West Virginia v. EPA

597 U.S. 697

• Court: Supreme Court of the United States
• Decided: June 30, 2022
• Docket: 20-1530
• Status: Published

Opinion

(Slip Opinion) OCTOBER TERM, 2021 1

Syllabus

NOTE: Where it is feasible, a syllabus (headnote) will be released, as is being done in connection with this case, at the time the opinion is issued. The syllabus constitutes no part of the opinion of the Court but has been prepared by the Reporter of Decisions for the convenience of the reader. See United States v. Detroit Timber & Lumber Co., 200 U. S. 321, 337.

SUPREME COURT OF THE UNITED STATES

WEST VIRGINIA ET AL. v. ENVIRONMENTAL PROTECTION AGENCY ET AL.

CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE DISTRICT OF COLUMBIA CIRCUIT

No. 20–1530. Argued February 28, 2022—Decided June 30, 2022* In 2015, the Environmental Protection Agency (EPA) promulgated the Clean Power Plan rule, which addressed carbon dioxide emissions from existing coal- and natural-gas-fired power plants. For authority, the Agency cited Section 111 of the Clean Air Act, which, although known as the New Source Performance Standards program, also au- thorizes regulation of certain pollutants from existing sources under Section 111(d). 42 U. S. C. §7411(d). Prior to the Clean Power Plan, EPA had used Section 111(d) only a handful of times since its enact- ment in 1970. Under that provision, although the States set the actual enforceable rules governing existing sources (such as power plants), EPA determines the emissions limit with which they will have to com- ply. The Agency derives that limit by determining the “best system of emission reduction . . . that has been adequately demonstrated,” or the BSER, for the kind of existing source at issue. §7411(a)(1). The limit then reflects the amount of pollution reduction “achievable through the application of” that system. Ibid. In the Clean Power Plan, EPA determined that the BSER for exist- ing coal and natural gas plants included three types of measures, which the Agency called “building blocks.” 80 Fed. Reg. 64667. The first building block was “heat rate improvements” at coal-fired plants—essentially practices such plants could undertake to burn coal —————— * Together with No. 20–1531, North American Coal Corp. v. Environ- mental Protection Agency et al., No. 20–1778, Westmoreland Mining Holdings LLC v. Environmental Protection Agency et al., and No. 20– 1780, North Dakota v. Environmental Protection Agency et al., also on certiorari to the same court. 2 WEST VIRGINIA v. EPA

more cleanly. Id., at 64727. This sort of source-specific, efficiency- improving measure was similar in kind to those that EPA had previ- ously identified as the BSER in other Section 111 rules. Building blocks two and three were quite different, as both involved what EPA called “generation shifting” at the grid level—i.e., a shift in electricity production from higher-emitting to lower-emitting produc- ers. Building block two was a shift in generation from existing coal- fired power plants, which would make less power, to natural-gas-fired plants, which would make more. Ibid. This would reduce carbon di- oxide emissions because natural gas plants produce less carbon dioxide per unit of electricity generated than coal plants. Building block three worked like building block two, except that the shift was from both coal and gas plants to renewables, mostly wind and solar. Id., at 64729, 64748. The Agency explained that, to implement the needed shift in generation to cleaner sources, an operator could reduce the regulated plant’s own production of electricity, build or invest in a new or exist- ing natural gas plant, wind farm, or solar installation, or purchase emission allowances or credits as part of a cap-and-trade regime. Id., at 64731–64732. Taking any of these steps would implement a sector- wide shift in electricity production from coal to natural gas and renew- ables. Id., at 64731. Having decided that the BSER was one that would reduce carbon pollution mostly by moving production to cleaner sources, EPA then set about determining “the degree of emission limitation achievable through the application” of that system. §7411(a)(1). The Agency rec- ognized that, in translating the BSER into an operational emissions limit, it could choose whether to require anything from a little genera- tion shifting to a great deal. It settled on what it regarded as a “rea- sonable” amount of shift, which it based on modeling how much more electricity both natural gas and renewable sources could supply with- out causing undue cost increases or reducing the overall power supply. Id., at 64797–64811. The Agency ultimately projected, for instance, that it would be feasible to have coal provide 27% of national electricity generation by 2030, down from 38% in 2014. From these projected changes, EPA determined the applicable emissions performance rates, which were so strict that no existing coal plant would have been able to achieve them without engaging in one of the three means of gener- ation shifting. The Government projected that the rule would impose billions in compliance costs, raise retail electricity prices, require the retirement of dozens of coal plants, and eliminate tens of thousands of jobs. This Court stayed the Clean Power Plan in 2016, preventing the rule from taking effect. It was later repealed after a change in Presidential administrations. Specifically, in 2019, EPA found that the Clean Cite as: 597 U. S. ____ (2022) 3

Power Plan had exceeded the Agency’s statutory authority under Sec- tion 111(d), which it interpreted to “limit[ ] the BSER to those systems that can be put into operation at a building, structure, facility, or in- stallation.” 84 Fed. Reg. 32524. EPA explained that the Clean Power Plan, rather than setting the standard “based on the application of equipment and practices at the level of an individual facility,” had in- stead based it on “a shift in the energy generation mix at the grid level,” id., at 32523. The Agency determined that the interpretive question raised by the Clean Power Plan fell under the major questions doctrine. Under that doctrine, it determined, a clear statement is nec- essary for a court to conclude that Congress intended to delegate au- thority “of this breadth to regulate a fundamental sector of the econ- omy.” Id., at 32529. It found none. The Agency replaced the Clean Power Plan by promulgating a different Section 111(d) regulation, known as the Affordable Clean Energy (ACE) rule. Id., at 32532. In that rule, EPA determined that the BSER would be akin to building block one of the Clean Power Plan: a combination of equipment up- grades and operating practices that would improve facilities’ heat rates. Id., at 32522, 32537. A number of States and private parties filed petitions for review in the D. C. Circuit, challenging EPA’s repeal of the Clean Power Plan and its enactment of the replacement ACE rule. The Court of Appeals consolidated the cases and held that EPA’s “repeal of the Clean Power Plan rested critically on a mistaken reading of the Clean Air Act”— namely, that generation shifting cannot be a “system of emission re- duction” under Section 111. 985 F. 3d 914, 995. The court vacated the Agency’s repeal of the Clean Power Plan and remanded to the Agency for further consideration. It also vacated and remanded the ACE rule for the same reason. The court’s decision was followed by another change in Presidential administrations, and EPA moved the court to partially stay its mandate as to the Clean Power Plan while the Agency considered whether to promulgate a new Section 111(d) rule. No party opposed the motion, and the Court of Appeals agreed to stay its vaca- tur of the Agency’s repeal of the Clean Power Plan. Held: 1. This case remains justiciable notwithstanding the Government’s contention that no petitioner has Article III standing, given EPA’s stated intention not to enforce the Clean Power Plan and to instead engage in new rulemaking. In considering standing to appeal, the question is whether the appellant has experienced an injury “fairly traceable to the judgment below.” Food Marketing Institute v. Argus Leader Media, 588 U. S. ___, ___. If so, and a “favorable ruling” from the appellate court “would redress [that] injury,” then the appellant has a cognizable Article III stake. Ibid. Here, the judgment below 4 WEST VIRGINIA v. EPA

vacated the ACE rule and its embedded repeal of the Clean Power Plan, and accordingly purports to bring the Clean Power Plan back into legal effect. There is little question that the petitioner States are injured, since the rule requires them to more stringently regulate power plant emissions within their borders. The Government counters that EPA’s current posture has mooted the prior dispute. The distinc- tion between mootness and standing matters, however, because the Government bears the burden to establish that a once-live case has become moot. The Government’s argument in this case boils down to its representation that EPA does not intend to enforce the Clean Power Plan prior to promulgating a new Section 111(d) rule. But “voluntary cessation does not moot a case” unless it is “absolutely clear that the allegedly wrongful behavior could not reasonably be expected to recur.” Parents Involved in Community Schools v. Seattle School Dist. No. 1, 551 U. S. 701, 719. Here, the Government “nowhere suggests that if this litigation is resolved in its favor it will not” reimpose emissions limits predicated on generation shifting. Ibid. Pp. 14–16. 2. Congress did not grant EPA in Section 111(d) of the Clean Air Act the authority to devise emissions caps based on the generation shifting approach the Agency took in the Clean Power Plan. Pp. 16–31. (a) In devising emissions limits for power plants, EPA “deter- mines” the BSER that—taking into account cost, health, and other fac- tors—it finds “has been adequately demonstrated,” and then quanti- fies “the degree of emission limitation achievable” if that best system were applied to the covered source. §7411(a)(1). The issue here is whether restructuring the Nation’s overall mix of electricity genera- tion, to transition from 38% to 27% coal by 2030, can be the BSER within the meaning of Section 111. Precedent teaches that there are “extraordinary cases” in which the “history and the breadth of the authority that [the agency] has as- serted,” and the “economic and political significance” of that assertion, provide a “reason to hesitate before concluding that Congress” meant to confer such authority. FDA v. Brown & Williamson Tobacco Corp., 529 U. S. 120, 159–160. See, e.g., Alabama Assn. of Realtors v. Depart- ment of Health and Human Servs., 594 U. S. ___, ___; Utility Air Reg- ulatory Group v. EPA, 573 U. S. 302, 324; Gonzales v. Oregon, 546 U. S. 243, 267; National Federation of Independent Business v. OSHA, 595 U. S. ___, ___. Under this body of law, known as the major ques- tions doctrine, given both separation of powers principles and a prac- tical understanding of legislative intent, the agency must point to “clear congressional authorization” for the authority it claims. Utility Air, 573 U. S., at 324. Pp. 16–20. (b) This is a major questions case. EPA claimed to discover an Cite as: 597 U. S. ____ (2022) 5

unheralded power representing a transformative expansion of its reg- ulatory authority in the vague language of a long-extant, but rarely used, statute designed as a gap filler. That discovery allowed it to adopt a regulatory program that Congress had conspicuously declined to enact itself. Given these circumstances, there is every reason to “hesitate before concluding that Congress” meant to confer on EPA the authority it claims under Section 111(d). Brown & Williamson, 529 U. S., at 160. Prior to 2015, EPA had always set Section 111 emissions limits based on the application of measures that would reduce pollution by causing the regulated source to operate more cleanly, see, e.g., 41 Fed. Reg. 48706—never by looking to a “system” that would reduce pollu- tion simply by “shifting” polluting activity “from dirtier to cleaner sources.” 80 Fed. Reg. 64726. The Government quibbles with this his- tory, pointing to the 2005 Mercury Rule as one Section 111 rule that it says relied upon a cap-and-trade mechanism to reduce emissions. See 70 Fed. Reg. 28616. But in that regulation, EPA set the emissions limit—the “cap”—based on the use of “technologies [that could be] in- stalled and operational on a nationwide basis” in the relevant timeframe. Id., at 28620–28621. By contrast, and by design, there are no particular controls a coal plant operator can install and operate to attain the emissions limits established by the Clean Power Plan. In- deed, the Agency nodded to the novelty of its approach when it ex- plained that it was pursuing a “broader, forward-thinking approach to the design” of Section 111 regulations that would “improve the overall power system,” rather than the emissions performance of individual sources, by forcing a shift throughout the power grid from one type of energy source to another. 80 Fed. Reg. 64703 (emphasis added). This view of EPA’s authority was not only unprecedented; it also effected a “fundamental revision of the statute, changing it from [one sort of] scheme of . . . regulation” into an entirely different kind. MCI Tele- communications Corp. v. American Telephone & Telegraph Co., 512 U. S. 218, 231. The Government attempts to downplay matters, noting that the Agency must limit the magnitude of generation shift it demands to a level that will not be “exorbitantly costly” or “threaten the reliability of the grid.” Brief for Federal Respondents 42. This argument does not limit the breadth of EPA’s claimed authority so much as reveal it: On EPA’s view of Section 111(d), Congress implicitly tasked it, and it alone, with balancing the many vital considerations of national policy implicated in the basic regulation of how Americans get their energy. There is little reason to think Congress did so. EPA has admitted that issues of electricity transmission, distribution, and storage are not within its traditional expertise. And this Court doubts that “Congress 6 WEST VIRGINIA v. EPA

. . . intended to delegate . . . decision[s] of such economic and political significance,” i.e., how much coal-based generation there should be over the coming decades, to any administrative agency. Brown & Wil- liamson, 529 U. S., at 160. Nor can the Court ignore that the regula- tory writ EPA newly uncovered in Section 111(d) conveniently enabled it to enact a program, namely, cap-and-trade for carbon, that Congress had already considered and rejected numerous times. The importance of the policy issue and ongoing debate over its merits “makes the oblique form of the claimed delegation all the more suspect.” Gonzales, 546 U. S., at 267–268. Pp. 20–28. (c) Given that precedent counsels skepticism toward EPA’s claim that Section 111 empowers it to devise carbon emissions caps based on a generation shifting approach, the Government must point to “clear congressional authorization” to regulate in that manner. Utility Air, 573 U. S., at 324. The Government can offer only EPA’s authority to establish emissions caps at a level reflecting “the application of the best system of emission reduction . . . adequately demonstrated.” §7411(a)(1). The word “system” shorn of all context, however, is an empty vessel. Such a vague statutory grant is not close to the sort of clear authorization required. The Government points to other provi- sions of the Clean Air Act—specifically the Acid Rain and National Ambient Air Quality Standards (NAAQS) programs—that use the word “system” or “similar words” to describe sector-wide mechanisms for reducing pollution. But just because a cap-and-trade “system” can be used to reduce emissions does not mean that it is the kind of “system of emission reduction” referred to in Section 111. Finally, the Court has no occasion to decide whether the statutory phrase “system of emission reduction” refers exclusively to measures that improve the pollution performance of individual sources, such that all other actions are ineligible to qualify as the BSER. It is perti- nent to the Court’s analysis that EPA has acted consistent with such a limitation for four decades. But the only question before the Court is more narrow: whether the “best system of emission reduction” iden- tified by EPA in the Clean Power Plan was within the authority granted to the Agency in Section 111(d) of the Clean Air Act. For the reasons given, the answer is no. Pp. 28–31. 985 F. 3d 914, reversed and remanded.

ROBERTS, C. J., delivered the opinion of the Court, in which THOMAS, ALITO, GORSUCH, KAVANAUGH, and BARRETT, JJ., joined. GORSUCH, J., filed a concurring opinion, in which ALITO, J., joined. KAGAN, J., filed a dissenting opinion, in which BREYER and SOTOMAYOR, JJ., joined. Cite as: 597 U. S. ____ (2022) 1

Opinion of the Court

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Wash- ington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES _________________

Nos. 20–1530, 20–1531, 20–1778 and 20–1780 _________________

WEST VIRGINIA, ET AL., PETITIONERS 20–1530 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

THE NORTH AMERICAN COAL CORPORATION, PETITIONER 20–1531 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

WESTMORELAND MINING HOLDINGS LLC, PETITIONER 20–1778 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

NORTH DAKOTA, PETITIONER 20–1780 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL. ON WRITS OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE DISTRICT OF COLUMBIA CIRCUIT [June 30, 2022]

CHIEF JUSTICE ROBERTS delivered the opinion of the Court. The Clean Air Act authorizes the Environmental Protec- tion Agency to regulate power plants by setting a “standard of performance” for their emission of certain pollutants into 2 WEST VIRGINIA v. EPA

the air. 84 Stat. 1683, 42 U. S. C. §7411(a)(1). That stand- ard may be different for new and existing plants, but in each case it must reflect the “best system of emission reduc- tion” that the Agency has determined to be “adequately demonstrated” for the particular category. §§7411(a)(1), (b)(1), (d). For existing plants, the States then implement that requirement by issuing rules restricting emissions from sources within their borders. Since passage of the Act 50 years ago, EPA has exercised this authority by setting performance standards based on measures that would reduce pollution by causing plants to operate more cleanly. In 2015, however, EPA issued a new rule concluding that the “best system of emission reduction” for existing coal-fired power plants included a requirement that such facilities reduce their own production of electric- ity, or subsidize increased generation by natural gas, wind, or solar sources. The question before us is whether this broader conception of EPA’s authority is within the power granted to it by the Clean Air Act. I A The Clean Air Act establishes three main regulatory pro- grams to control air pollution from stationary sources such as power plants. Clean Air Amendments of 1970, 84 Stat. 1676, 42 U. S. C. §7401 et seq. One program is the New Source Performance Standards program of Section 111, at issue here. The other two are the National Ambient Air Quality Standards (NAAQS) program, set out in Sections 108 through 110 of the Act, 42 U. S. C. §§7408–7410, and the Hazardous Air Pollutants (HAP) program, set out in Section 112, §7412. To understand the place and function of Section 111 in the statutory scheme, some background on the other two programs is in order. The NAAQS program addresses air pollutants that “may Cite as: 597 U. S. ____ (2022) 3

reasonably be anticipated to endanger public health or wel- fare,” and “the presence of which in the ambient air results from numerous or diverse mobile or stationary sources.” §7408(a)(1). After identifying such pollutants, EPA estab- lishes a NAAQS for each. The NAAQS represents “the max- imum airborne concentration of [the] pollutant that the public health can tolerate.” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 465 (2001); see §7409(b). EPA, though, does not choose which sources must reduce their pollution and by how much to meet the ambient pollution target. Instead, Section 110 of the Act leaves that task in the first instance to the States, requiring each “to submit to [EPA] a plan designed to implement and maintain such standards within its boundaries.” Train v. Natural Re- sources Defense Council, Inc., 421 U. S. 60, 65 (1975); §7410. The second major program governing stationary sources is the HAP program. The HAP program primarily targets pollutants, other than those already covered by a NAAQS, that present “a threat of adverse human health effects,” in- cluding substances known or anticipated to be “carcino- genic, mutagenic, teratogenic, neurotoxic,” or otherwise “acutely or chronically toxic.” §7412(b)(2). EPA’s regulatory role with respect to these toxic pollu- tants is different in kind from its role in administering the NAAQS program. There, EPA is generally limited to deter- mining the maximum safe amount of covered pollutants in the air. As to each hazardous pollutant, by contrast, the Agency must promulgate emissions standards for both new and existing major sources. §7412(d)(1). Those standards must “require the maximum degree of reduction in emis- sions . . . that the [EPA] Administrator, taking into consid- eration the cost of achieving such emission reduction, and any non-air quality health and environmental impacts and energy requirements, determines is achievable . . . through application of measures, processes, methods, systems or techniques” of emission reduction. §7412(d)(2). In other 4 WEST VIRGINIA v. EPA

words, EPA must directly require all covered sources to re- duce their emissions to a certain level. And it chooses that level by determining the “maximum degree of reduction” it considers “achievable” in practice by using the best existing technologies and methods. §7412(d)(3). Thus, in the parlance of environmental law, Section 112 directs the Agency to impose “technology-based standard[s] for hazardous emissions,” Alaska Dept. of Environmental Conservation v. EPA, 540 U. S. 461, 485, n. 12 (2004) (em- phasis added). This sort of “ ‘technology-based’ approach fo- cuses upon the control technologies that are available to in- dustrial entities and requires the agency to . . . ensur[e] that regulated firms adopt the appropriate cleanup technol- ogy.” T. McGarity, Media-Quality, Technology, and Cost- Benefit Balancing Strategies for Health and Environmen- tal Regulation, 46 Law & Contemp. Prob. 159, 160 (Summer 1983) (McGarity). (Such “technologies” are not limited to literal technology, such as scrubbers; “changes in the design and operation” of the facility, or “in the way that employees perform their tasks,” are also available options. Id., at 163, n. 18.) The third air pollution control scheme is the New Source Performance Standards program of Section 111. §7411. That section directs EPA to list “categories of stationary sources” that it determines “cause[], or contribute[ ] signifi- cantly to, air pollution which may reasonably be anticipated to endanger public health or welfare.” §7411(b)(1)(A). Un- der Section 111(b), the Agency must then promulgate for each category “Federal standards of performance for new sources,” §7411(b)(1)(B). A “standard of performance” is one that “reflects the degree of emission limitation achievable through the application of the best system of emission reduction which (taking into account the cost of achiev- ing such reduction and any nonair quality health and Cite as: 597 U. S. ____ (2022) 5

environmental impact and energy requirements) the [EPA] Administrator determines has been adequately demonstrated.” §7411(a)(1). Thus, the statute directs EPA to (1) “determine[ ],” taking into account various factors, the “best system of emission reduction which . . . has been adequately demonstrated,” (2) ascertain the “degree of emission limitation achievable through the application” of that system, and (3) impose an emissions limit on new stationary sources that “reflects” that amount. Ibid.; see also 80 Fed. Reg. 64538 (2015). Generally speaking, a source may achieve that emissions cap any way it chooses; the key is that its pollution be no more than the amount “achievable through the application of the best system of emission reduction . . . adequately demonstrated,” or the BSER. §7411(a)(1); see §7411(b)(5). EPA undertakes this analysis on a pollutant-by-pollutant basis, establishing different standards of performance with respect to different pollutants emitted from the same source category. See, e.g., 73 Fed. Reg. 35838 (2008); 42 Fed. Reg. 22510 (1977). Although the thrust of Section 111 focuses on emissions limits for new and modified sources—as its title indicates— the statute also authorizes regulation of certain pollutants from existing sources. Under Section 111(d), once EPA “has set new source standards addressing emissions of a partic- ular pollutant under . . . section 111(b),” 80 Fed. Reg. 64711, it must then address emissions of that same pollutant by existing sources—but only if they are not already regulated under the NAAQS or HAP programs. §7411(d)(1). Existing power plants, for example, emit many pollutants covered by a NAAQS or HAP standard. Section 111(d) thus “operates as a gap-filler,” empowering EPA to regulate harmful emis- sions not already controlled under the Agency’s other au- thorities. American Lung Assn. v. EPA, 985 F. 3d 914, 932 (CADC 2021). 6 WEST VIRGINIA v. EPA

