state agencies are not to be trusted to do their part is unwarranted in another respect: EPA itself said so. As EPA concedes, States, by and large, take their statutory responsibility seriously, and EPA sees no reason to intervene in the vast majority of cases. Brief for Respondents 30, n. 9; 57 Fed. Reg. 28095 (1992) (“States have been largely successful in [‘administering and enforcing the various components of the PSD program’], and EPA’s involvement in interpretative and enforcement issues is limited …”). In light of this concession, EPA and amici not only fail to overcome the established presumption that States act in good faith, see Alden v. Maine, 527 U.S. 706, 755 (1999) (“We are unwilling to assume the States will refuse to honor … or obey the binding laws of the United States”), but also admit that their fears about a race to the bottom bear little relation to the real-world experience under the statute. See ante, at 36 (“We see no reason not to take EPA at its word”).


The statute contains safeguards to correct arbitrary and capricious BACT decisions when they do occur. Before EPA approves a State’s PSD permit program that allows a state agency to make BACT determinations, EPA must be satisfied that the State provides “an opportunity for state judicial review.” 61 Fed. Reg. 1882 (1996). Furthermore, before an individual permit may issue, the State must allow all “interested persons,” including “representatives of the [EPA] Administrator,” to submit comments on, among other things, “control technology requirements.” 42 U.S.C. § 7475(a)(2). To facilitate EPA’s participation in the State’s public comment process, the statute further provides that specific procedures be followed to inform the EPA Administrator of “every action” taken in the course of the permit approval process. §7475(d) (“Each State shall transmit to the Administrator a copy of each permit application relating to a major emitting facility received by such State and provide notice to the Administrator of every action related to the consideration of such permit”). Any person who participated in the comment process can pursue an administrative appeal of the State’s decision, followed, as mentioned, by judicial review in state courts.

EPA followed none of the normal procedures here. Only after the period for public comments expired did it intervene and seek to overturn Alaska’s decision that Low NOx was BACT. To justify its decision to opt out of the State’s administrative and judicial review process and, instead, to issue a unilateral order after everyone had spoken, EPA complains that it has not before intervened in “any State administrative review proceedings in State courts” and should not now be forced to do so. Tr. of Oral Arg. 35. With scant analysis, the majority agrees. Ante, at 26 (“It would be unusual, to say the least, for Congress to remit a federal agency enforcing federal law solely to state court. We decline to read such an uncommon regime into the Act’s silence”). The problem, of course, is that it is all the more unusual to allow a federal agency to take unilateral action to set aside a State’s administrative decision.

Despite EPA’s protestations, the statute makes explicit provision for EPA to challenge a state agency’s BACT determination in state proceedings. The statute requires States to set up an administrative process for “interested persons” to submit comments. §7475(a)(2). “[I]nterested persons,” Congress took care to note, include “representatives of the [EPA] Administrator.” Ibid.; see also Alaska Stat. §46.14.990(20) (2002) (defining “person” to include “an agency of the United States”). Given that EPA itself requires, as a condition of approving a State’s PSD program, that this process culminate in judicial review in state courts, 61 Fed. Reg., at 1882, it follows that EPA, a subset of all “interested persons,” must take the same procedural steps and cannot evade the more painstaking state process by a mere stroke of the pen under the agency’s letterhead.

On a more fundamental level, EPA and the majority confuse a substantive environmental statute like the CAA with a general administrative law statute like the Administrative Procedure Act (APA). EPA, the federal agency charged only with the CAA’s implementation, has no roving commission to ferret out arbitrary and capricious conduct by state agencies under the state equivalent of the APA. That task is left to state courts. See Idaho v. Coeur d’Alene Tribe of Idaho, 521 U.S. 261, 276 (1997) (“[T]he elaboration of administrative law … is one of the primary responsibilities of the state judiciary”).

Like federal courts, state courts are charged with reviewing agency actions to ensure that they comport with principles of rationality and due process. See, e.g., 5 U.S.C. § 706(2)(A); Alaska Stat. §44.62.570(b)(3) (2002). Counsel for respondents were unable to identify, either in their briefs or at oral argument, a single State that “does not have in its law the requirement that its own agencies … act rationally.” Tr. of Oral Arg. 30. Although it remains an open question whether EPA can bypass the state judiciary and go directly into federal district court under 28 U.S.C. § 1345 the availability of state judicial review defeats the Government’s argument that, absent EPA’s oversight, there is a legal vacuum where BACT decisions are not subject to review.

