American Lung Association v. EPA

United States Court of Appeals
FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued October 8, 2020 Decided January 19, 2021

No. 19-1140

AMERICAN LUNG ASSOCIATION AND AMERICAN PUBLIC
HEALTH ASSOCIATION,
PETITIONERS

v.

ENVIRONMENTAL PROTECTION AGENCY AND ANDREW
WHEELER, ADMINISTRATOR,
RESPONDENTS

AEP GENERATING COMPANY, ET AL.,
INTERVENORS

Consolidated with 19-1165, 19-1166, 19-1173, 19-1175,
19-1176, 19-1177, 19-1179, 19-1185, 19-1186, 19-1187,
19-1188

On Petitions for Review of a Final Action
of the Environmental Protection Agency

Steven C. Wu, Deputy Solicitor General, Office of the
Attorney General for the State of New York, argued the cause
for the State and Municipal petitioners and intervenor Nevada.
 2
With him on the briefs were Letitia James, Attorney General,
Barbara D. Underwood, Solicitor General, Matthew W.
Grieco, Assistant Solicitor General, Michael J. Myers, Senior
Counsel, Andrew G. Frank, Assistant Attorney General of
Counsel, Xavier Becerra, Attorney General, Office of the
Attorney General for the State of California, Robert W. Byrne,
Senior Assistant Attorney General, David A. Zonana,
Supervising Deputy Attorney General, Jonathan A. Wiener, M.
Elaine Meckenstock, Timothy E. Sullivan, Elizabeth B. Rumsey,
and Theodore A.B. McCombs, Deputy Attorneys General,
William Tong, Attorney General, Office of the Attorney
General for the State of Connecticut, Matthew I. Levine and
Scott N. Koschwitz, Assistant Attorneys General, Kathleen
Jennings, Attorney General, Office of the Attorney General for
the State of Delaware, Valerie S. Edge, Deputy Attorney
General, Philip J. Weiser, Attorney General, Office of the
Attorney General for the State of Colorado, Eric R. Olson,
Solicitor General, Robyn L. Wille, Senior Assistant Attorney
General, Clare E. Connors, Attorney General, Office of the
Attorney General for the State of Hawaii, William F. Cooper,
Deputy Attorney General, Aaron M. Frey, Attorney General,
Office of the Attorney General for the State of Maine, Laura
E. Jensen, Assistant Attorney General, Brian E. Frosh,
Attorney General, Office of the Attorney General for the State
of Maryland, John B. Howard, Jr., Joshua M. Segal, and
Steven J. Goldstein, Special Assistant Attorneys General,
Maura Healey, Attorney General, Office of the Attorney
General for the Commonwealth of Massachusetts, Melissa A.
Hoffer and Christophe Courchesne, Assistant Attorneys
General, Megan M. Herzog and David S. Frankel, Special
Assistant Attorneys General, Dana Nessel, Attorney General,
Office of the Attorney General for the State of Michigan,
Gillian E. Wener, Assistant Attorney General, Keith Ellison,
Attorney General, Office of the Attorney General for the State
of Minnesota, Peter N. Surdo, Special Assistant Attorney
 3
General, Aaron D. Ford, Attorney General, Office of the
Attorney General for the State of Nevada, Heidi Parry Stern,
Solicitor General, Gurbir S. Grewal, Attorney General, Office
of the Attorney General for the State of New Jersey, Lisa J.
Morelli, Deputy Attorney General, Hector Balderas, Attorney
General, Office of the Attorney General for the State of New
Mexico, Tania Maestas, Chief Deputy Attorney General,
Joshua H. Stein, Attorney General, Office of the Attorney
General for the State of North Carolina, Asher Spiller,
Assistant Attorney General, Ellen F. Rosenblum, Attorney
General, Office of the Attorney General for the State of
Oregon, Paul Garrahan, Attorney-in-Charge, Steve Novick,
Special Assistant Attorney General, Josh Shapiro, Attorney
General, Office of the Attorney General for the
Commonwealth of Pennsylvania, Ann R. Johnston, Senior
Deputy Attorney General, Aimee D. Thomson, Deputy
Attorney General, Peter F. Neronha, Attorney General, Office
of the Attorney General for the State of Rhode Island, Gregory
S. Schultz, Special Assistant Attorney General, Thomas J.
Donovan, Jr., Attorney General, Office of the Attorney
General for the State of Vermont, Nicholas F. Persampieri,
Assistant Attorney General, Mark Herring, Attorney General,
Office of the Attorney General for the Commonwealth of
Virginia, Donald D. Anderson, Deputy Attorney General, Paul
Kugelman, Jr., Senior Assistant Attorney General and Chief,
Environmental Section, Caitlin Colleen Graham O=Dwyer,
Assistant Attorney General, Robert W. Ferguson, Attorney
General, Office of the Attorney General for the State of
Washington, Christopher H. Reitz and Emily C. Nelson,
Assistant Attorneys General, Joshua L. Kaul, Attorney
General, Office of the Attorney General for the State of
Wisconsin, Gabe Johnson-Karp, Assistant Attorney General,
Karl A. Racine, Attorney General, Office of the Attorney
General for the District of Columbia, Loren L. AliKhan,
Solicitor General, Tom Carr, City Attorney, Office of the City
 4
Attorney for the City of Boulder, Debra S. Kalish, Senior
Counsel, Mark A. Flessner, Corporation Counsel, Office of the
Corporation Counsel for the City of Chicago, Benna Ruth
Solomon, Deputy Corporation Counsel, Jared Policicchio,
Supervising Assistant Corporation Counsel, Kristin M.
Bronson, City Attorney, Office of the City Attorney for the
City and County of Denver, Lindsay S. Carder and Edward J.
Gorman, Assistant City Attorneys, Michael N. Feuer, City
Attorney, Office of the City Attorney for the City of Los
Angeles, Michael J. Bostrom, Assistant City Attorney, James
E. Johnson, Corporation Counsel, New York City Law
Department, Christopher G. King, Senior Counsel, Marcel S.
Pratt, City Solicitor, City of Philadelphia Law Department,
Scott J. Schwarz and Patrick K. O’Neill, Divisional Deputy
City Solicitors, and Thomas F. Pepe, City Attorney, City of
South Miami. Morgan A. Costello and Brian M. Lusignan,
Assistant Attorneys General, Office of the Attorney General
for the State of New York, Gavin G. McGabe, Deputy Attorney
General, Anne Minard, Special Assistant Attorney General,
Office of the Attorney General for the State of New Mexico,
Cynthia M. Weisz, Assistant Attorney General, Office of the
Attorney General for the State of Maryland, entered
appearances.