Although the States set the actual rules governing exist- ing power plants, EPA itself still retains the primary regu- latory role in Section 111(d). The Agency, not the States, decides the amount of pollution reduction that must ulti- mately be achieved. It does so by again determining, as when setting the new source rules, “the best system of emis- sion reduction . . . that has been adequately demonstrated for [existing covered] facilities.” 40 CFR §60.22(b)(5) (2021); see also 80 Fed. Reg. 64664, and n. 1. The States then sub- mit plans containing the emissions restrictions that they intend to adopt and enforce in order not to exceed the per- missible level of pollution established by EPA. See §§60.23, 60.24; 42 U. S. C. §7411(d)(1). Reflecting the ancillary nature of Section 111(d), EPA has used it only a handful of times since the enactment of the statute in 1970. See 80 Fed. Reg. 64703, and n. 275 (past regulations pertained to “four pollutants from five source categories”). For instance, the Agency has established emissions limits on acid mist from sulfuric acid production, 41 Fed. Reg. 48706 (1976) (identifying “fiber mist elimina- tor” technology as BSER); sulfide gases released by kraft pulp mills, 44 Fed. Reg. 29829 (1979) (determining BSER to be a combination of scrubbers, incineration, filtration systems, and temperature control); and emissions of vari- ous harmful gases from municipal landfills, 61 Fed. Reg. 9907 (1996) (setting BSER as use of a flare to combust the gases). It was thus only a slight overstatement for one of the architects of the 1990 amendments to the Clean Air Act to refer to Section 111(d) as an “obscure, never-used section of the law.” Hearings on S. 300 et al. before the Subcom- mittee on Environmental Protection of the Senate Commit- tee on Environment and Public Works, 100th Cong., 1st Sess., 13 (1987) (remarks of Sen. Durenberger). B Things changed in October 2015, when EPA promulgated Cite as: 597 U. S. ____ (2022) 7

two rules addressing carbon dioxide pollution from power plants—one for new plants under Section 111(b), the other for existing plants under Section 111(d). Both were prem- ised on the Agency’s earlier finding that carbon dioxide is an “air pollutant” that “may reasonably be anticipated to endanger public health or welfare” by causing climate change. 80 Fed. Reg. 64530. Carbon dioxide is not subject to a NAAQS and has not been listed as a toxic pollutant. The first rule announced by EPA established federal car- bon emissions limits for new power plants of two varieties: fossil-fuel-fired electric steam generating units (mostly coal fired) and natural-gas-fired stationary combustion tur- bines. Id., at 64512. Following the statutory process set out above, the Agency determined the BSER for the two cat- egories of sources. For steam generating units, for instance, EPA determined that the BSER was a combination of high- efficiency production processes and carbon capture technol- ogy. See 80 Fed. Reg. 64512. EPA then set the emissions limit based on the amount of carbon dioxide that a plant would emit with these technologies in place. Id., at 64513. The second rule was triggered by the first: Because EPA was now regulating carbon dioxide from new coal and gas plants, Section 111(d) required EPA to also address carbon emissions from existing coal and gas plants. See §7411(d)(1). It did so through what it called the Clean Power Plan rule. In that rule, EPA established “final emission guidelines for states to follow in developing plans” to regulate existing power plants within their borders. Id., at 64662. To arrive at the guideline limits, EPA did the same thing it does when imposing federal regulations on new sources: It identified the BSER. The BSER that the Agency selected for existing coal-fired power plants, however, was quite different from the BSER it had chosen for new sources. The BSER for existing plants included three types of measures, which the Agency called 8 WEST VIRGINIA v. EPA

“building blocks.” Id., at 64667. The first building block was “heat rate improvements” at coal-fired plants—essen- tially practices such plants could undertake to burn coal more efficiently. Id., at 64727. But such improvements, EPA stated, would “lead to only small emission reductions,” because coal-fired power plants were already operating near optimum efficiency. Ibid. On the Agency’s view, “much larger emission reductions [were] needed from [coal- fired plants] to address climate change.” Ibid. So the Agency included two additional building blocks in its BSER, both of which involve what it called “generation shifting from higher-emitting to lower-emitting” producers of electricity. Id., at 64728. Building block two was a shift in electricity production from existing coal-fired power plants to natural-gas-fired plants. Ibid. Because natural gas plants produce “typically less than half as much” carbon dioxide per unit of electricity created as coal-fired plants, the Agency explained, “this generation shift [would] re- duce[ ] CO2 emissions.” Ibid. Building block three worked the same way, except that the shift was from both coal- and gas-fired plants to “new low- or zero-carbon generating ca- pacity,” mainly wind and solar. Id., at 64729, 64748. “Most of the CO2 controls” in the rule came from the application of building blocks two and three. Id., at 64728. The Agency identified three ways in which a regulated plant operator could implement a shift in generation to cleaner sources. Id., at 64731. First, an operator could simply reduce the regulated plant’s own production of elec- tricity. Second, it could build a new natural gas plant, wind farm, or solar installation, or invest in someone else’s exist- ing facility and then increase generation there. Ibid. Fi- nally, operators could purchase emission allowances or credits as part of a cap-and-trade regime. Id., at 64731– 64732. Under such a scheme, sources that achieve a reduc- tion in their emissions can sell a credit representing the value of that reduction to others, who are able to count it Cite as: 597 U. S. ____ (2022) 9

toward their own applicable emissions caps. EPA explained that taking any of these steps would im- plement a sector-wide shift in electricity production from coal to natural gas and renewables. Id., at 64731. Given the integrated nature of the power grid, “adding electricity to the grid from one generator will result in the instantane- ous reduction in generation from other generators,” and “re- ductions in generation from one generator lead to the in- stantaneous increase in generation” by others. Id., at 64769. So coal plants, whether by reducing their own pro- duction, subsidizing an increase in production by cleaner sources, or both, would cause a shift toward wind, solar, and natural gas. Having decided that the “best system of emission reduc- tion . . . adequately demonstrated” was one that would re- duce carbon pollution mostly by moving production to cleaner sources, EPA then set about determining “the de- gree of emission limitation achievable through the applica- tion” of that system. 42 U. S. C. §7411(a)(1). The Agency recognized that—given the nature of generation shifting— it could choose from “a wide range of potential stringencies for the BSER.” 80 Fed. Reg. 64730. Put differently, in translating the BSER into an operational emissions limit, EPA could choose whether to require anything from a little generation shifting to a great deal. The Agency settled on what it regarded as a “reasonable” amount of shift, which it based on modeling of how much more electricity both natu- ral gas and renewable sources could supply without causing undue cost increases or reducing the overall power supply. Id., at 64797–64811. Based on these changes, EPA pro- jected that by 2030, it would be feasible to have coal provide 27% of national electricity generation, down from 38% in 2014. Id., at 64665, 64694; see Dept. of Energy, U. S. En- ergy Information Admin., Monthly Energy Review (May 2015), Electricity Net Generation: Electric Power Sector, p. 106 (Table 7.2b). 10 WEST VIRGINIA v. EPA

From these significant projected reductions in genera- tion, EPA developed a series of complex equations to “deter- mine the emission performance rates” that States would be required to implement. 80 Fed. Reg. 64815. The calcula- tions resulted in numerical emissions ceilings so strict that no existing coal plant would have been able to achieve them without engaging in one of the three means of shifting gen- eration described above. Indeed, the emissions limit the Clean Power Plan established for existing power plants was actually stricter than the cap imposed by the simultane- ously published standards for new plants. Compare id., at 64742, with id., at 64513. The point, after all, was to compel the transfer of power generating capacity from existing sources to wind and solar. The White House stated that the Clean Power Plan would “drive a[n] . . . aggressive transformation in the domestic energy industry.” White House Fact Sheet, App. in Ameri- can Lung Assn. v. EPA, No. 19–1140 etc. (CADC), p. 2076. EPA’s own modeling concluded that the rule would entail billions of dollars in compliance costs (to be paid in the form of higher energy prices), require the retirement of dozens of coal-fired plants, and eliminate tens of thousands of jobs across various sectors. EPA, Regulatory Impact Analysis for the Clean Power Plan Final Rule 3–22, 3–30, 3–33, 6– 24, 6–25 (2015). The Energy Information Administration reached similar conclusions, projecting that the rule would cause retail electricity prices to remain persistently 10% higher in many States, and would reduce GDP by at least a trillion 2009 dollars by 2040. Dept. of Energy, Analysis of the Impacts of the Clean Power Plan 21, 63–64 (May 2015). C These projections were never tested, because the Clean Power Plan never went into effect. The same day that EPA promulgated the rule, dozens of parties (including 27 States) petitioned for review in the D. C. Circuit. After that Cite as: 597 U. S. ____ (2022) 11

court declined to enter a stay of the rule, the challengers sought the same relief from this Court. We granted a stay, preventing the rule from taking effect. West Virginia v. EPA, 577 U. S. 1126 (2016). The Court of Appeals later heard argument on the merits en banc. But before it could issue a decision, there was a change in Presidential admin- istrations. The new administration requested that the liti- gation be held in abeyance so that EPA could reconsider the Clean Power Plan. The D. C. Circuit obliged, and later dis- missed the petitions for review as moot. EPA eventually repealed the rule in 2019, concluding that the Clean Power Plan had been “in excess of its statu- tory authority” under Section 111(d). 84 Fed. Reg. 32523 (2019). Specifically, the Agency concluded that generation shifting should not have been considered as part of the BSER. The Agency interpreted Section 111 as “limit[ing] the BSER to those systems that can be put into operation at a building, structure, facility, or installation,” such as “add-on controls” and “inherently lower-emitting pro- cesses/practices/designs.” Id., at 32524. It then explained that the Clean Power Plan, rather than setting the stand- ard “based on the application of equipment and practices at the level of an individual facility,” had instead based it on “a shift in the energy generation mix at the grid level,” id., at 32523—not the sort of measure that has “a potential for application to an individual source.” Id., at 32524. The Agency determined that “the interpretative question raised” by the Clean Power Plan—“i.e., whether a ‘system of emission reduction’ can consist of generation-shifting measures”—fell under the “major question doctrine.” Id., at 32529. Under that doctrine, EPA explained, courts “ex- pect Congress to speak clearly if it wishes to assign to an agency decisions of vast economic and political signifi- cance.” Ibid. (quoting Utility Air Regulatory Group v. EPA, 573 U. S. 302, 324 (2014) (internal quotation marks omit- ted)). The Agency concluded that the Clean Power Plan was 12 WEST VIRGINIA v. EPA

such a decision, for a number of reasons. Its “generation- shifting scheme was projected to have billions of dollars of impact.” 84 Fed. Reg. 32529. “[N]o section 111 rule of the scores issued ha[d] ever been based on generation shifting.” Ibid. And that novel reading of the statute would empower EPA “to order the wholesale restructuring of any industrial sector” based only on its discretionary assessment of “such factors as ‘cost’ and ‘feasibility.’ ” Ibid. EPA argued that under the major questions doctrine, a clear statement was necessary to conclude that Congress intended to delegate authority “of this breadth to regulate a fundamental sector of the economy.” Ibid. It found none. “Indeed,” it concluded, given the text and structure of the statute, “Congress has directly spoken to this precise ques- tion and precluded” the use of measures such as generation shifting. Ibid. In the same rulemaking, the Agency replaced the Clean Power Plan by promulgating a different Section 111(d) reg- ulation, known as the Affordable Clean Energy (ACE) Rule. Id., at 32532. Based on its view of what measures may per- missibly make up the BSER, EPA determined that the best system would be akin to building block one of the Clean Power Plan: a combination of equipment upgrades and op- erating practices that would improve facilities’ heat rates. Id., at 32522, 32537. The ACE Rule determined that the application of its BSER measures would result in only small reductions in carbon dioxide emissions. Id., at 32561. D A number of States and private parties immediately filed petitions for review in the D. C. Circuit, challenging EPA’s repeal of the Clean Power Plan and its enactment of the re- placement ACE Rule. Other States and private entities— including petitioners here West Virginia, North Dakota, Westmoreland Mining Holdings LLC, and The North Amer- ican Coal Corporation (NACC)—intervened to defend both Cite as: 597 U. S. ____ (2022) 13

actions. The Court of Appeals consolidated all 12 petitions for re- view into one case. It then held that EPA’s “repeal of the Clean Power Plan rested critically on a mistaken reading of the Clean Air Act”—namely, that generation shifting can- not be a “system of emission reduction” under Section 111. 985 F. 3d, at 995. To the contrary, the court concluded, the statute could reasonably be read to encompass generation shifting. As part of that analysis, the Court of Appeals con- cluded that the major questions doctrine did not apply, and thus rejected the need for a clear statement of congressional intent to delegate such power to EPA. Id., at 959–968. Having found that EPA misunderstood the scope of its au- thority under the Clean Air Act, the Court vacated the Agency’s repeal of the Clean Power Plan and remanded to the Agency for further consideration. Id., at 995. It also vacated and remanded the replacement rule, the ACE Rule, for the same reason. Ibid. The court’s decision, handed down on January 19, 2021, was quickly followed by another change in Presidential ad- ministrations. One month later, EPA moved the Court of Appeals to partially stay the issuance of its mandate as it pertained to the Clean Power Plan. The Agency did so to ensure that the Clean Power Plan would not immediately go back into effect. Respondents’ Motion for a Partial Stay of Issuance of the Mandate in American Lung Assn. v. EPA, No. 19–1140 etc. (CADC), p. 4. EPA believed that such a result would not make sense while it was in the process of considering whether to promulgate a new Section 111(d) rule. Ibid. No party opposed the motion, and the court ac- cordingly stayed its vacatur of the Agency’s repeal of the Clean Power Plan. Westmoreland, NACC, and the States defending the re- peal of the Clean Power Plan all filed petitions for certio- rari. We granted the petitions and consolidated the cases. 595 U. S. ___ (2021). 14 WEST VIRGINIA v. EPA

II We first consider the Government’s contention that no pe- titioner has Article III standing to seek our review. Although most disputes over standing concern whether a plaintiff has satisfied the requirement when filing suit, “Ar- ticle III demands that an actual controversy persist throughout all stages of litigation.” Hollingsworth v. Perry, 570 U. S. 693, 705 (2013) (internal quotation marks omit- ted). The requirement of standing “must be met by persons seeking appellate review, just as it must be met by persons appearing in courts of first instance.” Arizonans for Official English v. Arizona, 520 U. S. 43, 64 (1997). In considering a litigant’s standing to appeal, the question is whether it has experienced an injury “fairly traceable to the judgment below.” Food Marketing Institute v. Argus Leader Media, 588 U. S. ___, ___ (2019) (slip op., at 4) (emphasis added; internal quotation marks omitted). If so, and a “favorable ruling” from the appellate court “would redress [that] in- jury,” then the appellant has a cognizable Article III stake. Ibid. Here, it is apparent that at least one group of petition- ers—the state petitioners—are injured by the Court of Ap- peals’ judgment. That judgment vacated “the ACE rule and its embedded repeal of the Clean Power Plan,” 985 F. 3d, at 995 (emphasis added), and accordingly purports to bring the Clean Power Plan back into legal effect. Thus, to the extent the Clean Power Plan harms the States, the D. C. Circuit’s judgment inflicts the same injury. And there can be “little question” that the rule does injure the States, since they are “the object of ” its requirement that they more stringently regulate power plant emissions within their borders. Lujan v. Defenders of Wildlife, 504 U. S. 555, 561– 562 (1992). The Government counters that “agency and judicial ac- tions” subsequent to the court’s entry of judgment have “eliminated any . . . possibility” of injury. Brief for Federal Cite as: 597 U. S. ____ (2022) 15

Respondents 16. First, after the decision, EPA informed the Court of Appeals that it does not intend to enforce the Clean Power Plan because it has decided to promulgate a new Section 111(d) rule. Second, on EPA’s request, the lower court stayed the part of its judgment that vacated the repeal, pending that new rulemaking. “These circum- stances,” says the Government, “have mooted the prior dis- pute as to the CPP Repeal Rule’s legality.” Id., at 17 (em- phasis added). That Freudian slip, however, reveals the basic flaw in the Government’s argument: It is the doctrine of mootness, not standing, that addresses whether “an intervening circum- stance [has] deprive[d] the plaintiff of a personal stake in the outcome of the lawsuit.” Genesis HealthCare Corp. v. Symczyk, 569 U. S. 66, 72 (2013) (internal quotation marks omitted); see also Friends of the Earth, Inc. v. Laidlaw En- vironmental Services (TOC), Inc., 528 U. S. 167, 189–192 (2000). The distinction matters because the Government, not petitioners, bears the burden to establish that a once- live case has become moot. Id., at 189; Adarand Construc- tors, Inc. v. Slater, 528 U. S. 216, 222 (2000) (per curiam). That burden is “heavy” where, as here, “[t]he only con- ceivable basis for a finding of mootness in th[e] case is [the respondent’s] voluntary conduct.” Friends of the Earth, 528 U. S., at 189. Although the Government briefly argues that the lower court’s stay of its mandate extinguished the con- troversy, it cites no authority for that proposition, and it does not make sense: Lower courts frequently stay their mandates when notified that the losing party intends to seek our certiorari review. So the Government’s mootness argument boils down to its representation that EPA has no intention of enforcing the Clean Power Plan prior to prom- ulgating a new Section 111(d) rule. But “voluntary cessation does not moot a case” unless it is “absolutely clear that the allegedly wrongful behavior 16 WEST VIRGINIA v. EPA

could not reasonably be expected to recur.” Parents In- volved in Community Schools v. Seattle School Dist. No. 1, 551 U. S. 701, 719 (2007). Here the Government “nowhere suggests that if this litigation is resolved in its favor it will not” reimpose emissions limits predicated on generation shifting; indeed, it “vigorously defends” the legality of such an approach. Ibid. We do not dismiss a case as moot in such circumstances. See City of Mesquite v. Aladdin’s Cas- tle, Inc., 455 U. S. 283, 288–289 (1982). The case thus re- mains justiciable, and we may turn to the merits. III A In devising emissions limits for power plants, EPA first “determines” the “best system of emission reduction” that— taking into account cost, health, and other factors—it finds “has been adequately demonstrated.” 42 U. S. C. §7411(a)(1). The Agency then quantifies “the degree of emission limitation achievable” if that best system were ap- plied to the covered source. Ibid.; see also 80 Fed. Reg. 64719. The BSER, therefore, “is the central determination that the EPA must make in formulating [its emission] guidelines” under Section 111. Id., at 64723. The issue here is whether restructuring the Nation’s overall mix of electricity generation, to transition from 38% coal to 27% coal by 2030, can be the “best system of emission reduction” within the meaning of Section 111. “It is a fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.” Davis v. Michigan Dept. of Treasury, 489 U. S. 803, 809 (1989). Where the statute at issue is one that confers au- thority upon an administrative agency, that inquiry must be “shaped, at least in some measure, by the nature of the question presented”—whether Congress in fact meant to confer the power the agency has asserted. FDA v. Brown & Cite as: 597 U. S. ____ (2022) 17