Requiring EPA to seek administrative and judicial review of a State’s BACT determination, instead of allowing it to be overturned by fiat, avoids the anomaly of shifting the burden of pleading and of initiating litigation from EPA to the State. Whether the BACT decision is reviewed in state court, or in federal district court if that option is available, see supra, at 8, EPA, as petitioner, bears the initial burden and costs of filing a petition for review alleging that the State acted arbitrarily. Under the scheme endorsed by the majority today, the tables are turned. Once EPA has issued an enforcement order, and the State seeks to invalidate that order, the State bears the burden of alleging that EPA acted arbitrarily. EPA and the majority concede that, because States enjoy substantial discretion in making BACT determinations, courts reviewing EPA’s order must ask not simply whether EPA acted arbitrarily but the convoluted question whether EPA acted arbitrarily in finding the State acted arbitrarily. Even under this unwieldy standard of review, and even if the burdens of persuasion and production remain with EPA, see ante, at 27—28, the initial burden of pleading and litigation now belongs to the State.

To make its decision more palatable, the majority holds that EPA still bears the burdens of production and persuasion, but there is little authority for this. The Court purports to rely on McCormick on Evidence for the proposition that “ ‘looking for the burden of pleading is not a foolproof guide to the allocation of burdens of proof.’ Ante, at 28, n. 17 (quoting 2 J. Strong, McCormick on Evidence §337, pp. 411—412 (5th ed. 1999)). The example–affirmative defense–discussed in that passage of the treatise, however, is far afield from the issues raised in this case. In fact, the treatise instructs that “[i]n most cases, the party who has the burden of pleading a fact will have the burdens of producing evidence and of persuading the jury of its existence as well.” Id., at 411. This is because “[t]he burdens of pleading and proof with regard to most facts have been and should be assigned to the plaintiff who generally seeks to change the present state of affairs and who therefore naturally should be expected to bear the risk of failure of proof or persuasion.” Id., at 412. In this case, EPA changed the status quo ante by issuing an order invalidating ADEC’s decision. Without upsetting accepted evidentiary principles, the majority cannot explain why EPA, as respondent in federal court–as opposed to the State, as petitioner alleging that EPA’s fait accompli was arbitrary–should bear the burdens of persuasion and production, or how this unusual reallocation of burdens should work in practice.

In any event, even the majority accepts that, under its reading of the statute, the State now bears the burden of pleading. With this burden-shifting benefit alone, EPA is most unlikely to follow the procedure, prescribed by federal law, of participating in the State’s administrative process and seeking judicial review in state courts. Instead, EPA can simply issue a unilateral order invalidating the State’s BACT determination and put the burden on the State to challenge EPA’s order. This end run around the State’s process is sure to undermine it. Unless Congress was on a fool’s errand, the loophole the majority finds goes only to demonstrate the inconsistency between its approach and the statutory scheme.

There is a further, and serious, flaw in the Court’s ruling. Suppose, before EPA issued its orders setting aside the State’s BACT determination, an Alaska state court had reviewed the matter and found no error of law or abuse of discretion in ADEC’s determination. The majority’s interpretation of the statute would allow EPA to intervene at this point for the first time, announce that ADEC’s determination is unreasoned under the CAA, and issue its own orders nullifying the state court’s ruling. This reworking of the balance between State and Federal Governments, not to mention the reallocation of authority between the Executive and Judicial Branches, shows the implausibility of the majority’s reasoning.

If a federal agency were to exercise an analogous power to review the decisions of federal courts, the arrangement would violate the well-established rule that the judgments of Article III courts cannot be revised by the Executive or Legislative Branches. See Hayburn’s Case, 2 Dall. 409, 410, n. (1792) (“ ‘[B]y the Constitution, neither the Secretary [of] War, nor any other Executive officer, nor even the Legislature, are authorized to sit as a court of errors on … judicial acts or opinions …’ ”); see also Plaut v. Spendthrift Farm, Inc., 514 U.S. 211 (1995). The principle that judicial decisions cannot be reopened at the whim of the Executive or the Legislature is essential to preserving separation of powers and judicial independence. Judges cannot, without sacrificing the autonomy of their office, put onto the scales of justice some predictive judgment about the probability that an administrator might reverse their rulings.