Kevin Poloncarz argued the cause for Power Company
Petitioners. With him on the briefs were Donald L. Ristow and
Jake Levine.

Mark W. DeLaquil argued the cause for Coal Industry
Petitioners. With him on the briefs were Shay Dvoretzky,
Charles T. Wehland, Jeffery D. Ubersax, Robert D. Cheren,
and Andrew Grossman.
 5
Theodore Hadzi-Antich argued the cause for Robinson
Enterprises Petitioners. With him on the briefs were Robert
Henneke and Ryan D. Walters.

Sean H. Donahue and Michael J. Myers argued the causes
for Public Health and Environmental Petitioners. On the briefs
were Ann Brewster Weeks, James P. Duffy, Susannah L.
Weaver, Joanne Spalding, Andres Restrepo, Vera Pardee,
Clare Lakewood, Howard M. Crystal, Elizabeth Jones,
Brittany E. Wright, Jon A. Mueller, David Doniger, Benjamin
Longstreth, Melissa J. Lynch, Lucas May, Vickie L. Patton,
Tomas Carbonell, Benjamin Levitan, Howard Learner, and
Scott Strand. Alejandra Nunez entered an appearance.

David M. Williamson argued the cause and filed the briefs
for Biogenic Petitioners.

Gene Grace, Jeff Dennis, and Rick Umoff were on the brief
for petitioners American Wind Energy Association, et al.

Theodore E. Lamm and Sean B. Hecht were on the brief
for amicus curiae Thomas C. Jorling in support of petitioners.

Gabriel Pacyniak, Brent Chapman, and Graciela Esquivel
were on the brief for amici curiae the Coalition to Protect
America=s National Parks and the National Parks Conservation
Association in support of petitioners.

Deborah A. Sivas and Matthew J. Sanders were on the
brief for amici curiae Administrative Law Professors in
support of petitioners.

Hope M. Babcock was on the brief for amici curiae the
American Thoracic Society, et al. in support of petitioners.
 6
Richard L. Revesz and Jack Lienke were on the brief for
amicus curiae the Institute for Policy Integrity at New York
University School of Law in support of petitioners.

Steph Tai was on the brief for amici curiae Climate
Scientists in support of petitioners.

Michael Burger and Collyn Peddie were on the brief for
amici curiae the National League of Cities, et al. in support of
petitioners.

Keri R. Steffes was on the brief for amici curiae Faith
Organizations in support of petitioners.

Shaun A. Goho was on the brief for amici curiae
Maximilian Auffhammer, et al. in support of petitioners.

Ethan G. Shenkman and Stephen K. Wirth were on the
brief for amici curiae Patagonia Works and Columbia
Sportswear Company in support of petitioners.

Mark Norman Templeton, Robert Adam Weinstock,
Alexander Valdes, and Benjamin Nickerson were on the brief
for amicus curiae Professor Michael Greenstone in support of
petitioners.

Nicole G. Berner and Renee M. Gerni were on the brief for
amicus curiae the Service Employees International Union in
support of petitioners.

Elizabeth B. Wydra, and Brianne J. Gorod were on the
brief for amici curiae Members of Congress in support of
petitioners.
 7
Jonas J. Monast was on the brief for amici curiae Energy
Modelers in support of petitioners.

Katherine Konschnik was on the brief for amici curiae
Former Commissioners of the Federal Energy Regulatory
Commission in support of petitioners.

Michael Landis, Elizabeth S. Merritt, and Wyatt G.
Sassman were on the brief for amici curiae Environment
America and National Trust for Historic Preservation in
support of petitioners.

Cara A. Horowitz was on the brief for amici curiae Grid
Experts in support of petitioners.

Eric Alan Isaacson was on the brief for amici curiae U.S.
Senators in support of petitioners.