Williamson Tobacco Corp., 529 U. S. 120, 159 (2000). In the ordinary case, that context has no great effect on the appro- priate analysis. Nonetheless, our precedent teaches that there are “extraordinary cases” that call for a different ap- proach—cases in which the “history and the breadth of the authority that [the agency] has asserted,” and the “eco- nomic and political significance” of that assertion, provide a “reason to hesitate before concluding that Congress” meant to confer such authority. Id., at 159–160. Such cases have arisen from all corners of the adminis- trative state. In Brown & Williamson, for instance, the Food and Drug Administration claimed that its authority over “drugs” and “devices” included the power to regulate, and even ban, tobacco products. Id., at 126–127. We re- jected that “expansive construction of the statute,” conclud- ing that “Congress could not have intended to delegate” such a sweeping and consequential authority “in so cryptic a fashion.” Id., at 160. In Alabama Assn. of Realtors v. De- partment of Health and Human Servs., 594 U. S. ___, ___ (2021) (per curiam) (slip op., at 3), we concluded that the Centers for Disease Control and Prevention could not, un- der its authority to adopt measures “necessary to prevent the . . . spread of ” disease, institute a nationwide eviction moratorium in response to the COVID–19 pandemic. We found the statute’s language a “wafer-thin reed” on which to rest such a measure, given “the sheer scope of the CDC’s claimed authority,” its “unprecedented” nature, and the fact that Congress had failed to extend the moratorium after previously having done so. Id., at ___–___ (slip op., at 6–8). Our decision in Utility Air addressed another question re- garding EPA’s authority—namely, whether EPA could con- strue the term “air pollutant,” in a specific provision of the Clean Air Act, to cover greenhouse gases. 573 U. S., at 310. Despite its textual plausibility, we noted that the Agency’s interpretation would have given it permitting authority over millions of small sources, such as hotels and office 18 WEST VIRGINIA v. EPA

buildings, that had never before been subject to such re- quirements. Id., at 310, 324. We declined to uphold EPA’s claim of “unheralded” regulatory power over “a significant portion of the American economy.” Id., at 324. In Gonzales v. Oregon, 546 U. S. 243 (2006), we confronted the Attorney General’s assertion that he could rescind the license of any physician who prescribed a controlled substance for as- sisted suicide, even in a State where such action was legal. The Attorney General argued that this came within his statutory power to revoke licenses where he found them “in- consistent with the public interest,” 21 U. S. C. §823(f ). We considered the “idea that Congress gave [him] such broad and unusual authority through an implicit delegation . . . not sustainable.” 546 U. S., at 267. Similar considerations informed our recent decision invalidating the Occupational Safety and Health Administration’s mandate that “84 mil- lion Americans . . . either obtain a COVID–19 vaccine or un- dergo weekly medical testing at their own expense.” Na- tional Federation of Independent Business v. Occupational Safety and Health Administration, 595 U. S. ___, ___ (2022) (per curiam) (slip op., at 5). We found it “telling that OSHA, in its half century of existence,” had never relied on its au- thority to regulate occupational hazards to impose such a remarkable measure. Id., at ___ (slip op., at 8). All of these regulatory assertions had a colorable textual basis. And yet, in each case, given the various circum- stances, “common sense as to the manner in which Con- gress [would have been] likely to delegate” such power to the agency at issue, Brown & Williamson, 529 U. S., at 133, made it very unlikely that Congress had actually done so. Extraordinary grants of regulatory authority are rarely ac- complished through “modest words,” “vague terms,” or “subtle device[s].” Whitman, 531 U. S., at 468. Nor does Congress typically use oblique or elliptical language to em- power an agency to make a “radical or fundamental change” to a statutory scheme. MCI Telecommunications Corp. v. Cite as: 597 U. S. ____ (2022) 19

American Telephone & Telegraph Co., 512 U. S. 218, 229 (1994). Agencies have only those powers given to them by Congress, and “enabling legislation” is generally not an “open book to which the agency [may] add pages and change the plot line.” E. Gellhorn & P. Verkuil, Controlling Chevron- Based Delegations, 20 Cardozo L. Rev. 989, 1011 (1999). We presume that “Congress intends to make major policy decisions itself, not leave those decisions to agencies.” United States Telecom Assn. v. FCC, 855 F. 3d 381, 419 (CADC 2017) (Kavanaugh, J., dissenting from denial of re- hearing en banc). Thus, in certain extraordinary cases, both separation of powers principles and a practical understanding of legisla- tive intent make us “reluctant to read into ambiguous stat- utory text” the delegation claimed to be lurking there. Util- ity Air, 573 U. S., at 324. To convince us otherwise, something more than a merely plausible textual basis for the agency action is necessary. The agency instead must point to “clear congressional authorization” for the power it claims. Ibid. The dissent criticizes us for “announc[ing] the arrival” of this major questions doctrine, and argues that each of the decisions just cited simply followed our “ordinary method” of “normal statutory interpretation,” post, at 13, 15 (opinion of KAGAN, J.). But in what the dissent calls the “key case” in this area, Brown & Williamson, post, at 15, the Court could not have been clearer: “In extraordinary cases . . . there may be reason to hesitate” before accepting a reading of a statute that would, under more “ordinary” circum- stances, be upheld. 529 U. S., at 159. Or, as we put it more recently, we “typically greet” assertions of “extravagant statutory power over the national economy” with “skepti- cism.” Utility Air, 573 U. S., at 324. The dissent attempts to fit the analysis in these cases within routine statutory interpretation, but the bottom line—a requirement of “clear 20 WEST VIRGINIA v. EPA

congressional authorization,” ibid.—confirms that the ap- proach under the major questions doctrine is distinct. As for the major questions doctrine “label[ ],” post, at 13, it took hold because it refers to an identifiable body of law that has developed over a series of significant cases all ad- dressing a particular and recurring problem: agencies as- serting highly consequential power beyond what Congress could reasonably be understood to have granted. Scholars and jurists have recognized the common threads between those decisions. So have we. See Utility Air, 573 U. S., at 324 (citing Brown & Williamson and MCI); King v. Burwell, 576 U. S. 473, 486 (2015) (citing Utility Air, Brown & Wil- liamson, and Gonzales). B Under our precedents, this is a major questions case. In arguing that Section 111(d) empowers it to substantially re- structure the American energy market, EPA “claim[ed] to discover in a long-extant statute an unheralded power” rep- resenting a “transformative expansion in [its] regulatory authority.” Utility Air, 573 U. S., at 324. It located that newfound power in the vague language of an “ancillary pro- vision[ ]” of the Act, Whitman, 531 U. S., at 468, one that was designed to function as a gap filler and had rarely been used in the preceding decades. And the Agency’s discovery allowed it to adopt a regulatory program that Congress had conspicuously and repeatedly declined to enact itself. Brown & Williamson, 529 U. S., at 159–160; Gonzales, 546 U. S., at 267–268; Alabama Assn., 594 U. S., at ___, ___ (slip op., at 2, 8). Given these circumstances, there is every rea- son to “hesitate before concluding that Congress” meant to confer on EPA the authority it claims under Section 111(d). Brown & Williamson, 529 U. S., at 159–160. Prior to 2015, EPA had always set emissions limits under Section 111 based on the application of measures that would reduce pollution by causing the regulated source to Cite as: 597 U. S. ____ (2022) 21

operate more cleanly. See, e.g., 41 Fed. Reg. 48706 (requir- ing “degree of control achievable through the application of fiber mist eliminators”); see also supra, at 6. It had never devised a cap by looking to a “system” that would reduce pollution simply by “shifting” polluting activity “from dirt- ier to cleaner sources.” 80 Fed. Reg. 64726; see id., at 64738 (“[O]ur traditional interpretation . . . has allowed regulated entities to produce as much of a particular good as they de- sire provided that they do so through an appropriately clean (or low-emitting) process.”). And as Justice Frankfurter has noted, “just as established practice may shed light on the extent of power conveyed by general statutory lan- guage, so the want of assertion of power by those who pre- sumably would be alert to exercise it, is equally significant in determining whether such power was actually con- ferred.” FTC v. Bunte Brothers, Inc., 312 U. S. 349, 352 (1941). The Government quibbles with this description of the his- tory of Section 111(d), pointing to one rule that it says relied upon a cap-and-trade mechanism to reduce emissions. See 70 Fed. Reg. 28616 (2005) (Mercury Rule). The legality of that choice was controversial at the time and was never ad- dressed by a court. See New Jersey v. EPA, 517 F. 3d 574 (CADC 2008) (vacating on other grounds). Even assuming the Rule was valid, though, it still does not help the Gov- ernment. In that regulation, EPA set the actual “emission cap”—i.e., the limit on emissions that sources would be re- quired to meet—“based on the level of [mercury] emissions reductions that w[ould] be achievable by” the use of “tech- nologies [that could be] installed and operational on a na- tionwide basis” in the relevant timeframe—namely, wet scrubbers. 70 Fed. Reg. 28620–28621. In other words, EPA set the cap based on the application of particular controls, and regulated sources could have complied by installing them. By contrast, and by design, there is no control a coal 22 WEST VIRGINIA v. EPA

plant operator can deploy to attain the emissions limits es- tablished by the Clean Power Plan. See supra, at 10. The Mercury Rule, therefore, is no precedent for the Clean Power Plan. To the contrary, it was one more entry in an unbroken list of prior Section 111 rules that devised the en- forceable emissions limit by determining the best control mechanisms available for the source.1 This consistent understanding of “system[s] of emission reduction” tracked the seemingly universal view, as stated by EPA in its inaugural Section 111(d) rulemaking, that “Congress intended a technology-based approach” to regu- lation in that Section. 40 Fed. Reg. 53343 (1975); see id., at 53341 (“degree of control to be reflected in EPA’s emission guidelines” will be based on “application of best adequately demonstrated control technology”).2 A technology-based standard, recall, is one that focuses on improving the emis- sions performance of individual sources. EPA “commonly —————— 1 The dissent cites other ostensible precedents, see post, at 25–26, but

they are also inapposite. A few allowed cap-and-trade or similar averag- ing measures as compliance mechanisms, like the Mercury Rule. See, e.g., 60 Fed. Reg. 65402 (1995). The others were not Section 111 rules. 2 See McGarity 165 (EPA promulgates “technology-based new source

performance standards that require the implementation of the ‘best available demonstrated’ technology”); P. McCubbin, The Risk in Technology- Based Standards, 16 Duke Env. L. & Pol’y Forum 1, 46, n. 180 (2003) (Section 111 standards “are another set of technology-based standards”); W. Wagner, The Triumph of Technology-Based Standards, 2000 U. Ill. L. Rev 83, 84, n. 4 (“Technology-based standards made their initial ap- pearance” in “Section 111 of the Clean Air Act,” which “requires the EPA to set technology-based emission limitations”). The dissent points to a 1977 amendment to Section 111 as evidence that the 1970 Congress did not intend for EPA to establish this sort of source-specific standard. Post, at 10–11. But it is clear that the 1977 amendment was merely intended to prohibit power plants from adopting one specific kind of at-the-source measure—a switch from burning high- sulfur coal to low-sulfur coal—and was not intended or understood to change the basic, source-focused regulatory approach. See Wisconsin Elec. Power Co. v. Reilly, 893 F. 2d 901, 919 (CA8 1990) (explaining the history); B. Ackerman & W. Hassler, Clean Coal/Dirty Air (1981) (same). Cite as: 597 U. S. ____ (2022) 23

referred to” the “level of control” required as a “best demon- strated technology (BDT)” standard, 73 Fed. Reg. 34073, and consistently applied it as such. E.g., 61 Fed. Reg. 9907 (declaring “BDT” to be “a well-designed and well-operated gas collection system and . . . a control device capable of re- ducing [harmful gases] in the collected gas by 98 weight- percent.”). Indeed, EPA nodded to this history in the Clean Power Plan itself, describing the sort of “systems of emission re- duction” it had always before selected—“efficiency improve- ments, fuel-switching,” and “add-on controls”—as “more traditional air pollution control measures.” 80 Fed. Reg. 64784. The Agency noted that it had “considered” such measures as potential systems of emission reduction for carbon dioxide, ibid., including a measure it ultimately adopted as a “component” of the BSER, namely, heat rate improvements. Id., at 64727. But, the Agency explained, in order to “control[ ] CO2 from affected [plants] at levels . . . necessary to mitigate the dan- gers presented by climate change,” it could not base the emissions limit on “measures that improve efficiency at the power plants.” Id., at 64728. “The quantity of emissions reductions resulting from the application of these measures” would have been “too small.” Id., at 64727. In- stead, to attain the necessary “critical CO2 reductions,” EPA adopted what it called a “broader, forward-thinking approach to the design” of Section 111 regulations. Id., at 64703. Rather than focus on improving the performance of individual sources, it would “improve the overall power sys- tem by lowering the carbon intensity of power generation.” Ibid. (emphasis added). And it would do that by forcing a shift throughout the power grid from one type of energy source to another. In the words of the then-EPA Adminis- trator, the rule was “not about pollution control” so much as it was “an investment opportunity” for States, especially “investments in renewables and clean energy.” Oversight 24 WEST VIRGINIA v. EPA

Hearing on EPA’s Proposed Carbon Pollution Standards for Existing Power Plants before the Senate Committee on En- vironment and Public Works, 113th Cong., 2d Sess., p. 33 (2014). This view of EPA’s authority was not only unprecedented; it also effected a “fundamental revision of the statute, changing it from [one sort of] scheme of . . . regulation” into an entirely different kind. MCI, 512 U. S., at 231. Under the Agency’s prior view of Section 111, its role was limited to ensuring the efficient pollution performance of each indi- vidual regulated source. Under that paradigm, if a source was already operating at that level, there was nothing more for EPA to do. Under its newly “discover[ed]” authority, Utility Air, 573 U. S., at 324, however, EPA can demand much greater reductions in emissions based on a very dif- ferent kind of policy judgment: that it would be “best” if coal made up a much smaller share of national electricity gen- eration. And on this view of EPA’s authority, it could go further, perhaps forcing coal plants to “shift” away virtually all of their generation—i.e., to cease making power alto- gether.3 The Government attempts to downplay the magnitude of this “unprecedented power over American industry.” In- dustrial Union Dept., AFL–CIO v. American Petroleum In- stitute, 448 U. S. 607, 645 (1980) (plurality opinion). The amount of generation shifting ordered, it argues, must be “adequately demonstrated” and “best” in light of the statu-

—————— 3 The dissent suggests that EPA could bring about the same result by,

for example, simply requiring coal plants to become natural gas plants, and that this would fit within the prior regulatory approach of efficiency- improving, at-the-source measures. Post, at 24. Of course, EPA has never ordered anything remotely like that, and we doubt it could. Sec- tion 111(d) empowers EPA to guide States in “establish[ing] standards of performance” for “existing source[s],” §7411(d)(1), not to direct existing sources to effectively cease to exist. Cite as: 597 U. S. ____ (2022) 25

tory factors of “cost,” “nonair quality health and environ- mental impact,” and “energy requirements.” 42 U. S. C. §7411(a)(1). EPA therefore must limit the magnitude of generation shift it demands to a level that will not be “exor- bitantly costly” or “threaten the reliability of the grid.” Brief for Federal Respondents 42. But this argument does not so much limit the breadth of the Government’s claimed authority as reveal it. On EPA’s view of Section 111(d), Congress implicitly tasked it, and it alone, with balancing the many vital considerations of na- tional policy implicated in deciding how Americans will get their energy. EPA decides, for instance, how much of a switch from coal to natural gas is practically feasible by 2020, 2025, and 2030 before the grid collapses, and how high energy prices can go as a result before they become unreasonably “exorbitant.” There is little reason to think Congress assigned such de- cisions to the Agency. For one thing, as EPA itself admitted when requesting special funding, “Understand[ing] and project[ing] system-wide . . . trends in areas such as elec- tricity transmission, distribution, and storage” requires “technical and policy expertise not traditionally needed in EPA regulatory development.” EPA, Fiscal Year 2016: Jus- tification of Appropriation Estimates for the Committee on Appropriations 213 (2015) (emphasis added). “When [an] agency has no comparative expertise” in making certain policy judgments, we have said, “Congress presumably would not” task it with doing so. Kisor v. Wilkie, 588 U. S. ___, ___ (2019) (slip op., at 17); see also Gonzales, 546 U. S., at 266–267. We also find it “highly unlikely that Congress would leave” to “agency discretion” the decision of how much coal- based generation there should be over the coming decades. MCI, 512 U. S., at 231; see also Brown & Williamson, 529 U. S., at 160 (“We are confident that Congress could not have intended to delegate a decision of such economic and 26 WEST VIRGINIA v. EPA

political significance to an agency in so cryptic a fashion.”). The basic and consequential tradeoffs involved in such a choice are ones that Congress would likely have intended for itself. See W. Eskridge, Interpreting Law: A Primer on How To Read Statutes and the Constitution 288 (2016) (“Even if Congress has delegated an agency general rule- making or adjudicatory power, judges presume that Con- gress does not delegate its authority to settle or amend ma- jor social and economic policy decisions.”). Congress certainly has not conferred a like authority upon EPA any- where else in the Clean Air Act. The last place one would expect to find it is in the previously little-used backwater of Section 111(d). The dissent contends that there is nothing surprising about EPA dictating the optimal mix of energy sources na- tionwide, since that sort of mandate will reduce air pollu- tion from power plants, which is EPA’s bread and butter. Post, at 20–22. But that does not follow. Forbidding evic- tions may slow the spread of disease, but the CDC’s order- ing such a measure certainly “raise[s] an eyebrow.” Post, at 18. We would not expect the Department of Homeland Se- curity to make trade or foreign policy even though doing so could decrease illegal immigration. And no one would con- sider generation shifting a “tool” in OSHA’s “toolbox,” post, at 21, even though reducing generation at coal plants would reduce workplace illness and injury from coal dust. The dissent also cites our decision in American Elec. Power Co. v. Connecticut, 564 U. S. 410 (2011). Post, at 20. The question there, however, was whether Congress wanted district court judges to decide, under unwritten fed- eral nuisance law, “whether and how to regulate carbon- dioxide emissions from powerplants.” 564 U. S., at 426. We answered no, given the existence of Section 111(d). But we said nothing about the ways in which Congress intended EPA to exercise its power under that provision. And it is doubtful we had in mind that it would claim the authority Cite as: 597 U. S. ____ (2022) 27

to require a large shift from coal to natural gas, wind, and solar. After all, EPA had never regulated in that manner, despite having issued many prior rules governing power plants under Section 111. See, e.g., 71 Fed. Reg. 9866 (2006); 70 Fed. Reg. 28616; 44 Fed. Reg. 33580; 36 Fed. Reg. 24875 (1973).4 Finally, we cannot ignore that the regulatory writ EPA newly uncovered conveniently enabled it to enact a program that, long after the dangers posed by greenhouse gas emis- sions “had become well known, Congress considered and re- jected” multiple times. Brown & Williamson, 529 U. S., at 144; see also Alabama Assn., 594 U. S., at ___ (slip op., at 2); Bunte Brothers, 312 U. S., at 352 (lack of authority not previously exercised “reinforced by [agency’s] unsuccessful attempt . . . to secure from Congress an express grant of [the challenged] authority”). At bottom, the Clean Power Plan essentially adopted a cap-and-trade scheme, or set of state cap-and-trade schemes, for carbon. See 80 Fed. Reg. 64734 (“Emissions trading is . . . an integral part of our BSER analysis.”). Congress, however, has consistently re- jected proposals to amend the Clean Air Act to create such a program. See, e.g., American Clean Energy and Security