The Court today denies state judicial systems the same judicial independence it has long guarded for itself–only that the injury here is worse. Under the majority’s holding, decisions by state courts would be subject to being overturned, not just by any agency, but by an agency established by a different sovereign. We should be reluctant to interpret a congressional statute to deny to States the judicial independence guaranteed by their own constitutions. See Buckalew v. Holloway, 604 P.2d 240, 245 (Alaska 1979) (“There is no doubt that judicial independence was a paramount concern of the delegates [to the Alaska Constitutional Convention]”); see also, e.g., Cal. Const., Art. III, §3 (“The powers of state government are legislative, executive, and judicial. Persons charged with the exercise of one power may not exercise either of the others except as permitted by this Constitution”); see also 7 B. Witkin, Summary of California Law 159—160 (9th ed. 1988) (“[Under] the principle of separation of powers … , one [department] cannot exercise or interfere with the functions of either of the others”). The Federal Government is free, within its vast legislative authority, to impose federal standards. For States to have a role, however, their own governing processes must be respected. New York v. United States, 505 U.S. 144 (1992). If, by some course of reasoning, state courts must live with the insult that their judgments can be revised by a federal agency, the Court should at least insist upon a clear instruction from Congress. That directive cannot be found here. Cf. Gregory v. Ashcroft, 501 U.S. 452, 460 (1991) (“[I]f Congress intends to alter the usual constitutional balance between the States and the Federal Government, it must make its intention to do so unmistakably clear in the language of the statute” (internal quotation marks omitted)).

There is a final deficiency in the scheme the majority finds in the statute. Nothing in the Court’s analysis prevents EPA from issuing an order setting aside a BACT determination months, or even years, later. Congress cannot have intended this result. After all, when Congress provides for EPA’s involvement, it directs the agency to act sooner rather than later by establishing a pre-authorization procedure. 42 U.S.C. § 7475(a)(8). The majority misses the point when it faults ADEC for “overlook[ing] the obvious difference between a statutory requirement … and a statutory authorization.” Ante, at 25 (emphasis deleted). ADEC does not overlook the difference between approval before the fact and oversight after the fact. Rather, ADEC, unlike the majority, recognizes that the Act’s explicit provision for a preauthorization process underscores the need for finality in state permitting decisions, making implausible an interpretation of the statute that would allow a post hoc veto procedure that upsets the same reliance and expectation interests.

The majority’s initial response that “[t]his case threatens no such development [because] [i]t involves preconstruction orders issued by EPA … , not postconstruction federal Agency directives,” ante, at 29, provides no assurance that the logic of its reasoning would not in the future allow EPA’s belated interventions. When the majority confronts the problem, it concludes that “EPA, we are confident, could not indulge in the inequitable conduct ADEC and the dissent hypothesize while the federal courts sit to review EPA’s actions.” Ibid. The authority it cites for this proposition, however, consists of nothing more than a religious exemption case that is far removed from the issues presented here and a dissent from a case that has been overruled in part. Ibid. State agencies rely on this dictum at their own risk.

The majority’s reassurance to the States will likely be to no avail. “The principle that the United States are not bound by any statute of limitations, nor barred by any laches of their officers, however gross, in a suit brought by them as a sovereign Government to enforce a public right, or to assert a public interest, is established past all controversy or doubt.” United States v. Beebe, 127 U.S. 338, 344 (1888); see also United States v. Summerlin, 310 U.S. 414, 416 (1940) (“It is well settled that the United States is not bound by state statutes of limitation or subject to the defense of laches in enforcing its rights”); Utah Power & Light Co. v. United States, 243 U.S. 389, 409 (1917) (“[L]aches or neglect of duty on the part of officers of the Government is no defense to a suit by it to enforce a public right or protect a public interest … . A suit by the United States to enforce and maintain its policy … stands upon a different plane in this and some other respects from the ordinary private suit …”). Section 167, moreover, is mandatory. Once a violation of a statutory “requirement” is found, “[t]he Administrator shall … take such measures, including issuance of an order, or seeking injunctive relief, as necessary to prevent the construction or modification of a major emitting facility which does not conform to the requirements of this part … .” 42 U.S.C. § 7477. In short, EPA’s enforcement authority can–indeed, must–be exercised at any point. In light of our precedents a court would be hard pressed to hold otherwise.

The majority seeks to limit the consequence of its holding by quoting the response by respondents’ counsel at oral argument that ADEC could “absolutely” arrive at the same BACT determination if only it would pile on another layer of procedure and justify its decision on an “ ‘appropriate record.’ Ante, at 36 (quoting Tr. of Oral Arg. 35). As the Court of Appeals recognized in a prior case, however, this option gives no solace to the States:

The hardship is the process itself. Process costs money. If a federal licensee must spend years attempting to satisfy an elaborate, shifting array of state procedural requirements, then he must borrow a fortune to pay lawyers, economists, accountants, archaeologists, historians, engineers, recreational consultants, environmental consultants, biologists and others, with no revenue, no near-term prospect of revenue, and no certainty that there ever will be revenue. Meanwhile, politics, laws, interest rates, construction costs, and costs of alternatives change. Undue process may impose cost and uncertainty sufficient to thwart the federal determination that a power project should proceed.” Sayles Hydro Associates v. Maughan, 985 F.2d 451, 454 (CA9 1993).