Jonathan D. Brightbill, Principal Deputy Assistant
Attorney General, U.S. Department of Justice, and Meghan E.
Greenfield and Benjamin Carlisle, Attorneys, argued the
causes for respondents. With them on the brief was Jeffrey
Bossert Clark, Assistant Attorney General.

Lindsay S. See, Solicitor General, Office of the Attorney
General for the State of West Virginia, argued the cause for
State and Industry intervenors in support of respondents
regarding Affordable Clean Energy Rule. With her on the brief
were Patrick Morrisey, Attorney General, Thomas T.
Lampman, Assistant Solicitors General, Thomas A. Lorenzen,
Elizabeth B. Dawson, Rae Cronmiller, Kevin G. Clarkson,
Attorney General at the time the brief was filed, Office of the
Attorney General for the State of Alaska, Clyde Sniffen Jr.,
Attorney General, Leslie Rutledge, Attorney General, Office of
the Attorney General for the State of Arkansas, Nicholas J.
 8
Bronni, Solicitor General, Vincent M. Wagner, Deputy
Solicitor General, Dylan L. Jacobs, Assistant Solicitor General,
Steve Marshall, Attorney General, Office of the Attorney
General for the State of Alabama, Edmund G. LaCour, Jr.,
Solicitor General, Christopher M. Carr, Attorney General,
Office of the Attorney General for the State of Georgia,
Andrew A. Pinson, Solicitor General, Derek Schmidt, Attorney
General, Office of the Attorney General for the State of Kansas,
Jeffrey A. Chanay, Chief Deputy Attorney General, Curtis T.
Hill, Jr., Attorney General, Office of the Attorney General of
Indiana, Thomas M. Fisher, Solicitor General, Andrew
Beshear, Governor, Office of the Governor for the
Commonwealth of Kentucky, S. Travis Mayo, Chief Deputy
General Counsel, Taylor Payne, Deputy General Counsel,
Joseph A. Newberg, Deputy General Counsel and Deputy
Executive Director, Jeff Landry, Attorney General, Office of
the Attorney General for the State of Louisiana, Elizabeth B.
Murrill, Solicitor General, Harry J. Vorhoff, Assistant
Attorney General, Eric S. Schmitt, Attorney General, Office of
the Attorney General for the State of Missouri, D. John Sauer,
Solicitor General, Julie Marie Blake, Deputy Solicitor General,
Timothy C. Fox, Attorney General at the time the brief was
filed, Office of the Attorney General for the State of Montana,
Matthew T. Cochenour, Deputy Solicitor General, Wayne
Stenehjem, Attorney General, Office of the Attorney General
for the State of North Dakota, Paul M. Seby, Special Assistant
Attorney General, Douglas J. Peterson, Attorney General,
Office of the Attorney General for the State of Nebraska, Justin
D. Lavene, Assistant Attorney General, Dave Yost, Attorney
General, Office of the Attorney General of the State of Ohio,
Benjamin M. Flowers, Solicitor General, Cameron F.
Simmons, Principal Assistant Attorney General, Mike Hunter,
Attorney General, Office of the Attorney General for the State
of Oklahoma, Mithun Mansinghani, Solicitor General, Jason
R. Ravnsborg, Attorney General, Office of the Attorney
 9
General for the State of South Dakota, Steven R. Blair,
Assistant Attorney General, Alan Wilson, Attorney General,
Office of the Attorney General for the State of South Carolina,
James Emory Smith, Jr., Deputy Solicitor General, Ken
Paxton, Attorney General, Office of the Attorney General for
the State of Texas, Kyle D. Hawkins, Solicitor General, Sean
Reyes, Attorney General, Office of the Attorney General for the
State of Utah, Tyler R. Green, Solicitor General, Bridget Hill,
Attorney General, Office of the Attorney General for the State
of Wyoming, James Kaste, Deputy Attorney General, Todd E.
Palmer, William D. Booth, Obianuju Okasi, Carroll W.
McGuffey, III, Misha Tseytlin, C. Grady Moore, III, Julia
Barber, F. William Brownell, Elbert Lin, Allison D. Wood,
Scott A. Keller, Jeffrey H. Wood, Jeremy Evan Maltz, Steven P.
Lehotsky, Michael B. Schon, Emily Church Schilling, Kristina
R. Van Bockern, David M. Flannery, Kathy G. Beckett, Edward
L. Kropp, Amy M. Smith, Janet J. Henry, Melissa Horne,
Angela Jean Levin, Eugene M. Trisko, John A. Rego, Reed W.
Sirak, Michael A. Zody, Jacob Santini, Robert D. Cheren, Mark
W. DeLaquil, and Andrew M. Grossman. C. Frederick Beckner,
III, James R. Bedell, Margaret C. Campbell, Erik D. Lange,
and John D. Lazzaretti entered an appearance.

James P. Duffy argued the cause for Public Health and
Environmental Intervenors in support of respondents. With
him on the brief were Ann Brewster Weeks, Sean H. Donahue,
Susannah L. Weaver, Joanne Spalding, Andres Restrepo, Vera
Pardee, Clare Lakewood, Elizabeth Jones, Brittany E. Wright,
Jon A. Mueller, David Doniger, Benjamin Longstreth, Melissa
J. Lynch, Lucas May, Vickie L. Patton, Tomas Carbonell,
Benjamin Levitan, Howard Learner, and Scott Strand.