—————— 4 According to the dissent, “EPA is always controlling the mix of energy

sources” under Section 111 because all of the Agency’s rules impose some costs on regulated plants, and therefore (all else equal) cause those plants to lose some share of the electricity market. Post, at 22. But there is an obvious difference between (1) issuing a rule that may end up caus- ing an incidental loss of coal’s market share, and (2) simply announcing what the market share of coal, natural gas, wind, and solar must be, and then requiring plants to reduce operations or subsidize their competitors to get there. No one has ever thought that the Clean Power Plan was just business as usual. See American Lung Assn. v. EPA, 985 F. 3d 914, 1000 (CADC 2021) (Walker, J., dissenting) (“Leaders of the environmen- tal movement considered the rule ‘groundbreaking,’ called its announce- ment ‘historic,’ and labeled it a ‘critically important catalyst.’ ” (footnotes omitted)). 28 WEST VIRGINIA v. EPA

Act of 2009, H. R. 2454, 111th Cong., 1st Sess.; Clean En- ergy Jobs and American Power Act, S. 1733, 111th Cong., 1st Sess. (2009). It has also declined to enact similar measures, such as a carbon tax. See, e.g., Climate Protec- tion Act of 2013, S. 332, 113th Cong., 1st Sess.; Save our Climate Act of 2011, H. R. 3242, 112th Cong., 1st Sess. “The importance of the issue,” along with the fact that the same basic scheme EPA adopted “has been the subject of an earnest and profound debate across the country, . . . makes the oblique form of the claimed delegation all the more sus- pect.” Gonzales, 546 U. S., at 267–268 (internal quotation marks omitted). C Given these circumstances, our precedent counsels skep- ticism toward EPA’s claim that Section 111 empowers it to devise carbon emissions caps based on a generation shifting approach. To overcome that skepticism, the Government must—under the major questions doctrine—point to “clear congressional authorization” to regulate in that manner. Utility Air, 573 U. S., at 324. All the Government can offer, however, is the Agency’s authority to establish emissions caps at a level reflecting “the application of the best system of emission reduction . . . adequately demonstrated.” 42 U. S. C. §7411(a)(1). As a matter of “definitional possibilities,” FCC v. AT&T Inc., 562 U. S. 397, 407 (2011), generation shifting can be described as a “system”—“an aggregation or assemblage of objects united by some form of regular interaction,” Brief for Fed- eral Respondents 31—capable of reducing emissions. But of course almost anything could constitute such a “system”; shorn of all context, the word is an empty vessel. Such a vague statutory grant is not close to the sort of clear author- ization required by our precedents. The Government, echoed by the other respondents, looks to other provisions of the Clean Air Act for support. It Cite as: 597 U. S. ____ (2022) 29

points out that the Act elsewhere uses the word “system” or “similar words” to describe cap-and-trade schemes or other sector-wide mechanisms for reducing pollution. Ibid. The Acid Rain program set out in Title IV of the Act establishes a cap-and-trade scheme for reducing sulfur dioxide emis- sions, which the statute refers to as an “emission allocation and transfer system.” §7651(b) (emphasis added). And Sec- tion 110 of the NAAQS program specifies that “marketable permits” and “auctions of emissions rights” qualify as “con- trol measures, means, or techniques” that States may adopt in their state implementation plans in order “to meet the applicable requirements of ” a NAAQS. §7410(a)(2)(A). If the word “system” or similar words like “technique” or “means” can encompass cap-and-trade, the Government maintains, why not in Section 111? But just because a cap-and-trade “system” can be used to reduce emissions does not mean that it is the kind of “sys- tem of emission reduction” referred to in Section 111. In- deed, the Government’s examples demonstrate why it is not. First, unlike Section 111, the Acid Rain and NAAQS pro- grams contemplate trading systems as a means of comply- ing with an already established emissions limit, set either directly by Congress (as with Acid Rain, see 42 U. S. C. §7651c) or by reference to the safe concentration of the pol- lutant in the ambient air (as with the NAAQS). In Section 111, by contrast, it is EPA’s job to come up with the cap itself: the “numerical limit on emissions” that States must apply to each source. 80 Fed. Reg. 64768. We doubt that Congress directed the Agency to set an emissions cap at the level “which reflects the degree of emission limitation achievable through the application of [a cap-and-trade] sys- tem,” §7411(a)(1), for that degree is indeterminate. It is one thing for Congress to authorize regulated sources to use trading to comply with a preset cap, or a cap that must be based on some scientific, objective criterion, such as the 30 WEST VIRGINIA v. EPA

NAAQS. It is quite another to simply authorize EPA to set the cap itself wherever the Agency sees fit. Second, Congress added the above authorizations for the use of emissions trading programs in 1990, simultaneous with amending Section 111 to its present form. At the time, cap-and-trade was a novel and highly touted concept. The Acid Rain program was “the nation’s first-ever emissions trading program.” L. Heinzerling & R. Steinzor, A Perfect Storm: Mercury and the Bush Administration, 34 Env. L. Rep. 10297, 10309 (2004). And Congress went out of its way to amend the NAAQS statute to make absolutely clear that the “measures, means, [and] techniques” States could use to meet the NAAQS included cap-and-trade. §7410(a)(2)(A). Yet “not a peep was heard from Congress about the possibility that a trading regime could be in- stalled under §111.” Id., at 10309. Finally, the Government notes that other parts of the Clean Air Act, past and present, have “explicitly limited the permissible components of a particular ‘system’ ” of emis- sion reduction in some regard. Brief for Federal Respond- ents 32. For instance, a separate section of the statute em- powers EPA to require the “degree of reduction achievable through the retrofit application of the best system of contin- uous emission reduction.” §7651f(b)(2) (emphasis added). The comparatively unadorned use of the phrase “best sys- tem of emission reduction” in Section 111, the Government urges, “suggest[s] a conscious congressional” choice not to limit the measures that may constitute the BSER to those applicable at or to an individual source. Id., at 32. These arguments, however, concern an interpretive ques- tion that is not at issue. We have no occasion to decide whether the statutory phrase “system of emission reduc- tion” refers exclusively to measures that improve the pollu- tion performance of individual sources, such that all other actions are ineligible to qualify as the BSER. To be sure, it is pertinent to our analysis that EPA has acted consistent Cite as: 597 U. S. ____ (2022) 31

with such a limitation for the first four decades of the stat- ute’s existence. But the only interpretive question before us, and the only one we answer, is more narrow: whether the “best system of emission reduction” identified by EPA in the Clean Power Plan was within the authority granted to the Agency in Section 111(d) of the Clean Air Act. For the reasons given, the answer is no.5 * * * Capping carbon dioxide emissions at a level that will force a nationwide transition away from the use of coal to generate electricity may be a sensible “solution to the crisis of the day.” New York v. United States, 505 U. S. 144, 187 (1992). But it is not plausible that Congress gave EPA the authority to adopt on its own such a regulatory scheme in Section 111(d). A decision of such magnitude and conse- quence rests with Congress itself, or an agency acting pur- suant to a clear delegation from that representative body. The judgment of the Court of Appeals for the District of Co- lumbia Circuit is reversed, and the cases are remanded for further proceedings consistent with this opinion.

It is so ordered.

—————— 5 We find it odd that the dissent accuses us of champing at the bit to

“constrain EPA’s efforts to address climate change,” post, at 4, yet also chides us for “mak[ing] no effort” to opine—in what would be plain dicta—on “how far [our] opinion constrain[s] EPA,” post, at 12. Cite as: 597 U. S. ____ (2022) 1

GORSUCH, J., concurring

WEST VIRGINIA, ET AL., PETITIONERS 20–1530 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

THE NORTH AMERICAN COAL CORPORATION, PETITIONER 20–1531 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

WESTMORELAND MINING HOLDINGS LLC, PETITIONER 20–1778 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

NORTH DAKOTA, PETITIONER 20–1780 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL. ON WRITS OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE DISTRICT OF COLUMBIA CIRCUIT [June 30, 2022]

JUSTICE GORSUCH, with whom JUSTICE ALITO joins, con- curring. To resolve today’s case the Court invokes the major ques- tions doctrine. Under that doctrine’s terms, administrative agencies must be able to point to “ ‘clear congressional au- thorization’ ” when they claim the power to make decisions of vast “ ‘economic and political significance.’ ” Ante, at 17, 19. Like many parallel clear-statement rules in our law, this one operates to protect foundational constitutional 2 WEST VIRGINIA v. EPA

guarantees. I join the Court’s opinion and write to offer some additional observations about the doctrine on which it rests. I A One of the Judiciary’s most solemn duties is to ensure that acts of Congress are applied in accordance with the Constitution in the cases that come before us. To help fulfill that duty, courts have developed certain “clear-statement” rules. These rules assume that, absent a clear statement otherwise, Congress means for its laws to operate in con- gruence with the Constitution rather than test its bounds. In this way, these clear-statement rules help courts “act as faithful agents of the Constitution.” A. Barrett, Substan- tive Canons and Faithful Agency, 90 B. U. L. Rev. 109, 169 (2010) (Barrett). Consider some examples. The Constitution prohibits Congress from passing laws imposing various types of ret- roactive liability. See Art. I, § 9; Landgraf v. USI Film Products, 511 U. S. 244, 265–266 (1994). Consistent with this rule, Chief Justice Marshall long ago advised that “a court . . . ought to struggle hard against a [statutory] con- struction which will, by a retrospective operation, affect the rights of parties.” United States v. Schooner Peggy, 1 Cranch 103, 110 (1801). Justice Paterson likewise in- sisted that courts must interpret statutes to apply only pro- spectively “unless they are so clear, strong, and imperative, that no other meaning can be annexed to them.” United States v. Heth, 3 Cranch 399, 413 (1806). The Constitution also incorporates the doctrine of sover- eign immunity. See, e.g., Hans v. Louisiana, 134 U. S. 1, 12–17 (1890). To enforce that doctrine, courts have consist- ently held that “nothing but express words, or an insur- mountable implication” would justify the conclusion that Cite as: 597 U. S. ____ (2022) 3

lawmakers intended to abrogate the States’ sovereign im- munity. Chisholm v. Georgia, 2 Dall. 419, 450 (1793) (Ire- dell, J., dissenting); see Seminole Tribe of Fla. v. Florida, 517 U. S. 44, 55 (1996). In a similar vein, Justice Story ob- served that “[i]t is a general rule in the interpretation of legislative acts not to construe them to embrace the sover- eign power or government, unless expressly named or in- cluded by necessary implication.” United States v. Greene, 26 F. Cas. 33, 34 (No. 15, 258) (CC Me. 1827). The major questions doctrine works in much the same way to protect the Constitution’s separation of powers. Ante, at 19. In Article I, “the People” vested “[a]ll” federal “legislative powers . . . in Congress.” Preamble; Art. I, § 1. As Chief Justice Marshall put it, this means that “im- portant subjects . . . must be entirely regulated by the leg- islature itself,” even if Congress may leave the Executive “to act under such general provisions to fill up the details.” Wayman v. Southard, 10 Wheat. 1, 42–43 (1825). Doubt- less, what qualifies as an important subject and what con- stitutes a detail may be debated. See, e.g., Gundy v. United States, 588 U. S. ___, ___–___ (2019) (plurality opinion) (slip op., at 4–6); id., at ___–___ (GORSUCH, J., dissenting) (slip op., at 10–12). But no less than its rules against retroactive legislation or protecting sovereign immunity, the Constitu- tion’s rule vesting federal legislative power in Congress is “vital to the integrity and maintenance of the system of gov- ernment ordained by the Constitution.” Marshall Field & Co. v. Clark, 143 U. S. 649, 692 (1892). It is vital because the framers believed that a republic— a thing of the people—would be more likely to enact just laws than a regime administered by a ruling class of largely unaccountable “ministers.” The Federalist No. 11, p. 85 (C. Rossiter ed. 1961) (A. Hamilton). From time to time, some 4 WEST VIRGINIA v. EPA

have questioned that assessment.1 But by vesting the law- making power in the people’s elected representatives, the Constitution sought to ensure “not only that all power [w]ould be derived from the people,” but also “that those [e]ntrusted with it should be kept in dependence on the peo- ple.” Id., No. 37, at 227 (J. Madison). The Constitution, too, placed its trust not in the hands of “a few, but [in] a number of hands,” ibid., so that those who make our laws would bet- ter reflect the diversity of the people they represent and have an “immediate dependence on, and an intimate sym- pathy with, the people.” Id., No. 52, at 327 (J. Madison). Today, some might describe the Constitution as having de- signed the federal lawmaking process to capture the wis- dom of the masses. See P. Hamburger, Is Administrative Law Unlawful? 502–503 (2014). Admittedly, lawmaking under our Constitution can be difficult. But that is nothing particular to our time nor any accident. The framers believed that the power to make new laws regulating private conduct was a grave one that could, if not properly checked, pose a serious threat to individual liberty. See The Federalist No. 48, at 309–312 (J. Madison); see also id., No. 73, at 441–442 (A. Hamilton). As a result,

—————— 1 For example, Woodrow Wilson famously argued that “popular sover-

eignty” “embarrasse[d]” the Nation because it made it harder to achieve “executive expertness.” The Study of Administration, 2 Pol. Sci. Q. 197, 207 (1887) (Administration). In Wilson’s eyes, the mass of the people were “selfish, ignorant, timid, stubborn, or foolish.” Id., at 208. He ex- pressed even greater disdain for particular groups, defending “[t]he white men of the South” for “rid[ding] themselves, by fair means or foul, of the intolerable burden of governments sustained by the votes of igno- rant [African-Americans].” 9 W. Wilson, History of the American People 58 (1918). He likewise denounced immigrants “from the south of Italy and men of the meaner sort out of Hungary and Poland,” who possessed “neither skill nor energy nor any initiative of quick intelligence.” 5 id., at 212. To Wilson, our Republic “tr[ied] to do too much by vote.” Admin- istration 214. Cite as: 597 U. S. ____ (2022) 5

the framers deliberately sought to make lawmaking diffi- cult by insisting that two houses of Congress must agree to any new law and the President must concur or a legislative supermajority must override his veto. The difficulty of the design sought to serve other ends too. By effectively requiring a broad consensus to pass legisla- tion, the Constitution sought to ensure that any new laws would enjoy wide social acceptance, profit from input by an array of different perspectives during their consideration, and thanks to all this prove stable over time. See id., No. 10, at 82–84 (J. Madison). The need for compromise inher- ent in this design also sought to protect minorities by en- suring that their votes would often decide the fate of pro- posed legislation—allowing them to wield real power alongside the majority. See id., No. 51, at 322–324 (J. Mad- ison). The difficulty of legislating at the federal level aimed as well to preserve room for lawmaking “by governments more local and more accountable than a distant federal” au- thority, National Federation of Independent Business v. Sebelius, 567 U. S. 519, 536 (2012) (plurality opinion), and in this way allow States to serve as “laborator[ies]” for “novel social and economic experiments,” New State Ice Co. v. Liebmann, 285 U. S. 262, 311 (1932) (Brandeis, J., dis- senting); see J. Sutton, 51 Imperfect Solutions: States and the Making of American Constitutional Law 11 (2018). Permitting Congress to divest its legislative power to the Executive Branch would “dash [this] whole scheme.” De- partment of Transportation v. Association of American Rail- roads, 575 U. S. 43, 61 (2015) (ALITO, J., concurring). Leg- islation would risk becoming nothing more than the will of the current President, or, worse yet, the will of unelected officials barely responsive to him. See S. Breyer, Making Our Democracy Work: A Judge’s View 110 (2010) (“[T]he president may not have the time or willingness to review [agency] decisions”). In a world like that, agencies could churn out new laws more or less at whim. Intrusions on 6 WEST VIRGINIA v. EPA

liberty would not be difficult and rare, but easy and profuse. See The Federalist No. 47, at 303 (J. Madison); id., No. 62, at 378 (J. Madison). Stability would be lost, with vast num- bers of laws changing with every new presidential admin- istration. Rather than embody a wide social consensus and input from minority voices, laws would more often bear the support only of the party currently in power. Powerful spe- cial interests, which are sometimes “uniquely” able to influ- ence the agendas of administrative agencies, would flourish while others would be left to ever-shifting winds. T. Merrill, Capture Theory and the Courts: 1967–1983, 72 Chi.-Kent L. Rev. 1039, 1043 (1997). Finally, little would remain to stop agencies from moving into areas where state authority has traditionally predominated. See, e.g., Solid Waste Agency of Northern Cook Cty. v. Army Corps of Engineers, 531 U. S. 159, 173–174 (2001) (SWANC). That would be a particularly ironic outcome, given that so many States have robust nondelegation doctrines designed to ensure demo- cratic accountability in their state lawmaking processes. See R. May, The Nondelegation Doctrine is Alive and Well in the States, The Reg. Rev. (Oct. 15, 2020). B Much as constitutional rules about retroactive legislation and sovereign immunity have their corollary clear-state- ment rules, Article I’s Vesting Clause has its own: the ma- jor questions doctrine. See Gundy, 588 U. S., at ___–___ (GORSUCH, J., dissenting) (slip op., at 20–21). Some version of this clear-statement rule can be traced to at least 1897, when this Court confronted a case involving the Interstate Commerce Commission, the federal government’s “first modern regulatory agency.” S. Dudley, Milestones in the Evolution of the Administrative State 3 (Nov. 2020). The ICC argued that Congress had endowed it with the power to set carriage prices for railroads. See ICC v. Cincinnati, N. O. & T. P. R. Co., 167 U. S. 479, 499 (1897). The Court Cite as: 597 U. S. ____ (2022) 7

deemed that claimed authority “a power of supreme deli- cacy and importance,” given the role railroads then played in the Nation’s life. Id., at 505. Therefore, the Court ex- plained, a special rule applied: “That Congress has transferred such a power to any ad- ministrative body is not to be presumed or implied from any doubtful and uncertain language. The words and phrases efficacious to make such a delegation of power are well understood, and have been frequently used, and if Congress had intended to grant such a power to the [agency], it cannot be doubted that it would have used language open to no misconstruction, but clear and direct.” Ibid. (emphasis added). With the explosive growth of the administrative state since 1970, the major questions doctrine soon took on spe- cial importance.2 In 1980, this Court held it “unreasonable to assume” that Congress gave an agency “unprecedented power[s]” in the “absence of a clear [legislative] mandate.” Industrial Union Dept., AFL–CIO v. American Petroleum Institute, 448 U. S. 607, 645 (plurality opinion). In the years that followed, the Court routinely enforced “the non- delegation doctrine” through “the interpretation of statu-