If there is to be a second look, notwithstanding the 18 months ADEC spent analyzing BACT, a third or fourth look is just as permissible. The majority creates a sort of Zeno’s paradox for state agencies. Because there can always be an additional procedure to ensure that the preceding process was followed, no matter how many steps States take toward the objective, they may never reach it.

This is a most regrettable result. In the proper discharge of their responsibilities to implement the CAA in different conditions and localities nationwide, the States maintain permanent staffs within special agencies. These state employees, who no doubt take pride in their own resourcefulness, expertise, and commitment to the law, are the officials directed by Congress to make case-by-case, site-specific, determinations under the Act. Regulated persons and entities should be able to consult an agency staff with certainty and confidence, giving due consideration to agency recommendations and guidance. After today’s decision, however, a state agency can no longer represent itself as the real governing body. No matter how much time was spent in consultation and negotiation, a single federal administrator can in the end set all aside by a unilateral order. This is a great step backward in Congress’ design to grant States a significant stake in developing and enforcing national environmental objectives.

If EPA were to announce that permit applications subject to BACT review must be submitted to it in the first instance and can be forwarded to the State only with EPA’s advance approval, I should assume even the majority would find the basic structure of the BACT provisions undercut. In practical terms, however, the majority displaces state agencies, and degrades their role, in much the same way. In the case before us the applicant made elaborate submissions to ADEC. For over a year and a half, there ensued the constructive discourse that is the very object of the agency process, with both the ADEC staff and the applicant believing the State’s decision would be dispositive. EPA did not participate in the administrative process, but waited until after the record was closed to intervene by issuing an order setting aside the BACT determination.

We are advised that an applicant sometimes must spend up to $500,000 on the permit process and that, for a complex project, the time for approval can take from five to seven years. Brief for National Environmental Development Association et al. as Amici Curiae 8. Under the new multiple-tiered process, permit expenditures become less justified, state officials less credible, reliance less certain. The Court should be under no illusion that its decision respects the State’s administrative process.

The federal balance is remitted, in many instances, to Congress. Here the Court remits it to a single agency official. This is inconsistent with the assurance Congress gave to regulated entities when it allowed state agencies to decide upon the grant or denial of a permit under the BACT provisions of the CAA.


In the end EPA appears to realize the weakness of its arguments and asks us simply to defer to its expertise in light of the purported statutory ambiguity. See Brief for Respondents 41—43 (asking for deference under Chevron U.S. A. Inc. v. Natural Resources Defense Council, Inc., 467 U.S. 837 (1984)). To its credit, the majority holds Chevron deference inapplicable. Deference is inappropriate for all the reasons the majority recites, ante, at 21—22, plus one more: The statute is not in any way ambiguous. As a result, our inquiry should proceed no further.

Actions, however, speak louder than words, and the majority ends up giving EPA the very Chevron deference–and more–it says should be denied. The Court’s opinion is chock full of Chevron-like language. Compare 467 U.S., at 843 (“whether the agency’s answer is based on a permissible construction of the statute”); id., at 845 (“whether the Administrator’s view … is a reasonable one”), with ante, at 22 (“[EPA’s] arguments do not persuade us to reject [them] as impermissible”); ante, at 27 (“ That rational interpretation, we agree, is surely permissible”). So deficient are its statutory arguments that the majority must hide behind Chevron’s vocabulary, despite its explicit holding that Chevron does not apply. In applying Chevron de facto under these circumstances, however, the majority undermines the well-established distinction our precedents draw between Chevron and less deferential forms of judicial review.

The broader implication of today’s decision is more unfortunate still. The CAA is not the only statute that relies on a close and equal partnership between federal and state authorities to accomplish congressional objectives. See, e.g., New York v. United States, 505 U.S., at 167 (listing examples). Under the majority’s reasoning, these other statutes, too, could be said to confer on federal agencies ultimate decisionmaking authority, relegating States to the role of mere provinces or political corporations, instead of coequal sovereigns entitled to the same dignity and respect. Cf. Alden v. Maine, 527 U.S. 706 (1999). If cooperative federalism, Hodel v. Virginia Surface Mining & Reclamation Assn., Inc., 452 U.S. 264, 289 (1981), is to achieve Congress’ goal of allowing state governments to be accountable to the democratic process in implementing environmental policies, federal agencies cannot consign States to the ministerial tasks of information gathering and making initial recommendations, while reserving to themselves the authority to make final judgments under the guise of surveillance and oversight.

For these reasons, and with all respect, I dissent from the opinion and the judgment of the Court.