Letitia James, Attorney General, Office of the Attorney
General for the State of New York, Michael J. Myers, Senior
Counsel, Brian Lusignan, Assistant Attorney General of
 10
Counsel, Barbara D. Underwood, Solicitor General, Steven C.
Wu, Deputy Solicitor General, Matthew W. Grieco, Assistant
Solicitor General, Xavier Becerra, Attorney General, Office of
the Attorney General for the State of California, Robert W.
Byrne, Senior Assistant Attorney General, David A. Zonana,
Supervising Deputy Attorney General, Jonathan A. Wiener, M.
Elaine Meckenstock, Timothy E. Sullivan, Elizabeth B. Rumsey,
and Theodore A.B. McCombs, Deputy Attorneys General,
William Tong, Attorney General, Office of the Attorney
General for the State of Connecticut, Matthew I. Levine and
Scott N. Koschwitz, Assistant Attorneys General, Kathleen
Jennings, Attorney General, Office of the Attorney General for
the State of Delaware, Valerie S. Edge, Deputy Attorney
General, Philip J. Weiser, Attorney General, Office of the
Attorney General for the State of Colorado, Eric R. Olson,
Solicitor General, Robyn L. Wille, Senior Assistant Attorney
General, Clare E. Connors, Attorney General, Office of the
Attorney General for the State of Hawaii, William F. Cooper,
Deputy Attorney General, Aaron M. Frey, Attorney General,
Office of the Attorney General for the State of Maine, Laura
E. Jensen, Assistant Attorney General, Brian E. Frosh,
Attorney General, Office of the Attorney General for the State
of Maryland, John B. Howard, Jr., Joshua M. Segal, and Steven
J. Goldstein, Special Assistant Attorneys General, Maura
Healey, Attorney General, Office of the Attorney General for
the Commonwealth of Massachusetts, Melissa A. Hoffer and
Christophe Courchesne, Assistant Attorneys General, Megan
M. Herzog and David S. Frankel, Special Assistant Attorneys
General, Dana Nessel, Attorney General, Office of the
Attorney General for the State of Michigan, Gillian E. Wener,
Assistant Attorney General, Keith Ellison, Attorney General,
Office of the Attorney General for the State of Minnesota,
Peter N. Surdo, Special Assistant Attorney General, Aaron D.
Ford, Attorney General, Office of the Attorney General for the
State of Nevada, Heidi Parry Stern, Solicitor General, Gurbir
 11
S. Grewal, Attorney General, Office of the Attorney General
for the State of New Jersey, Lisa J. Morelli, Deputy Attorney
General, Hector Balderas, Attorney General, Office of the
Attorney General for the State of New Mexico, Tania Maestas,
Chief Deputy Attorney General, Joshua H. Stein, Attorney
General, Office of the Attorney General for the State of North
Carolina, Asher Spiller, Assistant Attorney General, Ellen F.
Rosenblum, Attorney General, Office of the Attorney General
for the State of Oregon, Paul Garrahan, Attorney-in-Charge,
Steve Novick, Special Assistant Attorney General, Josh
Shapiro, Attorney General, Office of the Attorney General for
the Commonwealth of Pennsylvania, Ann R. Johnston, Senior
Deputy Attorney General, Aimee D. Thomson, Deputy
Attorney General, Peter F. Neronha, Attorney General, Office
of the Attorney General for the State of Rhode Island, Gregory
S. Schultz, Special Assistant Attorney General, Thomas J.
Donovan, Jr., Attorney General, Office of the Attorney
General for the State of Vermont, Nicholas F. Persampieri,
Assistant Attorney General, Mark Herring, Attorney General,
Office of the Attorney General for the Commonwealth of
Virginia, Donald D. Anderson, Deputy Attorney General, Paul
Kugelman, Jr., Senior Assistant Attorney General and Chief,
Environmental Section, Caitlin Colleen Graham O=Dwyer,
Assistant Attorney General, Robert W. Ferguson, Attorney
General, Office of the Attorney General for the State of
Washington, Christopher H. Reitz and Emily C. Nelson,
Assistant Attorneys General, Karl A. Racine, Attorney
General, Office of the Attorney General for the District of
Columbia, Loren L. AliKhan, Solicitor General, Tom Carr,
City Attorney, Office of the City Attorney for the City of
Boulder, Debra S. Kalish, Senior Counsel, Mark A. Flessner,
Corporation Counsel, Office of the Corporation Counsel for the
City of Chicago, Benna Ruth Solomon, Deputy Corporation
Counsel, Jared Policicchio, Supervising Assistant Corporation
Counsel, Kristin M. Bronson, City Attorney, Office of the City
 12
Attorney for the City and County of Denver, Lindsay S. Carder
and Edward J. Gorman, Assistant City Attorneys, Michael N.
Feuer, City Attorney, Office of the City Attorney for the City
of Los Angeles, Michael J. Bostrom, Assistant City Attorney,
James E. Johnson, Corporation Counsel, New York City Law
Department, Christopher G. King, Senior Counsel, Marcel S.
Pratt, City Solicitor, City of Philadelphia Law Department,
Scott J. Schwarz and Patrick K. O’Neill, Divisional Deputy
City Solicitors, and Thomas F. Pepe, City Attorney, City of
South Miami were on the brief for the State and Municipal
Intervenors in support of respondents. Jeremiah Langston,
Assistant Attorney General, Office of the Attorney General for
the State of Montana, Stephen C. Meredith, Solicitor, Office of
the Attorney General for the Commonwealth of Kentucky,
Margaret I. Olson, Assistant Attorney General, Office of the
Attorney General for the State of North Dakota, and Erik E.
Petersen, Assistant Attorney General, Office of the Attorney
General for the State of Wyoming, and Robert A. Wolf entered
appearances.