—————— 2 In the 1960s and 1970s, Congress created dozens of new federal ad-

ministrative agencies. See W. Howell & D. Lewis, Agencies by Presiden- tial Design, 64 J. of Politics 1095, 1105 (Nov. 2002). Between 1970 and 1990, the Code of Federal Regulations grew from about 44,000 pages to about 106,000. See C. DeMuth, Can the Administrative State be Tamed?, 8 J. Legal Analysis 121, 126 (Feb. 2016). Today, Congress is- sues “roughly two hundred to four hundred laws” every year, while “fed- eral administrative agencies adopt something on the order of three thou- sand to five thousand final rules.” R. Cass, Rulemaking Then and Now: From Management to Lawmaking, 28 Geo. Mason L. Rev. 683, 694 (2021). Beyond that, agencies regularly “produce thousands, if not mil- lions,” of guidance documents which, as a practical matter, bind affected parties too. See C. Coglianese, Illuminating Regulatory Guidance, 9 Mich. J. Env. & Admin. L. 243, 247–248 (2020). 8 WEST VIRGINIA v. EPA

tory texts, and, more particularly, [by] giving narrow con- structions to statutory delegations that might otherwise be thought to be unconstitutional.” Mistretta v. United States, 488 U. S. 361, 373, n. 7 (1989). In fact, this Court applied the major questions doctrine in “all corners of the adminis- trative state,” whether the issue at hand involved an agency’s asserted power to regulate tobacco products, ban drugs used in physician-assisted suicide, extend Clean Air Act regulations to private homes, impose an eviction mora- torium, or enforce a vaccine mandate. Ante, at 17; see FDA v. Brown & Williamson Tobacco Corp., 529 U. S. 120, 160 (2000); Gonzales v. Oregon, 546 U. S. 243, 267 (2006); Util- ity Air Regulatory Group v. EPA, 573 U. S. 302, 324 (2014); Alabama Assn. of Realtors v. Department of Health and Hu- man Servs., 594 U. S. ___, ___ (2021) (per curiam) (slip op., at 6); National Federation of Independent Business v. OSHA, 595 U. S. ___, ___ (2022) (per curiam) (slip op., at 6).3 The Court has applied the major questions doctrine for the same reason it has applied other similar clear-state- ment rules—to ensure that the government does “not inad- vertently cross constitutional lines.” Barrett 175. And the constitutional lines at stake here are surely no less im- portant than those this Court has long held sufficient to jus- tify parallel clear-statement rules. At stake is not just a question of retroactive liability or sovereign immunity, but basic questions about self-government, equality, fair notice,

—————— 3 At times, this Court applied the major questions doctrine more like

an ambiguity canon. See FDA v. Brown & Williamson Tobacco Corp., 529 U. S. 120, 159 (2000). Ambiguity canons merely instruct courts on how to “choos[e] between equally plausible interpretations of ambiguous text,” and are thus weaker than clear-statement rules. Barrett 109. But our precedents have usually applied the doctrine as a clear-statement rule, and the Court today confirms that is the proper way to apply it. See ante, at 19–20, 28. Cite as: 597 U. S. ____ (2022) 9

federalism, and the separation of powers. See Part I–A, su- pra. The major questions doctrine seeks to protect against “unintentional, oblique, or otherwise unlikely” intrusions on these interests. NFIB v. OSHA, 595 U. S., at ___ (GORSUCH, J., concurring) (slip op., at 5). The doctrine does so by ensuring that, when agencies seek to resolve major questions, they at least act with clear congressional author- ization and do not “exploit some gap, ambiguity, or doubtful expression in Congress’s statutes to assume responsibilities far beyond” those the people’s representatives actually con- ferred on them. Ibid. As the Court aptly summarizes it today, the doctrine addresses “a particular and recurring problem: agencies asserting highly consequential power be- yond what Congress could reasonably be understood to have granted.” Ante, at 20. II A Turning from the doctrine’s function to its application, it seems to me that our cases supply a good deal of guidance about when an agency action involves a major question for which clear congressional authority is required. First, this Court has indicated that the doctrine applies when an agency claims the power to resolve a matter of great “political significance,” NFIB v. OSHA, 595 U. S., at ___ (slip op., at 6) (internal quotation marks omitted), or end an “earnest and profound debate across the country,” Gonzales, 546 U. S., at 267–268 (internal quotation marks omitted); see ante, at 17. So, for example, in Gonzales, the Court found that the doctrine applied when the Attorney General issued a regulation that would have effectively banned most forms of physician-assisted suicide even as certain States were considering whether to permit the prac- tice. 546 U. S., at 267. And in NFIB v. OSHA, the Court held the doctrine applied when an agency sought to man- date COVID–19 vaccines nationwide for most workers at a 10 WEST VIRGINIA v. EPA

time when Congress and state legislatures were engaged in robust debates over vaccine mandates. 595 U. S., at ___ (slip op., at 5); id., at ___ (GORSUCH, J., concurring) (slip op., at 3). Relatedly, this Court has found it telling when Con- gress has “ ‘considered and rejected’ ” bills authorizing something akin to the agency’s proposed course of action. Ante, at 20, 27 (quoting Brown & Williamson, 529 U. S., at 144). That too may be a sign that an agency is attempting to “ ‘work [a]round’ ” the legislative process to resolve for it- self a question of great political significance. NFIB v. OSHA, 595 U. S., at ___ (GORSUCH, J., concurring) (slip op., at 3).4 Second, this Court has said that an agency must point to clear congressional authorization when it seeks to regulate “ ‘a significant portion of the American economy,’ ” ante, at 18 (quoting Utility Air, 573 U. S., at 324), or require “bil- lions of dollars in spending” by private persons or entities, King v. Burwell, 576 U. S. 473, 485 (2015). The Court has held that regulating tobacco products, eliminating rate reg- ulation in the telecommunications industry, subjecting pri- vate homes to Clean Air Act restrictions, and suspending local housing laws and regulations can sometimes check this box. See Brown & Williamson, 529 U. S., at 160; MCI Telecommunications Corp. v. American Telephone & Tele- graph Co., 512 U. S. 218, 231 (1994) (MCI); Utility Air, 573 U. S., at 324; Alabama Assn. of Realtors, 594 U. S., at ___ (slip op., at 6).

—————— 4 In the dissent’s view, the Court has erred both today and in the past

by pointing to failed legislation. Post, at 27–28 (opinion of KAGAN, J.). But the Court has not pointed to failed legislation to resolve what a duly enacted statutory text means, only to help resolve the antecedent ques- tion whether the agency’s challenged action implicates a major question. The dissent endorses looking to extrinsic evidence to resolve that ques- tion too. See post, at 21–22 (discussing whether there is a “mismatch” between an agency’s expertise and its challenged action). Cite as: 597 U. S. ____ (2022) 11

Third, this Court has said that the major questions doc- trine may apply when an agency seeks to “intrud[e] into an area that is the particular domain of state law.” Ibid. Of course, another longstanding clear-statement rule—the federalism canon—also applies in these situations. To pre- serve the “proper balance between the States and the Fed- eral Government” and enforce limits on Congress’s Com- merce Clause power, courts must “ ‘be certain of Congress’s intent’ ” before finding that it “legislate[d] in areas tradi- tionally regulated by the States.” Gregory v. Ashcroft, 501 U. S. 452, 459–460 (1991). But unsurprisingly, the major questions doctrine and the federalism canon often travel to- gether. When an agency claims the power to regulate vast swaths of American life, it not only risks intruding on Con- gress’s power, it also risks intruding on powers reserved to the States. See SWANC, 531 U. S., at 162, 174. While this list of triggers may not be exclusive, each of the signs the Court has found significant in the past is pre- sent here, making this a relatively easy case for the doc- trine’s application. The EPA claims the power to force coal and gas-fired power plants “to cease [operating] altogether.” Ante, at 24. Whether these plants should be allowed to op- erate is a question on which people today may disagree, but it is a question everyone can agree is vitally important. See ante, at 24–25. Congress has debated the matter fre- quently. Ibid.; see generally Climate Change, The History of a Consensus and the Causes of Inaction, Hearing before the Subcommittee on Environment of the House Committee on Oversight and Reform, 116th Cong., 1st Sess., pt. I (2019). And so far it has “conspicuously and repeatedly de- clined” to adopt legislation similar to the Clean Power Plan (CPP). Ante, at 20; see American Lung Assn. v. EPA, 985 F. 3d 914, 998, n. 19 (CADC 2021) (Walker, J., concurring in part, concurring in judgment in part, and dissenting in part) (cataloguing failed legislative proposals); cf. Brown & 12 WEST VIRGINIA v. EPA

Williamson, 529 U. S., at 144. It seems that fact has frus- trated the Executive Branch and led it to attempt its own regulatory solution in the CPP. See 985 F. 3d, at 998, n. 20 (President stating that “ ‘if Congress won’t act soon . . . I will’ ”); cf. United States Telecom Assn. v. FCC, 855 F. 3d 381, 423–424 (CADC 2017) (Kavanaugh, J., dissenting from denial of rehearing en banc) (noting a “President’s interven- tion [may] underscor[e] the enormous significance” of a reg- ulation). Other suggestive factors are present too. “The electric power sector is among the largest in the U. S. economy, with links to every other sector.” N. Richardson, Keeping Big Cases From Making Bad Law: The Resurgent “Major Questions” Doctrine, 49 Conn. L. Rev. 355, 388 (2016). The Executive Branch has acknowledged that its proposed rule would force an “aggressive transformation” of the electricity sector through “transition to zero-carbon renewable energy sources.” White House Fact Sheet, App. in American Lung Assn. v. EPA, No. 19–1140 (CADC), pp. 2076–2077. The Executive Branch has also predicted its rule would force dozens of power plants to close and eliminate thousands of jobs by 2025. See EPA, Regulatory Impact Analysis for the Clean Power Plan Final Rule 3–27, 3–30, 3–33, 6–25 (Oct. 23, 2015). And industry analysts have estimated the CPP would cause consumers’ electricity costs to rise by over $200 billion. See National Mining Assn., EPA’s Clean Power Plan: An Economic Impact Analysis 2, 4 (2015). Finally, the CPP unquestionably has an impact on federalism, as “the regulation of utilities is one of the most important of the functions traditionally associated with the police power of the States.” Arkansas Elec. Cooperative Corp. v. Arkansas Pub. Serv. Comm’n, 461 U. S. 375, 377 (1983). None of this is to say the policy the agency seeks to pursue is unwise or should not be pursued. It is only to say that the agency seeks to resolve for itself the sort of question normally re- served for Congress. As a result, we look for clear evidence Cite as: 597 U. S. ____ (2022) 13

that the people’s representatives in Congress have actually afforded the agency the power it claims. B At this point, the question becomes what qualifies as a clear congressional statement authorizing an agency’s ac- tion. Courts have long experience applying clear-statement rules throughout the law, and our cases have identified sev- eral telling clues in this context too. First, courts must look to the legislative provisions on which the agency seeks to rely “ ‘with a view to their place in the overall statutory scheme.’ ” Brown & Williamson, 529 U. S., at 133. “[O]blique or elliptical language” will not supply a clear statement. Ante, at 18; see Spector v. Nor- wegian Cruise Line Ltd., 545 U. S. 119, 139 (2005) (plural- ity opinion) (cautioning against reliance on “broad or gen- eral language”). Nor may agencies seek to hide “elephants in mouseholes,” Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001), or rely on “gap filler” provi- sions, ante, at 20. So, for example, in MCI this Court re- jected the Federal Communication Commission’s attempt to eliminate rate regulation for the telecommunications in- dustry based on a “subtle” provision that empowered the FCC to “ ‘modify’ ” rates. 512 U. S., at 231. In Brown & Williamson, the Court rejected the Food and Drug Admin- istration’s attempt to regulate cigarettes based a “cryptic” statutory provision that granted the agency the power to regulate “drugs” and “devices.” 529 U. S., at 126, 156, 160. And in Gonzales, the Court doubted that Congress gave the Attorney General “broad and unusual authority” to regu- late drugs for physician-assisted suicide through “oblique” statutory language. 546 U. S., at 267. Second, courts may examine the age and focus of the stat- ute the agency invokes in relation to the problem the agency seeks to address. As the Court puts it today, it is unlikely 14 WEST VIRGINIA v. EPA

that Congress will make an “[e]xtraordinary gran[t] of reg- ulatory authority” through “vague language” in “ ‘a long-ex- tant statute.’ ” Ante, at 18–20 (quoting Utility Air, 573 U. S., at 324). Recently, too, this Court found a clear state- ment lacking when OSHA sought to impose a nationwide COVID–19 vaccine mandate based on a statutory provision that was adopted 40 years before the pandemic and that fo- cused on conditions specific to the workplace rather than a problem faced by society at large. See NFIB v. OSHA, 595 U. S., at ___ (GORSUCH, J., concurring) (slip op., at 3). Of course, sometimes old statutes may be written in ways that apply to new and previously unanticipated situations. See Sedima, S. P. R. L. v. Imrex Co., 473 U. S. 479, 499 (1985). But an agency’s attempt to deploy an old statute focused on one problem to solve a new and different problem may also be a warning sign that it is acting without clear congres- sional authority. See ante, at 18. Third, courts may examine the agency’s past interpreta- tions of the relevant statute. See ante, at 20–21. A “con- temporaneous” and long-held Executive Branch interpreta- tion of a statute is entitled to some weight as evidence of the statute’s original charge to an agency. United States v. Philbrick, 120 U. S. 52, 59 (1887). Conversely, in NFIB v. OSHA, the Court found it “telling that OSHA, in its half century of existence, ha[d] never before adopted a broad public health regulation” under the statute that the agency sought to invoke as authority for a nationwide vaccine man- date. 595 U. S., at ___ (slip op., at 8); ante, at 18; see also Brown & Williamson, 529 U. S., at 158–159 (noting that for decades the FDA had said it lacked statutory power to reg- ulate cigarettes). As the Court states today, “ ‘the want of [an] assertion of power by those who presumably would be alert’ ” to it is “ ‘significant in determining whether such power was actually conferred.’ ” Ante, at 21. When an agency claims to have found a previously “unheralded power,” its assertion generally warrants “a measure of Cite as: 597 U. S. ____ (2022) 15

skepticism.” Utility Air, 573 U. S., at 324. Fourth, skepticism may be merited when there is a mis- match between an agency’s challenged action and its con- gressionally assigned mission and expertise. Ante, at 25. As the Court explains, “[w]hen an agency has no compara- tive expertise in making certain policy judgments, . . . Con- gress presumably would not task it with doing so.” Ibid. (internal quotation marks and alterations omitted). So, for example, in Alabama Assn. of Realtors, this Court rejected an attempt by a public health agency to regulate housing. 594 U. S., at ___ (slip op., at 5). And in NFIB v. OSHA, the Court rejected an effort by a workplace safety agency to or- dain “broad public health measures” that “f[ell] outside [its] sphere of expertise.” 595 U. S., at ___ (slip op., at 6).5 Asking these questions again yields a clear answer in our case. See ante, at 28–31. As the Court details, the agency before us cites no specific statutory authority allowing it to transform the Nation’s electrical power supply. See ante, at 28. Instead, the agency relies on a rarely invoked statu- tory provision that was passed with little debate and has been characterized as an “obscure, never-used section of the law.” Ante, at 6 (internal quotation marks omitted). Nor has the agency previously interpreted the relevant provi- sion to confer on it such vast authority; there is no original, longstanding, and consistent interpretation meriting judi-

—————— 5 The dissent not only agrees that a mismatch between an agency’s ex-

pertise and its challenged action is relevant to the major questions doc- trine analysis; the dissent suggests that such a mismatch is necessary to the doctrine’s application. See post, at 14–15. But this Court has never taken that view. See, e.g., ICC v. Cincinnati, N. O. & T. P. R. Co., 167 U. S. 479, 505 (1897) (interstate commerce agency regulating interstate railroad commerce); Industrial Union Dept., AFL–CIO v. American Pe- troleum Institute, 448 U. S. 607, 645 (1980) (plurality opinion) (workplace safety agency regulating workplace carcinogens); Brown & Williamson, 529 U. S., at 159–160 (drug agency regulating tobacco); King v. Burwell, 576 U. S. 473, 485–486 (2015) (tax agency administering tax credits). 16 WEST VIRGINIA v. EPA

cial respect. See ante, at 20–22. Finally, there is a “mis- match” between the EPA’s expertise over environmental matters and the agency’s claim that “Congress implicitly tasked it, and it alone, with balancing the many vital con- siderations of national policy implicated in deciding how Americans will get their energy.” Ante, at 25. Such a claimed power “requires technical and policy expertise not traditionally needed in [the] EPA’s regulatory develop- ment.” Ibid. (internal quotation marks omitted). Again, in observing this much, the Court does not purport to pass on the wisdom of the agency’s course. It acknowledges only that agency officials have sought to resolve a major policy question without clear legislative authorization to do so. III In places, the dissent seems to suggest that we should not be unduly “ ‘concerned’ ” with the Constitution’s assignment of the legislative power to Congress. Post, at 29 (opinion of KAGAN, J.). Echoing Woodrow Wilson, the dissent seems to think “a modern Nation” cannot afford such sentiments. Post, at 29–31. But recently, our dissenting colleagues acknowledged that the Constitution assigns “all legislative Powers” to Congress and “bar[s their] further delegation.” Gundy, 588 U. S., at ___ (plurality opinion of KAGAN, J.) (slip op., at 4) (internal quotation marks and alteration omitted). To be sure, in that case we disagreed about the exact nature of the “nondelegation inquiry” courts must em- ploy to vindicate the Constitution. Id., at ___ (slip op., at 5). But like Chief Justice Marshall, we all recognized that the Constitution does impose some limits on the delegation of legislative power. See ibid.; Wayman, 10 Wheat., at 42– 43. And while we all agree that administrative agencies have important roles to play in a modern nation, surely none of us wishes to abandon our Republic’s promise that the people and their representatives should have a mean- Cite as: 597 U. S. ____ (2022) 17

ingful say in the laws that govern them. Cf. Rucho v. Com- mon Cause, 588 U. S. ___, ___ (2019) (KAGAN, J., dissenting) (slip op., at 7) (“Republican liberty demands not only, that all power should be derived from the people; but that those entrusted with it should be kept in dependence on the peo- ple” (internal quotation marks and alteration omitted)).6 So what is our real point of disagreement? The dissent next suggests that the Court strays from its commitment to textualism by relying on a clear-statement rule (the major questions doctrine) to resolve today’s case. Post, at 28. But our law is full of clear-statement rules and has been since the founding. Our colleagues do not dispute the point. In fact, they have regularly invoked many of these rules.7 —————— 6 In the course of its argument, the dissent leans heavily on two recent

academic articles. Post, at 29. But if a battle of law reviews were the order of the day, it might be worth adding to the reading list. See, e.g., I. Wurman, Nondelegation at the Founding, 130 Yale L. J. 1490, 1493– 1494 (2021); D. Candeub, Preference and Administrative Law, 72 Admin. L. Rev. 607, 614–628 (2020); P. Hamburger, Delegation or Divesting?, 115 Nw. L. Rev. Online 88, 91–110 (2020); M. McConnell, The President Who Would Not Be King 326–335 (2020); A. Gordon, Nondelegation, 12 N. Y. U. J. L. & Liberty 718, 719 (2019); R. Cass, Delegation Reconsid- ered: A Delegation Doctrine for the Modern Administrative State, 40 Harv. J. L. & Pub. Pol’y 147, 155–161 (2017); G. Lawson & G. Seidman, “A Great Power of Attorney:” Understanding the Fiduciary Constitution 104–129 (2017); P. Hamburger, Is Administrative Law Unlawful? 377– 402 (2014); L. Alexander & S. Prakash, Reports of the Nondelegation Doctrine’s Death are Greatly Exaggerated, 70 U. Chi. L. Rev. 1297, 1298–1299 (2003); G. Lawson, Delegation and Original Meaning, 88 Va. L. Rev. 327, 335–343 (2002); D. Schoenbrod, The Delegation Doctrine: Could the Court Give It Substance? 83 Mich. L. Rev. 1223, 1252–1255, 1260–1261 (1985); see generally P. Wallison & J. Yoo, The Administra- tive State Before the Supreme Court: Perspectives on the Nondelegation Doctrine (2022). 7 See, e.g., United States v. Washington, 596 U. S. ___, ___ (2022) (slip

op., at 6) (intergovernmental immunity); Rehaif v. United States, 588 U. S. ___, ___ (2019) (slip op., at 3) (mens rea); Michigan v. Bay Mills Indian Community, 572 U. S. 782, 790–791 (2014) (sovereign immunity); Vartelas v. Holder, 566 U. S. 257, 261, 266–267 (2012) (retroactivity); Gonzalez v. Thaler, 565 U. S. 134, 141–142 (2012) (presumption against 18 WEST VIRGINIA v. EPA

If that’s not the problem, perhaps the dissent means to suggest that the major questions doctrine does not belong on the list of our clear-statement rules. At times, the dis- sent appears to dismiss the doctrine as a “get-out-of-text free car[d].” Ibid. The dissent even seems to suggest that the doctrine could threaten “the safety and efficacy of med- ications” or lead to “the routine adulteration of food.” Post, at 31. But then again, the dissent also acknowledges that the major questions doctrine should “sensibl[y]” apply in at least some situations. Post, at 14–15. The dissent even fa- vorably highlights one application of the doctrine that our colleagues criticized less than a year ago. See post, at 18 (citing Alabama Assn. of Realtors, 594 U. S. ___). And, of course, our colleagues have joined other applications of the major questions doctrine in the past. See, e.g., King, 576 U. S., at 485–486; Gonzales, 546 U. S., at 267–268. Nor does the dissent really seem to dispute that a major ques- tion is at stake in this case. As the dissent observes, the agency’s challenged action before us concerns one of “the greatest . . . challenge[s] of our time.” Post, at 21. If this case does not implicate a “question of deep economic and political significance,” King, 576 U. S., at 486 (internal quo- tation marks omitted), it is unclear what might.8 —————— rule being jurisdictional); Smith v. Bayer Corp., 564 U. S. 299, 307 (2011) (presumption that federal injunctions don’t bar parallel state proceed- ings); Fowler v. United States, 563 U. S. 668, 677 (2011) (federalism canon); Kucana v. Holder, 558 U. S. 233, 237 (2010) (presumption in fa- vor of judicial review); Holland v. Florida, 560 U. S. 631, 645–646 (2010) (presumption in favor of equitable tolling); Hamilton v. Lanning, 560 U. S. 505, 517 (2010) (presumption that Bankruptcy Code didn’t erode past practice). 8 The dissent seeks to invoke Justice Scalia as authority against the

major questions doctrine. See post, at 31–32. But the dissent neglects to mention that Justice Scalia authored or joined several of the Court’s major questions decisions, including Brown & Williamson, which the dis- sent describes as the “key case.” Post, at 15–16 (citing 529 U. S. 120); see also Whitman v. American Trucking Assns., Inc., 531 U. S. 457, 468 (2001); Utility Air, 573 U. S., at 307; A. Scalia, A Note on the Benzene Cite as: 597 U. S. ____ (2022) 19

In the end, our disagreement really seems to center on a difference of opinion about whether the statute at issue here clearly authorizes the agency to adopt the CPP. The dissent even complains that I have failed to conduct an ex- haustive analysis of the relevant statutory language. See post, at 28, n. 8. But in this concurrence, I have sought to provide some observations about the underlying doctrine on which today’s decision rests. On the merits of the case be- fore us, I join the Court’s opinion, which comprehensively sets forth why Congress did not clearly authorize the EPA to engage in a “generation shifting approach” to the produc- tion of energy in this country. Ante, at 28. In reaching its judgment, the Court hardly professes to “appoin[t] itself” “the decision-maker on climate policy.” Post, at 33. The Court acknowledges only that, under our Constitution, the people’s elected representatives in Congress are the deci- sionmakers here—and they have not clearly granted the agency the authority it claims for itself. Ante, at 31. * When Congress seems slow to solve problems, it may be only natural that those in the Executive Branch might seek to take matters into their own hands. But the Constitution does not authorize agencies to use pen-and-phone regula- tions as substitutes for laws passed by the people’s repre- sentatives. In our Republic, “[i]t is the peculiar province of the legislature to prescribe general rules for the govern- ment of society.” Fletcher v. Peck, 6 Cranch 87, 136 (1810). Because today’s decision helps safeguard that foundational constitutional promise, I am pleased to concur.

—————— Case, American Enterprise Institute, J. on Govt. & Soc., July–Aug. 1980, pp. 27–28. Cite as: 597 U. S. ____ (2022) 1

KAGAN, J., dissenting

WEST VIRGINIA, ET AL., PETITIONERS 20–1530 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

THE NORTH AMERICAN COAL CORPORATION, PETITIONER 20–1531 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

WESTMORELAND MINING HOLDINGS LLC, PETITIONER 20–1778 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL.

NORTH DAKOTA, PETITIONER 20–1780 v. ENVIRONMENTAL PROTECTION AGENCY, ET AL. ON WRITS OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE DISTRICT OF COLUMBIA CIRCUIT [June 30, 2022]

JUSTICE KAGAN, with whom JUSTICE BREYER and JUSTICE SOTOMAYOR join, dissenting. Today, the Court strips the Environmental Protection Agency (EPA) of the power Congress gave it to respond to “the most pressing environmental challenge of our time.” Massachusetts v. EPA, 549 U. S. 497, 505 (2007). Climate change’s causes and dangers are no longer sub- ject to serious doubt. Modern science is “unequivocal that 2 WEST VIRGINIA v. EPA

human influence”—in particular, the emission of green- house gases like carbon dioxide—“has warmed the atmos- phere, ocean and land.” Intergovernmental Panel on Cli- mate Change, Sixth Assessment Report, The Physical Science Basis: Headline Statements 1 (2021). The Earth is now warmer than at any time “in the history of modern civ- ilization,” with the six warmest years on record all occur- ring in the last decade. U. S. Global Change Research Pro- gram, Fourth National Climate Assessment, Vol. I, p. 10 (2017); Brief for Climate Scientists as Amici Curiae 8. The rise in temperatures brings with it “increases in heat- related deaths,” “coastal inundation and erosion,” “more frequent and intense hurricanes, floods, and other extreme weather events,” “drought,” “destruction of ecosystems,” and “potentially significant disruptions of food production.” American Elec. Power Co. v. Connecticut, 564 U. S. 410, 417 (2011) (internal quotation marks omitted). If the current rate of emissions continues, children born this year could live to see parts of the Eastern seaboard swallowed by the ocean. See Brief for Climate Scientists as Amici Curiae 6. Rising waters, scorching heat, and other severe weather conditions could force “mass migration events[,] political crises, civil unrest,” and “even state failure.” Dept. of De- fense, Climate Risk Analysis 8 (2021). And by the end of this century, climate change could be the cause of “4.6 mil- lion excess yearly deaths.” See R. Bressler, The Mortality Cost of Carbon, 12 Nature Communications 4467, p. 5 (2021). Congress charged EPA with addressing those potentially catastrophic harms, including through regulation of fossil- fuel-fired power plants. Section 111 of the Clean Air Act directs EPA to regulate stationary sources of any substance that “causes, or contributes significantly to, air pollution” and that “may reasonably be anticipated to endanger public health or welfare.” 42 U. S. C. §7411(b)(1)(A). Carbon di- oxide and other greenhouse gases fit that description. See Cite as: 597 U. S. ____ (2022) 3

American Elec. Power, 564 U. S., at 416–417; Massachu- setts, 549 U. S., at 528–532. EPA thus serves as the Na- tion’s “primary regulator of greenhouse gas emissions.” American Elec. Power, 564 U. S., at 428. And among the most significant of the entities it regulates are fossil-fuel- fired (mainly coal- and natural-gas-fired) power plants. To- day, those electricity-producing plants are responsible for about one quarter of the Nation’s greenhouse gas emissions. See EPA, Sources of Greenhouse Gas Emissions (Apr. 14, 2022), https://www.epa.gov/ghgemissions/sources-greenhouse- gas-emissions. Curbing that output is a necessary part of any effective approach for addressing climate change. To carry out its Section 111 responsibility, EPA issued the Clean Power Plan in 2015. The premise of the Plan— which no one really disputes—was that operational im- provements at the individual-plant level would either “lead to only small emission reductions” or would cost far more than a readily available regulatory alternative. 80 Fed. Reg. 64727–64728 (2015). That alternative—which fossil- fuel-fired plants were “already using to reduce their [carbon dioxide] emissions” in “a cost effective manner”—is called generation shifting. Id., at 64728, 64769. As the Court ex- plains, the term refers to ways of shifting electricity gener- ation from higher emitting sources to lower emitting ones— more specifically, from coal-fired to natural-gas-fired sources, and from both to renewable sources like solar and wind. See ante, at 8. A power company (like the many sup- porting EPA here) might divert its own resources to a cleaner source, or might participate in a cap-and-trade sys- tem with other companies to achieve the same emissions- reduction goals. This Court has obstructed EPA’s effort from the begin- ning. Right after the Obama administration issued the Clean Power Plan, the Court stayed its implementation. That action was unprecedented: Never before had the Court stayed a regulation then under review in the lower courts. 4 WEST VIRGINIA v. EPA

See Reply Brief for 29 States and State Agencies in No. 15A773, p. 33 (conceding the point). The effect of the Court’s order, followed by the Trump administration’s re- peal of the rule, was that the Clean Power Plan never went into effect. The ensuing years, though, proved the Plan’s moderation. Market forces alone caused the power industry to meet the Plan’s nationwide emissions target—through exactly the kinds of generation shifting the Plan contem- plated. See 84 Fed. Reg. 32561–32562 (2019); Brief for United States 47. So by the time yet another President took office, the Plan had become, as a practical matter, obsolete. For that reason, the Biden administration announced that, instead of putting the Plan into effect, it would commence a new rulemaking. Yet this Court determined to pronounce on the legality of the old rule anyway. The Court may be right that doing so does not violate Article III mootness rules (which are notoriously strict). See ante, at 14–16. But the Court’s docket is discretionary, and because no one is now subject to the Clean Power Plan’s terms, there was no reason to reach out to decide this case. The Court today issues what is really an advisory opinion on the proper scope of the new rule EPA is considering. That new rule will be subject anyway to immediate, pre-enforcement judi- cial review. But this Court could not wait—even to see what the new rule says—to constrain EPA’s efforts to ad- dress climate change. The limits the majority now puts on EPA’s authority fly in the face of the statute Congress wrote. The majority says it is simply “not plausible” that Congress enabled EPA to regulate power plants’ emissions through generation shift- ing. Ante, at 31. But that is just what Congress did when it broadly authorized EPA in Section 111 to select the “best system of emission reduction” for power plants. §7411(a)(1). The “best system” full stop—no ifs, ands, or buts of any kind relevant here. The parties do not dispute that generation shifting is indeed the “best system”—the Cite as: 597 U. S. ____ (2022) 5

most effective and efficient way to reduce power plants’ car- bon dioxide emissions. And no other provision in the Clean Air Act suggests that Congress meant to foreclose EPA from selecting that system; to the contrary, the Plan’s regulatory approach fits hand-in-glove with the rest of the statute. The majority’s decision rests on one claim alone: that gen- eration shifting is just too new and too big a deal for Con- gress to have authorized it in Section 111’s general terms. But that is wrong. A key reason Congress makes broad del- egations like Section 111 is so an agency can respond, ap- propriately and commensurately, to new and big problems. Congress knows what it doesn’t and can’t know when it drafts a statute; and Congress therefore gives an expert agency the power to address issues—even significant ones—as and when they arise. That is what Congress did in enacting Section 111. The majority today overrides that legislative choice. In so doing, it deprives EPA of the power needed—and the power granted—to curb the emission of greenhouse gases. I The Clean Air Act was major legislation, designed to deal with a major public policy issue. As Congress explained, its goal was to “speed up, expand, and intensify the war against air pollution” in all its forms. H. R. Rep. No. 91– 1146, p. 1 (1970). Or as this Court similarly recognized, the Act was a “drastic remedy to what was perceived as a seri- ous and otherwise uncheckable problem.” Union Elec. Co. v. EPA, 427 U. S. 246, 256 (1976). The Act, as the majority describes, established three major regulatory programs to control air pollution from stationary sources like power plants. See ante, at 2–6. The National Ambient Air Quality Standards (NAAQS) and Hazardous Air Pollutants (HAP) programs prescribe standards for specified pollutants, not including carbon dioxide. Section 111’s New Source Perfor- mance Standards program provides an additional tool for 6 WEST VIRGINIA v. EPA

regulating emissions from categories of stationary sources deemed to contribute significantly to pollution. As applied to existing (not new) sources, the program mandates—via Section 111(d)—that EPA set emissions levels for pollu- tants not covered by the NAAQS or HAP programs, includ- ing carbon dioxide. Section 111(d) thus ensures that EPA regulates existing power plants’ emissions of all pollutants. When the pollu- tant at issue falls within the NAAQS or HAP programs, EPA need do no more. But when the pollutant falls outside those programs, Section 111(d) requires EPA to set an emis- sions level for currently operating power plants (and other stationary sources). That means no pollutant from such a source can go unregulated: As the Senate Report explained, Section 111(d) guarantees that “there should be no gaps in control activities pertaining to stationary source emissions that pose any significant danger to public health or wel- fare.” S. Rep. No. 91–1196, p. 20 (1970). Reflecting that language, the majority calls Section 111(d) a “gap-filler.” Ante, at 5. It might also be thought of as a backstop or catch-all provision, protecting against pollutants that the NAAQS and HAP programs let go by. But the section is not, as the majority further claims, an “ancillary provision” or a statutory “backwater.” Ante, at 20, 26. That charac- terization is a non-sequitur. That something is a backstop does not make it a backwater. Even if they are needed only infrequently, see ante, at 6, 20, backstops can perform a critical function—and this one surely does. Again, Section 111(d) tells EPA that when a pollutant—like carbon diox- ide—is not regulated through other programs, EPA must undertake a further regulatory effort to control that sub- stance’s emission from existing stationary sources. In that way, Section 111(d) operates to ensure that the Act achieves comprehensive pollution control. Section 111 describes the prescribed regulatory effort in expansive terms. EPA must set for the relevant source Cite as: 597 U. S. ____ (2022) 7

(here, fossil-fuel-fired power plants) and the relevant pollu- tant (here, carbon dioxide) an emission level—more partic- ularly, “the degree of emission limitation achievable through the application of the best system of emission reduction which (taking into account the cost of achieving such reduction and any nonair quality health and environ- mental impact and energy requirements) the [EPA] Ad- ministrator determines has been adequately demon- strated.” §7411(a)(1). To take that language apart a bit, the provision instructs EPA to decide upon the “best system of emission reduction which . . . has been adequately demonstrated.” The provi- sion tells EPA, in making that determination, to take ac- count of both costs and varied “nonair” impacts (on health, the environment, and the supply of energy). And the provi- sion finally directs EPA to set the particular emissions limit achievable through use of the demonstrated “best system.” Taken as a whole, the section provides regulatory flexibility and discretion. It imposes, to be sure, meaningful con- straints: Take into account costs and nonair impacts, and make sure the best system has a proven track record.1 But the core command—go find the best system of emission re- duction—gives broad authority to EPA. If that flexibility is not apparent on the provision’s face, consider some dictionary definitions—supposedly a staple of this Court’s supposedly textualist method of reading stat- utes. A “system” is “a complex unity formed of many often diverse parts subject to a common plan or serving a common purpose.” Webster’s Third New International Dictionary 2322 (1971). Or again: a “system” is “[a]n organized and —————— 1 Those constraints have had real effect: They have led EPA in prior

rulemakings to exclude a number of pollution-control measures from the “best system of emission reduction.” See Brief for United States 49 (col- lecting citations). 8 WEST VIRGINIA v. EPA

coordinated method; a procedure.” American Heritage Dic- tionary 1768 (5th ed. 2018). The majority complains that a similar definition—cited to the Solicitor General’s brief but originally from another dictionary—is just too darn broad. Ante, at 28; see Brief for United States 31 (quoting Web- ster’s New International Dictionary 2562 (2d ed. 1959)). “[A]lmost anything” capable of reducing emissions, the ma- jority says, “could constitute such a ‘system’ ” of emission reduction. Ante, at 28. But that is rather the point. Con- gress used an obviously broad word (though surrounding it with constraints, see supra, at 7) to give EPA lots of latitude in deciding how to set emissions limits. And contra the ma- jority, a broad term is not the same thing as a “vague” one. Ante, at 18, 20, 28. A broad term is comprehensive, exten- sive, wide-ranging; a “vague” term is unclear, ambiguous, hazy. (Once again, dictionaries would tell the tale.) So EPA was quite right in stating in the Clean Power Plan that the “[p]lain meaning” of the term “system” in Section 111 refers to “a set of measures that work together to reduce emis- sions.” 80 Fed. Reg. 64762. Another of this Court’s opin- ions, involving a matter other than the bogeyman of envi- ronmental regulation, might have stopped there. For generation shifting fits comfortably within the con- ventional meaning of a “system of emission reduction.” Consider one of the most common mechanisms of genera- tion shifting: the use of a cap-and-trade scheme. Here is how the majority describes cap and trade: “Under such a scheme, sources that receive a reduction in their emissions can sell a credit representing the value of that reduction to others, who are able to count it toward their own applicable emissions caps.” Ante, at 8–9. Does that sound like a “sys- tem” to you? It does to me too. And it also has to this Court. In the past, we have explained that “[t]his type of ‘cap-and- trade’ system cuts costs while still reducing pollution to tar- get levels.” EPA v. EME Homer City Generation, L. P., 572 U. S. 489, 503, n. 10 (2014) (emphasis added). So what does Cite as: 597 U. S. ____ (2022) 9

the majority mean when it says that “[a]s a matter of defi- nitional possibilities, generation shifting can be described as a ‘system’ ”? Ante, at 28 (emphasis added; citation and some internal quotation marks omitted). Rarely has a stat- utory term so clearly applied. Other statutory provisions confirm the point. The Clean Air Act’s acid rain provision, for example, describes a cap- and-trade program as an “emission allocation and transfer system.” §7651(b) (emphasis added). So a “system,” accord- ing to the statute’s own usage, includes the kind of cap-and- trade mechanism that the Clean Power Plan relied on. And in a somewhat different way, the NAAQS provision shows that Section 111 encompasses such a regulatory technique. Under that provision, cap-and-trade schemes qualify as “control measures, means, or techniques” that state plans may use to reduce emissions. §7410(a)(2)(A). That lan- guage, of course, does not use the word “system.” But in specifying that cap and trade is allowable under the NAAQS program, the provision supports the same conclu- sion here—because Section 111 directs EPA to use “a pro- cedure similar to that provided by [the NAAQS].” §7411(d)(1). The majority discounts the relevance of both those provisions on the ground that they contemplate trad- ing systems only “as a means of complying with an already established emissions limit.” Ante, at 29 (emphasis in orig- inal). That is a distinction, to be sure. But to begin, it is far less of one than the majority thinks: In arguing that EPA’s claim of authority here would allow it to take the emissions limit as low as it wants, the majority ignores the varied constraints surrounding the “best system” language. See supra, at 7. And still more important for interpretive purposes, the distinction appears only in the majority’s opinion, not in any statutory language. That text, to the contrary, says to EPA: Do as you would do under the NAAQS and Acid Rain programs—go ahead and use cap and trade. 10 WEST VIRGINIA v. EPA

There is also a flipside point: Congress declined to include in Section 111 the restrictions on EPA’s authority contained in other Clean Air Act provisions. Most relevant here, quite a number of statutory sections confine EPA’s emissions- reduction efforts to technological controls—essentially, equipment or processes that can be put into place at a par- ticular facility. See ante, at 4 (describing those controls). So, for example, one provision tells EPA to set standards “reflect[ing] the greatest degree of emission reduction achievable through the application of technology.” §7521(a)(3)(A)(i). Others direct the use of the “best availa- ble retrofit technology,” or the “best available control tech- nology,” or the “maximum achievable control technology.” §§7491(b)(2)(A), (g)(2), 7475(a)(4), 7479(3), 7412(g)(2). There are still more. See, e.g., §§7411(h), 7511a(c)(7), 7651f(b)(2). None of those provisions would allow EPA to set emissions limits based on generation shifting, as the Agency acknowledges. See Brief for United States 32–33. But nothing like the language of those provisions is in- cluded in Section 111. That matters under normal rules of statutory interpretation. As Justice Scalia once wrote for the Court: “We do not lightly assume that Congress has omitted from its adopted text requirements that it nonethe- less intends to apply, and our reluctance is even greater when Congress has shown elsewhere in the same statute that it knows how to make such a requirement manifest.” Jama v. Immigration and Customs Enforcement, 543 U. S. 335, 341 (2005). Statutory history serves only to pile on: It shows that Congress has specifically declined to restrict EPA to technology-based controls in its regulation of existing sta- tionary sources. The key moment came in 1977, when Con- gress amended Section 111 to distinguish between new sources and existing ones. For new sources, EPA could se- lect only the “best technological system of continuous emis- sion reduction.” Clean Air Act Amendments, §109(c)(1)(A), Cite as: 597 U. S. ____ (2022) 11