Patrick Morrisey, Attorney General, Office of the
Attorney General for the State of West Virginia, Lindsay S.
See, Solicitor General, Thomas T. Lampman, Assistant
Solicitor General, Scott A. Keller, Jeffrey H. Wood, Jeremy
Evan Maltz, Steven P. Lehotsky, Michael B. Schon, Thomas A.
Lorenzen, Elizabeth B. Dawson, Rae Cronmiller, Steve
Marshall, Attorney General, Office of the Attorney General for
the State of Alabama, Edmund G. LaCour, Jr., Solicitor
General, Kevin G. Clarkson, Attorney General, Office of the
Attorney General for the State of Alaska at the time the brief
was filed, Clyde Sniffen, Jr., Attorney General, Leslie Rutledge,
Attorney General, Office of the Attorney General for the State
of Arkansas, Nicholas J. Bronni, Solicitor General, Vincent M.
Wagner, Deputy Solicitor General, Dylan L. Jacobs, Assistant
Solicitor General, Christopher M. Carr, Attorney General,
 13
Office of the Attorney General for the State of Georgia,
Andrew A. Pinson, Solicitor General, Derek Schmidt, Attorney
General, Office of the Attorney General for the State of Kansas,
Jeffrey A. Chanay, Chief Deputy Attorney General, Curtis T.
Hill, Jr., Attorney General, Office of the Attorney General of
Indiana, Thomas M. Fisher, Solicitor General, Andrew
Beshear, Governor, Office of the Governor for the
Commonwealth of Kentucky, S. Travis Mayo, Chief Deputy
General Counsel, Taylor Payne, Deputy General Counsel,
Joseph A. Newberg, Deputy General Counsel and Deputy
Executive Director, Jeff Landry, Attorney General, Office of
the Attorney General for the State of Louisiana, Elizabeth B.
Murrill, Solicitor General, Harry J. Vorhoff, Assistant
Attorney General, Eric S. Schmitt, Attorney General, Office of
the Attorney General for the State of Missouri, D. John Sauer,
Solicitor General, Julie Marie Blake, Deputy Solicitor General,
Timothy C. Fox, Attorney General at the time the brief was
filed, Office of the Attorney General for the State of Montana,
Matthew T. Cochenour, Deputy Solicitor General, Wayne
Stenehjem, Attorney General, Office of the Attorney General
for the State of North Dakota, Paul M. Seby, Special Assistant
Attorney General, Douglas J. Peterson, Attorney General,
Office of the Attorney General for the State of Nebraska, Justin
D. Lavene, Assistant Attorney General, Dave Yost, Attorney
General, Office of the Attorney General of the State of Ohio,
Benjamin M. Flowers, Solicitor General, Cameron F.
Simmons, Principal Assistant Attorney General, Mike Hunter,
Attorney General, Office of the Attorney General for the State
of Oklahoma, Mithun Mansinghani, Solicitor General, Jason
R. Ravnsborg, Attorney General, Office of the Attorney
General for the State of South Dakota, Steven R. Blair,
Assistant Attorney General, Alan Wilson, Attorney General,
Office of the Attorney General for the State of South Carolina,
James Emory Smith, Jr., Deputy Solicitor General, Ken
Paxton, Attorney General, Office of the Attorney General for
 14
the State of Texas, Kyle D. Hawkins, Solicitor General, Sean
Reyes, Attorney General, Office of the Attorney General for the
State of Utah, Tyler R. Green, Solicitor General, Bridget Hill,
Attorney General, Office of the Attorney General for the State
of Wyoming, James Kaste, Deputy Attorney General, Todd E.
Palmer, William D. Booth, Obianuju Okasi, Carroll W.
McGuffey, III, Misha Tseytlin, C. Grady Moore, III, Julia
Barber, F. William Brownell, Elbert Lin, Allison D. Wood,
Emily Church Schilling, Kristina R. Van Bockern, David M.
Flannery, Kathy G. Beckett, Edward L. Kropp, Amy M. Smith,
Janet J. Henry, Melissa Horne, Angela Jean Levin, Eugene M.
Trisko, John A. Rego, Reed W. Sirak, Michael A. Zody, Jacob
Santini, Robert D. Cheren, Mark W. DeLaquil, and Andrew M.
Grossman were on the brief for State and Industry Intervenors
in support of respondents regarding Clean Power Plan Repeal.

Wayne Stenehjem, Attorney General, Office of the
Attorney General for the State of North Dakota, and Paul M.
Seby, Special Assistant Attorney General, were on the brief for
intervenor State of North Dakota in support of the respondents.
Jerry Stouck entered an appearance.