91 Stat. 700 (emphasis added). But for existing sources, the word “technological” was struck out: EPA could select the “best system of continuous emission reduction.” Ibid. The House Report emphasized Congress’s deliberate choice: Whereas the standards set for new sources were to be based on “the best technological” controls, the “standards adopted for existing sources” were “to be based on available means of emission control (not necessarily technological).” H. R. Rep. No. 95–564, p. 129 (1977). The Report did not further explain the distinction. But presumably Congress gave EPA more flexibility over existing plants because imposing technological controls on old facilities is often not cost- effective.2 Thirteen years later, Congress followed up by de- leting from Section 111 the technological limitation apply- ing to new facilities. See Clean Air Act Amendments of 1990, §403(a), 104 Stat. 2631. Once again, then, Congress faced a choice: confine EPA to technological controls, or not. And replicating its earlier action for existing sources, Con- gress chose not. The majority breezes past that congressional choice on the ground that today’s opinion does not resolve whether EPA can regulate in some non-technological ways; instead, the opinion says only that the Clean Power Plan goes too —————— 2 The majority offers a theory for why Congress insisted on a techno-

logical system for new sources: It was, the majority says, to prevent EPA’s use of a particular kind of technological system (involving fuel switching) to achieve emissions reductions. See ante, at 22, n. 2, 23. To begin with: I don’t see how requiring EPA to select among technological systems precludes it from picking what the majority agrees is one such measure. See ante, at 4, 22, n. 2, 22–23. But more important, I can’t see why the majority’s explanation matters. Let’s assume the majority is right about Congress’s motive. The key point remains the same: What- ever that motive, Congress’s instruction to use technological systems ap- plied only to new sources, and not to existing ones. As to the latter, Con- gress allowed EPA more latitude: The Agency could use technological or non-technological methods, as it preferred. That distinction is what cre- ates interpretive difficulties for the majority—again, no matter why it arose. 12 WEST VIRGINIA v. EPA

far. See ante, at 30–31. That is a puzzling point. As an initial matter, it recharacterizes what this case has always been about. The Trump administration repealed the Clean Power Plan for one central reason: because (in its view) Sec- tion 111 confines EPA to facility-specific, technological measures. See 84 Fed. Reg. 32523–32529. In reviewing that repeal, the court below thus addressed that limit alone. See American Lung Assn. v. EPA, 985 F. 3d 914, 944 (CADC 2021). So add to the oddity of the Court’s declaring a de- funct regulation unlawful, see supra, at 4, the irregularity of its suggesting some kind of non-technological limit that no one (not EPA, not the parties, not the court below) has ever considered. More important here, both the nature and the statutory basis of that limit are left a mystery. If the majority is not distinguishing between technological con- trols and all others, what is it doing—and how far does its opinion constrain EPA? The majority makes no effort to say. And because that is so, the majority cannot even at- tempt to ground its limit in the statutory language. I’ve just shown that restricting EPA to technological controls is inconsistent with Section 111, especially when read in con- junction with other statutory provisions. And the majority provides no reason to think that its (possibly) different limit fares any better. Section 111 does not impose any con- straints—technological or otherwise—on EPA’s authority to regulate stationary sources (except for those stated, like cost). In somehow (and to some extent) saying otherwise, the majority flouts the statutory text. “Congress,” this Court has said, “knows to speak in plain terms when it wishes to circumscribe, and in capacious terms when it wishes to enlarge, agency discretion.” Ar- lington v. FCC, 569 U. S. 290, 296 (2013). In Section 111, Congress spoke in capacious terms. It knew that “without regulatory flexibility, changing circumstances and scien- tific developments would soon render the Clean Air Act ob- solete.” Massachusetts, 549 U. S., at 532. So the provision Cite as: 597 U. S. ____ (2022) 13

enables EPA to base emissions limits for existing stationary sources on the “best system.” That system may be techno- logical in nature; it may be whatever else the majority has in mind; or, most important here, it may be generation shifting. The statute does not care. And when Congress uses “expansive language” to authorize agency action, courts generally may not “impos[e] limits on [the] agency’s discretion.” Little Sisters of the Poor Saints Peter and Paul Home v. Pennsylvania, 591 U. S. ___, ___ (2020) (slip op., at 16). That constraint on judicial authority—that insistence on judicial modesty—should resolve this case. II The majority thinks not, contending that in “certain ex- traordinary cases”—of which this is one—courts should start off with “skepticism” that a broad delegation author- izes agency action. Ante, at 19. The majority labels that view the “major questions doctrine,” and claims to find sup- port for it in our caselaw. Ante, at 19–20, 28. But the rele- vant decisions do normal statutory interpretation: In them, the Court simply insisted that the text of a broad delega- tion, like any other statute, should be read in context, and with a modicum of common sense. Using that ordinary method, the decisions struck down agency actions (even though they plausibly fit within a delegation’s terms) for two principal reasons. First, an agency was operating far outside its traditional lane, so that it had no viable claim of expertise or experience. And second, the action, if allowed, would have conflicted with, or even wreaked havoc on, Con- gress’s broader design. In short, the assertion of delegated power was a misfit for both the agency and the statutory scheme. But that is not true here. The Clean Power Plan falls within EPA’s wheelhouse, and it fits perfectly—as I’ve just shown—with all the Clean Air Act’s provisions. That the Plan addresses major issues of public policy does not upend the analysis. Congress wanted EPA to do just that. 14 WEST VIRGINIA v. EPA

Section 111 entrusts important matters to EPA in the ex- pectation that the Agency will use that authority to combat pollution—and that courts will not interfere. A “[T]he words of a statute,” as the majority states, “must be read in their context and with a view to their place in the overall statutory scheme.” FDA v. Brown & Williamson To- bacco Corp., 529 U. S. 120, 133 (2000); see ante, at 16. We do not assess the meaning of a single word, phrase, or pro- vision in isolation; we also consider the overall statutory de- sign. And that is just as true of statutes broadly delegating power to agencies as of any other kind. In deciding on the scope of such a delegation, courts must assess how an agency action claimed to fall within the provision fits with other aspects of a statutory plan. So too, a court “must be guided to a degree by common sense as to the manner in which Congress is likely to dele- gate.” Brown & Williamson, 529 U. S., at 133. Assume that a policy decision, like this one, is a matter of significant “economic and political magnitude.” Ibid. We know that Congress delegates such decisions to agencies all the time— and often via broadly framed provisions like Section 111. See infra, at 29–31. But Congress does so in a sensible way. To decide whether an agency action goes beyond what Con- gress wanted, courts must assess (among other potentially relevant factors) the nature of the regulation, the nature of the agency, and the relationship of the two to each other. See, e.g., Barnhart v. Walton, 535 U. S. 212, 222 (2002). In particular, we have understood, Congress does not usually grant agencies the authority to decide significant issues on which they have no particular expertise. So when there is a mismatch between the agency’s usual portfolio and a given assertion of power, courts have reason to question whether Congress intended a delegation to go so far. The majority today goes beyond those sensible principles. Cite as: 597 U. S. ____ (2022) 15

It announces the arrival of the “major questions doctrine,” which replaces normal text-in-context statutory interpreta- tion with some tougher-to-satisfy set of rules. Ante, at 16– 31. Apparently, there is now a two-step inquiry. First, a court must decide, by looking at some panoply of factors, whether agency action presents an “extraordinary case[ ].” Ante, at 17; see ante, at 20–28. If it does, the agency “must point to clear congressional authorization for the power it claims,” someplace over and above the normal statutory ba- sis we require. Ante, at 19 (internal quotation marks omit- ted); see ante, at 28–31. The result is statutory interpreta- tion of an unusual kind. It is not until page 28 of a 31-page opinion that the majority begins to seriously discuss the meaning of Section 111. And even then, it does not address straight-up what should be the question: Does the text of that provision, when read in context and with a common- sense awareness of how Congress delegates, authorize the agency action here? The majority claims it is just following precedent, but that is not so. The Court has never even used the term “ma- jor questions doctrine” before. And in the relevant cases, the Court has done statutory construction of a familiar sort. It has looked to the text of a delegation. It has addressed how an agency’s view of that text works—or fails to do so— in the context of a broader statutory scheme. And it has asked, in a common-sensical (or call it purposive) vein, about what Congress would have made of the agency’s view—otherwise said, whether Congress would naturally have delegated authority over some important question to the agency, given its expertise and experience. In short, in assessing the scope of a delegation, the Court has consid- ered—without multiple steps, triggers, or special presump- tions—the fit between the power claimed, the agency claim- ing it, and the broader statutory design. The key case here is FDA v. Brown & Williamson. There, the Food and Drug Administration (FDA) asserted that its 16 WEST VIRGINIA v. EPA

power to regulate “drugs” and “devices” extended to tobacco products. The claim had something to it: FDA has broad authority over “drugs” and drug-delivery “devices,” and the definitions of those terms could be read to encompass nico- tine and cigarettes. But the asserted authority “simply [did] not fit” the overall statutory scheme. 529 U. S., at 143. FDA’s governing statute required the agency to ensure that regulated products were “safe” to be marketed—but there was no making tobacco products safe in the usual sense. Id., at 133–143. So FDA would have had to reinterpret what it meant to be “safe,” or else ban tobacco products al- together. Ibid. Both options, the Court thought, were pre- posterous. Until the agency action at issue, tobacco prod- ucts hadn’t been spoken of in the same breath as pharmaceuticals (FDA’s paradigmatic regulated product). And Congress had created in several statutes a “distinct regulatory scheme” for tobacco, not involving FDA. Id., at 155–156. So all the evidence was that Congress had never meant for FDA to have any—let alone total—control over the tobacco industry, with its “unique political history.” Id., at 159. Again, there was “simply” a lack of “fit” between the regulation at issue, the agency in question, and the broader statutory scheme. Id., at 143. The majority’s effort to find support in Brown & William- son for its interpretive approach fails. See ante, at 19. It may be helpful here to quote the full sentence that the ma- jority quotes half of. “In extraordinary cases,” the Court stated, “there may be reason to hesitate before concluding that Congress has intended such an implicit delegation.” 529 U. S., at 159. For anyone familiar with this Court’s Chevron doctrine, that language will ring a bell. The Court was saying only—and it was elsewhere explicit on this point—that there was reason to hesitate before giving FDA’s position Chevron deference. See id., at 132–133, 159–161. And what was that reason? The Court went on to explain that it would not defer to FDA because it read Cite as: 597 U. S. ____ (2022) 17

the relevant statutory provisions as negating the agency’s claimed authority. See id., at 160 (“[W]e are obliged to defer not to the agency’s expansive construction of the statute, but to Congress’ consistent judgment to deny the FDA this power”); id., at 133 (finding at Chevron’s first step that “Congress has directly spoken to the issue here and pre- cluded the FDA’s” asserted power). In reaching that con- clusion, the Court relied (as I’ve just explained) not on any special “clear authorization” demand, but on normal princi- ples of statutory interpretation: look at the text, view it in context, and use what the Court called some “common sense” about how Congress delegates. Ibid. That is how courts are to decide, in the majority’s language, whether an agency has asserted a “highly consequential power beyond what Congress could reasonably be understood to have granted.” Ante, at 20. The Court has applied the same kind of analysis in sub- sequent cases—holding in each that an agency exceeded the scope of a broadly framed delegation when it operated out- side the sphere of its expertise, in a way that warped the statutory text or structure. In Gonzales v. Oregon, 546 U. S. 243 (2006), we rejected the Attorney General’s assertion of authority (under a broad “public interest” standard) to re- scind doctors’ registrations for facilitating assisted suicide, even in States where doing so was legal. See id., at 243, 248–249, 261–275. We doubted Congress would have dele- gated such a “quintessentially medical judgment[ ]” to “an executive official who lacks medical expertise.” Id., at 266– 267. And we pointed to statutory provisions in which Con- gress—in opposition to the claimed power—had “painstak- ingly described the Attorney General’s limited authority” to deregister physicians. Id., at 262.3 —————— 3 Similarly, in King v. Burwell, 576 U. S. 473 (2015), we relied on

Brown & Williamson in declining to defer to the Internal Revenue Ser- vice’s construction of the Affordable Care Act. We thought it highly “un- likely that Congress would have delegated” an important decision about 18 WEST VIRGINIA v. EPA

Later, in Utility Air Regulatory Group v. EPA, 573 U. S. 302 (2014), the Court relied on similar reasoning to reject EPA’s efforts to regulate “millions of small” and previously unregulated sources of emissions—“including retail stores, offices, apartment buildings, shopping centers, schools, and churches.” Id., at 328. Key to that decision was the Court’s view that reading the delegation so expansively would be “inconsistent with” the statute’s broader “structure and de- sign.” Id., at 321. The Court explained that allowing the agency action to proceed would necessitate the “rewriting” of other “unambiguous statutory terms”—indeed, of “pre- cise numerical thresholds.” Id., at 321, 325–326. (In quot- ing one cryptic sentence of Utility Air as supporting its new approach, see ante, at 19, the majority ignores the nine pre- ceding pages of analysis of the statute’s text and context, see 573 U. S., at 315–324.) And last Term, the Court concluded that the Centers for Disease Control and Prevention (CDC) lacked the power to impose a nationwide eviction moratorium. Alabama Assn. of Realtors v. Department of Health and Human Servs., 594 U. S. ___, ___–___ (2021) (slip op., at 5–7). The Court held that other statutory language made it a “stretch” to read the relied-on delegation as covering the CDC’s action. Id., at ___ (slip op., at 6). And the Court raised an eyebrow at the thought of the CDC “intrud[ing]” into “the landlord-ten- ant relationship”—a matter outside the CDC’s usual “do- main.” Ibid.4 The eyebrow-raise is indeed a consistent presence in these cases, responding to something the Court found anomalous—looked at from Congress’s point of view—in a —————— healthcare pricing to an agency with “no expertise in crafting health in- surance policy.” 576 U. S., at 486. 4 Not every Justice, of course, agreed with the Court’s conclusions in

the above-discussed cases; to be frank, I dissented in a couple. But what matters here is the analysis those decisions undertook—and how, as I’ll describe, it supports EPA’s Clean Power Plan. Cite as: 597 U. S. ____ (2022) 19

particular agency’s exercise of authority. In each case, the Court thought, the agency had strayed out of its lane, to an area where it had neither expertise nor experience. The At- torney General making healthcare policy, the regulator of pharmaceutical concerns deciding the fate of the tobacco in- dustry, and so on. And in each case, the proof that the agency had roamed too far afield lay in the statutory scheme itself. The agency action collided with other statu- tory provisions; if the former were allowed, the latter could not mean what they said or could not work as intended. FDA having to declare tobacco “safe” to avoid shutting down an industry; or EPA having literally to change hard num- bers contained in the Clean Air Act. There, according to the Court, the statutory framework was “not designed to grant” the authority claimed. Utility Air, 573 U. S., at 324. The agency’s “singular” assertion of power “would render the statute unrecognizable to the Congress” that wrote it. Ibid. (internal quotation marks omitted). B The Court today faces no such singular assertion of agency power. As I have already explained, nothing in the Clean Air Act (or, for that matter, any other statute) con- flicts with EPA’s reading of Section 111. Notably, the ma- jority does not dispute that point. Of course, it views Sec- tion 111 (if for unexplained reasons) as less clear than I do. Compare ante, at 28, with supra, at 7–9. But nowhere does the majority provide evidence from within the statute itself that the Clean Power Plan conflicts with or undermines Congress’s design. That fact alone makes this case differ- ent from all the cases described above. As to the other crit- ical matter in those cases—is the agency operating outside its sphere of expertise?—the majority at least tries to say something. It claims EPA has no “comparative expertise” in “balancing the many vital considerations of national pol- icy” implicated in regulating electricity sources. Ante, at 20 WEST VIRGINIA v. EPA

25–26. But that is wrong. Start with what this Court has said before on the subject, reflecting Congress’s view of the matter. About a decade ago, we recognized that Congress had “delegated to EPA” in Section 111 “the decision whether and how to regulate carbon-dioxide emissions from powerplants.” American Elec. Power, 564 U. S., at 426. To stress the key word (be- cause the majority seems to miss it, see ante, at 26–27): not merely “whether” but also “how.” In making that delega- tion, we explained, Congress knew well what it was doing. Regulating power plant emissions is a complex undertak- ing. To do it right requires “informed assessment of com- peting interests”: “Along with the environmental benefit po- tentially achievable, our Nation’s energy needs and the possibility of economic disruption must weigh in the bal- ance.” 564 U. S., at 427; see §7411(a)(1) (instructing EPA to consider “energy requirements,” “cost,” and other fac- tors). Congress specifically “entrust[ed] such complex bal- ancing to EPA,” because that “expert agency” has the needed “scientific, economic, and technological resources” to carry it out. 564 U. S., at 427–428. So the balancing—in- cluding of the Nation’s “energy requirements”—that the majority says EPA has no “comparative expertise” in? §7411(a)(1); ante, at 25. We explained 11 short years ago, citing Congress, that it was smack in the middle of EPA’s wheelhouse. And we were right. Consider the Clean Power Plan’s component parts—let’s call them the what, who, and how— to see the rule’s normalcy. The “what” is the subject matter of the Plan: carbon dioxide emissions. This Court has al- ready found that those emissions fall within EPA’s domain. We said then: “[T]here is nothing counterintuitive to the no- tion that EPA can curtail the emission of substances that are putting the global climate out of kilter.” Massachusetts, 549 U. S., at 531. This is not the Attorney General regulat- ing medical care, or even the CDC regulating landlord- Cite as: 597 U. S. ____ (2022) 21

tenant relations. It is EPA (that’s the Environmental Pro- tection Agency, in case the majority forgot) acting to ad- dress the greatest environmental challenge of our time. So too, there is nothing special about the Plan’s “who”: fossil- fuel-fired power plants. In Utility Air, we thought EPA’s regulation of churches and schools highly unusual. See su- pra, at 18. But fossil-fuel-fired plants? Those plants pol- lute—a lot—and so they have long lived under the watchful eye of EPA. That was true even before EPA began regulat- ing carbon dioxide. See Train v. Natural Resources Defense Council, Inc., 421 U. S. 60, 78 (1975). Finally, the “how” of generation shifting creates no mis- match with EPA’s expertise. As the Plan noted, generation shifting has a well-established pedigree as a tool for reduc- ing pollution; even putting aside other federal regulation, see infra, at 25–26, both state regulators and power plants themselves have long used it to attain environmental goals. See 80 Fed. Reg. 64664; Brief for Power Company Respond- ents 47; see also S. Breyer, Regulation and Its Reform 444, n. 1 (1982) (citing literature on the subject from the 1970s). The technique is, so to speak, a tool in the pollution-control toolbox. And that toolbox is the one EPA uses. So that Agency, more than any other, has the desired “comparative expertise.” Ante, at 25. The majority cannot contest that point frontally: It knows that cap and trade and similar mechanisms are an ordinary part of modern environmental regulation. Instead, the majority protests that Congress would not have wanted EPA to “dictat[e],” through genera- tion shifting, the “mix of energy sources nationwide.” Ante, at 26. But that statement reflects a misunderstanding of how the electricity market works. Every regulation of power plants—even the most conventional, facility- specific controls—“dictat[es]” the national energy mix to one or another degree. That result follows because regula- tions affect costs, and the electrical grid works by taking up 22 WEST VIRGINIA v. EPA