Thomas J. Ward, Megan H. Berge, and Jared R. Wigginton
were on the brief for amicus curiae National Association of
Builders in support of respondents.

Before: MILLETT , PILLARD , and WALKER, Circuit Judges.

Opinion for the Court filed PER CURIAM.

Opinion concurring in part, concurring in the judgment in
part, and dissenting in part filed by Circuit Judge WALKER.
 15
TABLE OF C ONTENTS
I. Background .........................................................................17
A. The Clean Air Act ..........................................................17
B. Electricity and Climate Change .....................................21
1. Electricity.....................................................................21
2. Climate Change and the Federal Government ..........24
C. The Clean Power Plan ....................................................29
D. The ACE Rule ................................................................32
1. Repeal of the Clean Power Plan .................................32
2. Best System of Emission Reduction ..........................33
3. Degree of Emission Limitation Achievable ..............36
4. Implementing Regulations..........................................38
E. Petitions for Review .......................................................38
F. Jurisdiction and Standard of Review .............................39
II. Section 7411 .......................................................................40
A. Statutory Context............................................................40
1. Text ..............................................................................46
2. Statutory History, Structure, and Purpose .................59
3. Compliance Measures .................................................71
B. The Major Questions Doctrine ......................................74
1. The EPA’s Regulatory Mandate ................................75
2. Best System of Emission Reduction ..........................80
C. Federalism .......................................................................92
III. The EPA’s Authority to Regulate Carbon Dioxide
Emissions Under Section 7411 ..............................................98
A. The Coal Petitioners’ Challenges ..................................98
1. Endangerment Finding................................................99
2. Section 7411 and Section 7412’s
Parallel Operation...................................................... 111
B. The Robinson Petitioners’ Challenges ....................... 132
IV. Amendments to the Implementing Regulations ...... 138
V. Vacatur and Remand .................................................... 146
VI. Conclusion ..................................................................... 147
 16
As the Supreme Court recognized nearly fourteen years
ago, climate change has been called “the most pressing
environmental challenge of our time.” Massachusetts v. EPA,
549 U.S. 497, 505 (2007) (formatting modified). Soon
thereafter, the United States government determined that
greenhouse gas emissions are polluting our atmosphere and
causing significant and harmful effects on the human
environment. Endangerment and Cause or Contribute Findings
for Greenhouse Gases Under Section 202(a) of the Clean Air
Act (2009 Endangerment Finding), 74 Fed. Reg. 66,496,
66,497–66,499 (Dec. 15, 2009). And both Republican and
Democratic administrations have agreed: Power plants
burning fossil fuels like coal “are far and away” the largest
stationary source of greenhouse gases and, indeed, their role in
greenhouse gas emissions “dwarf[s] other categories[.]” EPA
Br. 169; see also Standards of Performance for Greenhouse
Gas Emissions from New, Modified, and Reconstructed
Stationary Sources: Electric Utility Generating Units (New
Source Rule), 80 Fed. Reg. 64,510, 64,522 (Oct. 23, 2015)
(fossil-fuel-fired power plants are “by far the largest emitters”
of greenhouse gases).

The question in this case is whether the Environmental
Protection Agency (EPA) acted lawfully in adopting the 2019
Affordable Clean Energy Rule (ACE Rule), 84 Fed. Reg.
32,520 (July 8, 2019), as a means of regulating power plants’
emissions of greenhouse gases. It did not. Although the EPA
has the legal authority to adopt rules regulating those
emissions, the central operative terms of the ACE Rule and the
repeal of its predecessor rule, the Clean Power Plan, 80 Fed.
Reg. 64,662 (Oct. 23, 2015), hinged on a fundamental
misconstruction of Section 7411(d) of the Clean Air Act. In
addition, the ACE Rule’s amendment of the regulatory
framework to slow the process for reduction of emissions is
arbitrary and capricious. For those reasons, the ACE Rule is
 17
vacated, and the record is remanded to the EPA for further
proceedings consistent with this opinion.

I. BACKGROUND

A. T HE C LEAN AIR ACT

In 1963, Congress passed the Clean Air Act, 42 U.S.C.
§ 7401 et seq., “to protect and enhance the quality of the
Nation’s air resources so as to promote the public health and
welfare and the productive capacity of its population[,]” id.
§ 7401(b)(1). Animating the Act was Congress’ finding that
“growth in the amount and complexity of air pollution brought
about by urbanization, industrial development, and the
increasing use of motor vehicles[] has resulted in mounting
dangers to the public health and welfare[.]” Id. § 7401(a)(2).

Section 111 of the Clean Air Act, which was added in 1970
and codified at 42 U.S.C. § 7411, directs the EPA to regulate
any new and existing stationary sources of air pollutants that
“cause[], or contribute[] significantly to, air pollution” and that
“may reasonably be anticipated to endanger public health or
welfare.” 42 U.S.C. § 7411(b)(1)(A); see id. § 7411(d), (f)
(providing that the EPA Administrator “shall” regulate existing
and new sources of air pollution). A “stationary source” is a
source of air pollution that cannot move, such as a power plant.
See id. § 7411(a)(3) (defining “stationary source” as “any
building, structure, facility, or installation which emits or may
emit any air pollutant[]”). An example of a common non-
stationary source of air pollution is a gas-powered motor
vehicle. See Utility Air Regulatory Group v. EPA (UARG), 573
U.S. 302, 308 (2014).

Within 90 days of the enactment of Section 7411, the EPA
Administrator was to promulgate a list of stationary source
categories that “cause[], or contribute[] significantly to, air
 18
pollution[.]” 42 U.S.C. § 7411(b)(1)(A). In 1971, the
Administrator included fossil-fuel-fired steam-generating
power plants on that list. Air Pollution Prevention and Control:
List of Categories of Stationary Sources, 36 Fed. Reg. 5,931
(March 31, 1971); see also New Source Rule, 80 Fed. Reg. at
64,527–64,528. Today’s power plants fall in that same
category. ACE Rule, 84 Fed. Reg. at 32,557 n.250.

Once a stationary source category is listed, the
Administrator must promulgate federal “standards of
performance” for all newly constructed sources in the category.
42 U.S.C. § 7411(b)(1)(B). The Act defines a “standard of
performance” as

a standard for emissions of air pollutants which
reflects 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 environmental impact and energy
requirements) the Administrator determines has been
adequately demonstrated.

Id. § 7411(a)(1).

Once such a new source regulation is promulgated, the
Administrator also must issue emission guidelines for already-
existing stationary sources within that same source category.
42 U.S.C. § 7411(d)(1)(A)(ii); see also American Elec. Power
Co., Inc. v. Connecticut (AEP), 564 U.S. 410, 424 (2011).

While the new source standards are promulgated and
enforced entirely by the EPA, the Clean Air Act prescribes a
process of cooperative federalism for the regulation of existing
sources. Under that structure, the statute delineates three
distinct regulatory steps involving three sets of actors—the
 19
EPA, the States, and regulated industry—each of which has a
flexible role in choosing how to comply. See 42 U.S.C.
§ 7411(a)(1), (d). This allows each State to work with the
stationary sources within its jurisdiction to devise a plan for
meeting the federally promulgated quantitative guideline for
emissions. See id. § 7411(d).

The process starts with the EPA first applying its expertise
to determine “the degree of emission limitation achievable
through the application of the best system of emission
reduction” that “has been adequately demonstrated.” 42 U.S.C.
§ 7411(a)(1); see 40 C.F.R. § 60.22a. That system must “tak[e]
into account the cost of achieving such reduction and any
nonair quality health and environmental impact and energy
requirements[.]” 42 U.S.C. § 7411(a)(1). Once the
Administrator identifies the best system of emission reduction,
she then determines the amount of emission reduction that
existing sources should be able to achieve based on the
application of that system and adopts corresponding emission
guidelines. Id.; see also, e.g., ACE Rule, 84 Fed. Reg. at
32,523; Clean Power Plan, 80 Fed. Reg. at 64,719.

Each State then submits to the EPA a plan that
(i) establishes standards of performance for that State’s
existing stationary sources’ air pollutants (excepting pollutants
already subject to separate federal emissions standards), and
(ii) “provides for the implementation and enforcement of such
standards of performance[]” by the State. 42 U.S.C.
§ 7411(d)(1); see 40 C.F.R. § 60.23a. The standards of
performance must “reflect[]” the emission targets that the EPA
has determined are achievable. 42 U.S.C. § 7411(a)(1). In this
context, a state standard need not adopt the best system
identified by the EPA to “reflect[]” it. Id.; see 40 C.F.R.
§ 60.24a(c). Instead, the Clean Air Act affords States
significant flexibility in designing and enforcing standards that
 20
employ other approaches so long as they meet the emission
guidelines prescribed by the Agency.

If a State fails to submit a satisfactory plan, the EPA may
prescribe a plan for that State. 42 U.S.C. § 7411(d)(2)(A); see
40 C.F.R. § 60.27a(c)-(e). Similarly, if the State submits a plan
but fails to enforce it, the EPA itself may enforce the plan’s
terms. Id. § 7411(d)(2)(B).

The third and final set of relevant actors are the regulated
entities themselves, to which, under the Act, the States may
afford leeway in crafting compliance measures. See Clean
Power Plan, 80 Fed. Reg. at 64,666; ACE Rule, 84 Fed. Reg.
at 32,555.

The EPA has exercised its authority under Section 7411
over the years to set emission limitations for different types of
air pollution from various categories of existing sources. See
42 Fed. Reg. 12,022 (March 1, 1977) (fluorides from phosphate
fertilizer plants); 42 Fed. Reg. 55,796 (Oct. 18, 1977) (acid
mist from sulfuric acid plants); 44 Fed. Reg. 29,828 (May 22,
1979) (total reduced sulfur from kraft pulp plants); 45 Fed.
Reg. 26,294 (April 17, 1980) (fluorides from primary
aluminum plants); 60 Fed. Reg. 65,387 (Dec. 19, 1995)
(various pollutants from municipal waste combustors); 61 Fed.
Reg. 9905 (March 12, 1996) (landfill gases from municipal
solid waste landfills); 70 Fed. Reg. 28,606 (May 18, 2005)
(mercury from coal-fired power plants).