energy from low-cost providers before high-cost ones. Con- sider an example: Suppose EPA requires coal-fired plants to use carbon-capture technology. That action increases those plants’ costs, and automatically (by virtue of the way the grid operates) reduces their share of the electricity mar- ket. So EPA is always controlling the mix of energy sources. In that sense (though the term has taken on a more special- ized meaning), everything EPA does is “generation shift- ing.” The majority’s idea that EPA has no warrant to direct such a shift just indicates that courts sometimes do not re- ally get regulation.5 Why, then, be “skeptic[al]” of EPA’s exercise of authority? Ante, at 28. When there is no misfit, of the kind apparent in our precedents, between the regulation, the agency, and the statutory design? Although the majority offers a flurry of complaints, they come down in the end to this: The Clean Power Plan is a big new thing, issued under a minor statu- tory provision. See ante, at 20, 24, 26 (labeling the Plan “transformative” and “unprecedented” and calling Section 111(d) an “ancillary” “backwater”). I have already ad- dressed the back half of that argument: In fact, there is —————— 5 The majority’s only response to the argument above similarly reveals

a misperception as to the practical impact of different regulatory tech- niques. According to the majority, there is an “obvious difference” be- tween changing the energy mix by conventional technological regulation and doing so by measures like cap and trade. Ante, at 27, n. 4. But in fact there is not. As I’ll detail later, generation shifting can effect a sig- nificant—or instead an insignificant—change in the energy mix; and the same is true of technological regulations. See infra, at 24–25. It all de- pends on the specifics: There is no necessary connection (in either direc- tion) between the kind of regulation and the magnitude of its effect. For example, a rule requiring the use of carbon-capture technology would have shifted far more electricity production from coal-fired plants than the Clean Power Plan would have. See ibid. In suggesting that cap-and- trade programs are somehow more suspect, the majority merely serves to disadvantage what is often the smartest kind of regulation: market- based programs that achieve the biggest bang for the buck. That is why so many power companies are on EPA’s side in this litigation. Cite as: 597 U. S. ____ (2022) 23

nothing insignificant about Section 111(d), which was in- tended to ensure that EPA would limit existing stationary sources’ emissions of otherwise unregulated pollutants (however few or many there were). See supra, at 6. And the front half of the argument doesn’t work either. The Clean Power Plan was not so big. It was not so new. And to the extent it was either, that should not matter. As to bigness—well, events have proved the opposite: The Clean Power Plan, we now know, would have had little or no impact. The Trump administration’s repeal of the Plan created a kind of controlled experiment: The Plan’s “magni- tude” (ante, at 24) could be measured by seeing how far short the industry fell of the Plan’s nationwide emissions target. Except that turned out to be the wrong question, because the industry didn’t fall short of the Plan’s goal; ra- ther, the industry exceeded that target, all on its own. See App. 265 (declaration of EPA official). And it did so mainly through the generation-shifting techniques that the Plan called for. See ibid.; Brief for United States 47. In effect, the Plan predicted market behavior, rather than altered it (as regulations usually do). Cf. Utility Air, 573 U. S., at 321–322 (discussing the “calamitous consequences” of the EPA approach there under review). And that fact has been understood for some years. At the time of the repeal, the Trump administration explained that “there [was] likely to be no difference between a world where the [Clean Power Plan was] implemented and one where it [was] not.” 84 Fed. Reg. 32561.6 It is small wonder, then, that the power —————— 6 Even when the Clean Power Plan was first issued, its projected im-

pact was far less than what the majority implies. The majority states, for example, that the rule would have “reduce[d] GDP by at least a tril- lion 2009 dollars by 2040.” Ante, at 10. That sounds like a lot, but it is in fact “equivalent to changes of a few tenths of 1 percent from baseline.” Dept. of Energy, Analysis of the Impacts of the Clean Power Plan 63–64 (2015). And the “billions of dollars in compliance costs” the majority highlights were vastly outweighed by the Plan’s projected benefits. Ante, at 10; see 80 Fed. Reg. 64679 (anticipating $5–$8 billion in costs and 24 WEST VIRGINIA v. EPA

industry overwhelmingly supports EPA in this case. See Brief for Power Company Respondents 2–3. In the regu- lated parties’ view, the rule aimed to achieve what most power companies also want: substantial reductions in car- bon dioxide emissions accomplished in a cost-effective way while maintaining a reliable electricity market. See id., at 26–27, 38, 41–42. The majority thus pivots to the massive consequences generation shifting could produce—but that claim fares just as poorly. On EPA’s view of its own authority, the majority worries, some future rule might “forc[e] coal plants to ‘shift’ away virtually all of their generation—i.e., to cease making power altogether.” Ante, at 24. But looking at the text of Section 111(d) might here come in handy. For the statute imposes, as already shown, a set of constraints—particu- larly involving costs and energy needs—that would pre- clude so extreme a regulation. See Brief for United States 41–42 (conceding the point); supra, at 7. And if the majority thinks those constraints do not really constrain, then it has a much bigger problem. For “traditional” technological con- trols, of the kind the majority approves, can have equally dramatic effects. Ante, at 23. Take, for example, the “fuel- switching” regulation the majority mentions. Ibid. Such a rule does just what you might think: It requires a plant to burn a different kind of fuel—say, natural gas instead of coal. So it too can significantly “restructur[e] the Nation’s overall mix of electricity generation.” Ante, at 16. Or take an even more technological-sounding approach: the use of carbon-capture equipment. Order the installation of that equipment, the Trump administration concluded, and the “exorbitant” costs “would almost certainly force the closure” of all affected “coal-fired power plants.” 84 Fed. Reg. 32548. —————— $32–$54 billion in benefits by 2030); see also EPA, Regulatory Impact Analysis for the Clean Power Plan Final Rule 6–35 (2015) (estimating that by 2030 jobs gained from the Plan would be some two or three times greater than jobs lost). Cite as: 597 U. S. ____ (2022) 25

The point is a simple one: If generation shifting can go big, so too can technological controls (assuming, once again, that the statute’s text is ignored). The problem (if any ex- ists) is not with the channel, but with the volume.7 The majority’s claim about the Clean Power Plan’s nov- elty—the most fleshed-out part of today’s opinion, see ante, at 20–24—is also exaggerated. As EPA explained when it issued the Clean Power Plan, an earlier Section 111(d) reg- ulation had determined that a cap-and-trade program was the “best system of emission reduction” for mercury. 70 Fed. Reg. 28616–28621 (2005); see 80 Fed. Reg. 64772. In the majority’s view, that rule was different because the “ac- tual emission cap” for the contemplated cap-and-trade scheme was based on the use of a plant-specific technol- ogy—namely, wet scrubbers. Ante, at 21 (internal quota- tion marks omitted). But the approval of cap and trade al- lowed EPA to make the emissions limits more stringent than it otherwise could have, because EPA knew that plants unable to cost-effectively install scrubbers could in- stead meet the limits through generation shifting. See 70 Fed. Reg. 28619. EPA could have designed the Clean Power Plan in the same way—say, by setting emissions limits based on carbon-capture technology, with the expectation that many plants would avail themselves of an approved cap-and-trade program instead. The majority gives no rea- son to think Section 111(d) allows that approach but disal- lows the Clean Power Plan. In both, generation shifting is —————— 7 The majority dismisses these hypotheticals as fantastical, protesting

that “EPA has never ordered anything remotely like [them], and we doubt it could.” Ante, at 24, n. 3. But that’s just the point. EPA hasn’t forced the elimination of coal plants—whether through technological controls or generation shifting—because the statutory constraints pre- vent it from doing so. The majority offers no reason to think that those constraints suffice for the measures it approves (fuel switching and car- bon capture) but not for the measure it rejects (generation shifting). Ei- ther the constraints are enough or they are not. The majority cannot have it both ways. 26 WEST VIRGINIA v. EPA

operating to increase the strictness of emissions limits. And the mercury rule itself was rooted in precedent. A decade earlier, EPA had determined that States could com- ply with a Section 111(d) regulation for municipal waste combustors by establishing cap-and-trade programs. See 40 CFR §§60.30a, 60.33b(d)(2) (1996). And beyond Section 111(d), trading and other tools of generation shifting be- come still more common. For decades, EPA has relied on those pollution-control techniques in rules covering new internal-combustion engines under Section 111(b), sources of nitrogen oxide under the NAAQS program, and motor ve- hicles under Section 202(a). See 73 Fed. Reg. 3595 (2008); 71 Fed. Reg. 39159 (2006); 63 Fed. Reg. 57358–57359 (1998); 48 Fed. Reg. 33456 (1983); see also Brief for Richard L. Revesz as Amicus Curiae 24–29 (collecting similar rules). No doubt the majority is right that scrubbers and other “add-on controls” are “more traditional air pollution control measures.” Ante, at 23. EPA readily acknowledged that fact in developing the Clean Power Plan. But the idea that the Plan’s reliance on generation shifting effected some kind of revolution in power-plant pollution control? No. As I’ve noted before, power plants themselves use that method. State environmental regulators use that method. And EPA has used that method, including under the statutory provi- sion invoked here. In any event, newness might be perfectly legitimate— even required—from Congress’s point of view. I do not dis- pute that an agency’s longstanding practice may inform a court’s interpretation of a statute delegating the agency power. See ante, at 20–21. But it is equally true, as Brown & Williamson recognized, that agency practices are “not carved in stone.” 529 U. S., at 156–157 (internal quotation marks omitted). Congress makes broad delegations in part so that agencies can “adapt their rules and policies to the demands of changing circumstances.” Id., at 157. To keep Cite as: 597 U. S. ____ (2022) 27

faith with that congressional choice, courts must give agen- cies “ample latitude” to revisit, rethink, and revise their regulatory approaches. Ibid. So it is here. Section 111(d) was written, as I’ve shown, to give EPA plenty of leeway. See supra, at 6–8. The enacting Congress told EPA to pick the “best system of emission reduction” (taking into account various factors). In selecting those words, Congress under- stood—it had to—that the “best system” would change over time. Congress wanted and instructed EPA to keep up. To ensure the statute’s continued effectiveness, the “best sys- tem” should evolve as circumstances evolved—in a way Congress knew it couldn’t then know. See Massachusetts, 549 U. S., at 532. EPA followed those statutory directions to the letter when it issued the Clean Power Plan. It se- lected a system (as the regulated parties agree) that achieved greater emissions reductions at lower cost than any technological alternative could have, while maintain- ing a reliable electricity market. Even if that system was novel, it was in EPA’s view better—actually, “best.” So it was the system that accorded with the enacting Congress’s choice. And contra the majority, it is that Congress’s choice which counts, not any later one’s. The majority says it “can- not ignore” that Congress in recent years has “considered and rejected” cap-and-trade schemes. Ante, at 27–28. But under normal principles of statutory construction, the ma- jority should ignore that fact (just as I should ignore that Congress failed to enact bills barring EPA from implement- ing the Clean Power Plan). As we have explained time and again, failed legislation “offers a particularly dangerous ba- sis on which to rest an interpretation of an existing law a different and earlier Congress” adopted. Bostock v. Clayton County, 590 U. S. ___, ___ (2020) (slip op., at 20) (internal quotation marks omitted); see Sullivan v. Finkelstein, 496 U. S. 617, 632 (1990) (Scalia, J., concurring in part) (“Argu- ments based on subsequent legislative history” should “not 28 WEST VIRGINIA v. EPA

be taken seriously, not even in a footnote”). Return to Brown & Williamson, which all agree is the key case in this sphere. It disclaimed any reliance on “Congress’ failure” to grant FDA jurisdiction over tobacco. 529 U. S., at 155. In- stead, the Court focused on the statutes Congress “ha[d] en- acted,” which created “a distinct regulatory scheme” for to- bacco, incompatible with FDA’s. Ibid. (emphasis added). Here, as I’ve shown and the majority effectively concedes, there is nothing equivalent. See supra, at 9–12. Search high and low, nothing in current law conflicts with, or oth- erwise casts doubt on, the Clean Power Plan. That leaves the Court in much the same place it was when deciding Massachusetts v. EPA. Said the Court then: “That subse- quent Congresses have eschewed enacting binding emis- sions limitations to combat global warming tells us nothing about what Congress meant” when it enacted the Clean Air Act. 549 U. S., at 529–530. And so the Court recognized EPA’s authority to regulate carbon dioxide. But that Court was not this Court; and this Court deprives EPA of the au- thority Congress gave it in Section 111(d) to respond to the same environmental danger. III Some years ago, I remarked that “[w]e’re all textualists now.” Harvard Law School, The Antonin Scalia Lecture Se- ries: A Dialogue with Justice Elena Kagan on the Reading of Statutes (Nov. 25, 2015). It seems I was wrong. The cur- rent Court is textualist only when being so suits it. When that method would frustrate broader goals, special canons like the “major questions doctrine” magically appear as get- out-of-text-free cards.8 Today, one of those broader goals —————— 8 The majority opinion at least addresses the statute’s text, though

overstating its ambiguity and approaching the action taken under it with unwarranted “skepticism.” Ante, at 28; see ante, at 28–31. The concur- rence, by contrast, concludes that the Clean Air Act does not clearly enough authorize EPA’s Plan without ever citing the statutory text. See Cite as: 597 U. S. ____ (2022) 29

makes itself clear: Prevent agencies from doing important work, even though that is what Congress directed. That anti-administrative-state stance shows up in the majority opinion, and it suffuses the concurrence. See ante, at 19, 25–26; e.g., ante, at 3–6 (GORSUCH, J., concurring). The kind of agency delegations at issue here go all the way back to this Nation’s founding. “[T]he founding era,” scholars have shown, “wasn’t concerned about delegation.” E. Posner & A. Vermeule, Interring the Nondelegation Doc- trine, 69 U. Chi. L. Rev. 1721, 1734 (2002) (Posner & Ver- meule). The records of the Constitutional Convention, the ratification debates, the Federalist—none of them suggests any significant limit on Congress’s capacity to delegate pol- icymaking authority to the Executive Branch. And neither does any early practice. The very first Congress gave sweeping authority to the Executive Branch to resolve some of the day’s most pressing problems, including questions of “territorial administration,” “Indian affairs,” “foreign and domestic debt,” “military service,” and “the federal courts.” J. Mortenson & N. Bagley, Delegation at the Founding, 121 Colum. L. Rev. 277, 349 (2021) (Mortenson & Bagley). That Congress, to use a few examples, gave the Executive power to devise a licensing scheme for trading with Indians; to craft appropriate laws for the Territories; and to decide how to pay down the (potentially ruinous) national debt. See id., at 334–338, 340–342, 344–345; C. Chabot, The Lost History of Delegation at the Founding, 56 Ga. L. Rev. 81, 113–134 (2021) (Chabot). Barely anyone objected on delegation grounds. See Mortenson & Bagley 281–282, 332, 339; Chabot 117–119; Posner & Vermeule 1733–1736. It is not surprising that Congress has always delegated, —————— ante, at 13–16. Nowhere will you find the concurrence ask: What does the phrase “best system of emission reduction” mean? §7411(a)(1). So much for “begin[ning], as we must, with a careful examination of the statutory text.” Henson v. Santander Consumer USA Inc., 582 U. S. 79, ___ (2017) (slip op., at 3). 30 WEST VIRGINIA v. EPA

and continues to do so—including on important policy is- sues. As this Court has recognized, it is often “unreasona- ble and impracticable” for Congress to do anything else. American Power & Light Co. v. SEC, 329 U. S. 90, 105 (1946). In all times, but ever more in “our increasingly com- plex society,” the Legislature “simply cannot do its job ab- sent an ability to delegate power under broad general direc- tives.” Mistretta v. United States, 488 U. S. 361, 372 (1989). Consider just two reasons why. First, Members of Congress often don’t know enough— and know they don’t know enough—to regulate sensibly on an issue. Of course, Members can and do provide overall direction. But then they rely, as all of us rely in our daily lives, on people with greater expertise and experience. Those people are found in agencies. Congress looks to them to make specific judgments about how to achieve its more general objectives. And it does so especially, though by no means exclusively, when an issue has a scientific or tech- nical dimension. Why wouldn’t Congress instruct EPA to select “the best system of emission reduction,” rather than try to choose that system itself? Congress knows that sys- tems of emission reduction lie not in its own but in EPA’s “unique expertise.” Martin v. Occupational Safety and Health Review Comm’n, 499 U. S. 144, 151 (1991). Second and relatedly, Members of Congress often can’t know enough—and again, know they can’t—to keep regula- tory schemes working across time. Congress usually can’t predict the future—can’t anticipate changing circum- stances and the way they will affect varied regulatory tech- niques. Nor can Congress (realistically) keep track of and respond to fast-flowing developments as they occur. Once again, that is most obviously true when it comes to scientific and technical matters. The “best system of emission reduc- tion” is not today what it was yesterday, and will surely be something different tomorrow. So for this reason too, a ra- tional Congress delegates. It enables an agency to adapt Cite as: 597 U. S. ____ (2022) 31

old regulatory approaches to new times, to ensure that a statutory program remains effective. See, e.g., National Federation of Independent Business v. OSHA, 595 U. S. ___, ___ (2022) (BREYER, SOTOMAYOR, and KAGAN, JJ., dissent- ing) (slip op., at 9) (observing that a statute’s broad lan- guage was meant to ensure that an agency had “the tools needed to confront emerging dangers”). Over time, the administrative delegations Congress has made have helped to build a modern Nation. Congress wanted fewer workers killed in industrial accidents. It wanted to prevent plane crashes, and reduce the deadliness of car wrecks. It wanted to ensure that consumer products didn’t catch fire. It wanted to stop the routine adulteration of food and improve the safety and efficacy of medications. And it wanted cleaner air and water. If an American could go back in time, she might be astonished by how much pro- gress has occurred in all those areas. It didn’t happen through legislation alone. It happened because Congress gave broad-ranging powers to administrative agencies, and those agencies then filled in—rule by rule by rule—Con- gress’s policy outlines. This Court has historically known enough not to get in the way. Maybe the best explanation of why comes from Justice Scalia. See Mistretta, 488 U. S., at 415–416 (dis- senting opinion). The context was somewhat different. He was responding to an argument that Congress could not constitutionally delegate broad policymaking authority; here, the Court reads a delegation with unwarranted skep- ticism, and thereby artificially constrains its scope. But Justice Scalia’s reasoning remains on point. He started with the inevitability of delegations: “[S]ome judgments in- volving policy considerations,” he stated, “must be left to [administrative] officers.” Id., at 415. Then he explained why courts should not try to seriously police those delega- tions, barring—or, I’ll add, narrowing—some on the ground that they went too far. The scope of delegations, he said, 32 WEST VIRGINIA v. EPA

“must be fixed according to common sense and the in- herent necessities of the governmental co-ordination. Since Congress is no less endowed with common sense than we are, and better equipped to inform itself of the necessities of government; and since the factors bear- ing upon those necessities are both multifarious and (in the nonpartisan sense) highly political . . . it is small wonder that we have almost never felt qualified to second-guess Congress regarding the permissible de- gree of policy judgment that can be left to those execut- ing or applying the law.” Id., at 416 (internal quotation marks omitted). In short, when it comes to delegations, there are good rea- sons for Congress (within extremely broad limits) to get to call the shots. Congress knows about how government works in ways courts don’t. More specifically, Congress knows what mix of legislative and administrative action conduces to good policy. Courts should be modest. Today, the Court is not. Section 111, most naturally read, authorizes EPA to develop the Clean Power Plan—in other words, to decide that generation shifting is the “best system of emission reduction” for power plants churning out carbon dioxide. Evaluating systems of emission reduc- tion is what EPA does. And nothing in the rest of the Clean Air Act, or any other statute, suggests that Congress did not mean for the delegation it wrote to go as far as the text says. In rewriting that text, the Court substitutes its own ideas about delegations for Congress’s. And that means the Court substitutes its own ideas about policymaking for Con- gress’s. The Court will not allow the Clean Air Act to work as Congress instructed. The Court, rather than Congress, will decide how much regulation is too much. The subject matter of the regulation here makes the Court’s intervention all the more troubling. Whatever else this Court may know about, it does not have a clue about Cite as: 597 U. S. ____ (2022) 33

how to address climate change. And let’s say the obvious: The stakes here are high. Yet the Court today prevents congressionally authorized agency action to curb power plants’ carbon dioxide emissions. The Court appoints it- self—instead of Congress or the expert agency—the decision- maker on climate policy. I cannot think of many things more frightening. Respectfully, I dissent.