The Clean Air Act is a comprehensive statute that includes
a variety of regulatory programs for tackling air pollution in
addition to Section 7411. Regulated parties may be subject to
one or more programs. As relevant here, the National Ambient
Air Quality Standards (NAAQS) provisions, 42 U.S.C.
§§ 7408–7410, govern the levels of specified air pollutants that
may be present in the atmosphere to protect air quality and the
 21
public health and welfare. The Hazardous Air Pollutants
program, id. § 7412, directs the EPA to establish strict emission
limitations for the most dangerous air pollutants emitted from
major sources. Section 7411’s cooperative federalism program
for existing sources operates as a gap-filler, requiring the EPA
to regulate harmful emissions not controlled under those other
two programs. Id. § 7411(d)(1)(i).

B. E LECTRICITY AND CLIMATE CHANGE

1. Electricity

Electricity powers the world. Chances are that you are
reading this opinion on a device that consumes electricity. Yet
two distinct characteristics of electricity make its production
and delivery in the massive quantities demanded by consumers
an exceptionally complex process. First, unlike most products,
electricity is a perfectly fungible commodity. Grid Experts
Amicus Br. 6. A watt of electricity is a watt of electricity, no
matter who makes it, how they make it, or where it is
purchased. Second, at least as of now, this highly demanded
product cannot be effectively stored at scale after it is created.
Paul L. Joskow, Creating a Smarter U.S. Electricity Grid, 26
J. ECON. PERSP. 29, 31–33 (2012).1 Instead, electricity must

1
Change in storage capacity is picking up speed. See generally
Richard L. Revesz & Burcin Unel, Managing the Future of the
Electricity Grid: Energy Storage and Greenhouse Gas Emissions,
42 HARV. ENV’T L. REV. 139, 140–141 (2018) (describing ongoing
declines in cost of storage); LAZARD, LAZARD’S LEVELIZED COST
OF STORAGE ANALYSIS—VERSION 6.0 (2020) (noting “storage costs
have declined across most use cases and technologies, particularly
for shorter-duration applications, in part driven by evolving
preferences in the industry”). Nevertheless, the grid’s production
capacity still far exceeds its present storage capacity. Univ. of Mich.
 22
constantly be produced, and is almost instantaneously
consumed. See Clean Power Plan, 80 Fed. Reg. at 64,677,
64,692; Grid Experts Amicus Br. 8.

Those unique attributes led to the creation of the American
electrical grid. 2 The grid has been called the “supreme
engineering achievement of the 20th century,” MASS. INST. OF
TECH., THE FUTURE OF THE ELECTRIC GRID 1 (2011)
(formatting modified), and it is an exceptionally complex,
interconnected system. “[A]ny electricity that enters the grid
immediately becomes a part of a vast pool of energy that is
constantly moving[.]” New York v. FERC, 535 U.S. 1, 7
(2002). That means that units of electricity as delivered to the
user are identical, no matter their source. On the grid, there is
no coal-generated electricity or renewable-generated
electricity; there is just electricity. See Clean Power Plan, 80
Fed. Reg. at 64,692; Grid Experts Amicus Br. 7–8. Also,
because storing electricity for any length of time remains
technically challenging and often costly, the components of the
grid must operate as a perfectly calibrated machine to deliver
the amount of electricity that all consumers across the United
States need at the moment they need it. Grid Experts Amicus

Ctr. for Sustainable Sys., U.S. GRID ENERGY STORAGE (Sept. 2020),
http://css.umich.edu/sites/default/files/US%20Grid%20Energy%20
Storage_CSS15-17_e2020.pdf (last visited Jan. 11, 2021) (United
States has 1,100 gigawatts of installed generation capacity and just
23 gigawatts of storage capacity).
2
Technically, “grids.” There are three regional grids in the
contiguous United States: Eastern, Western, and Texas. Grid
Experts Amicus Br. 9; see also United States Dep’t of Energy, North
American Electric Reliability Corporation Interconnections,
https://www.energy.gov/oe/downloads/north-american-electric-
reliability-corporation-interconnections (last visited Jan. 11, 2021).
 23
Br. 8, 10–11; see also 80 Fed. Reg. at 64,677. “If [someone]
in Atlanta on the Georgia [leg of the] system turns on a light,
every generator on Florida’s system almost instantly is caused
to produce some quantity of additional electric energy which
serves to maintain the balance in the interconnected system[.]”
Federal Power Comm’n v. Florida Power & Light Co., 404
U.S. 453, 460 (1972) (citation omitted). “Like orchestra
conductors signaling entrances and cut-offs, grid operators use
automated systems to signal particular generators to dispatch
more or less power to the grid as needed over the course of the
day, thus ensuring that power pooled on the grid rises and falls
to meet changing demand.” Grid Experts Amicus Br. 11.

Most generators of electricity on the American grid create
power by burning fossil fuels like coal, oil, and natural gas. See
United States Energy Information Administration (EIA),
Frequently Asked Questions: What Is U.S. Electricity
Generation by Energy Source? (Nov. 2, 2020),
https://www.eia.gov/tools/faqs/faq.php?id=427&t=3 (last
visited Jan. 11, 2021) (fossil fuels represented 62.6 percent of
electricity generation in 2019). Some of those power plants
take a fossil fuel (usually coal) and burn it in a water boiler to
make steam. Other power plants take a different fossil fuel
(usually natural gas), mix it with highly compressed air, and
ignite it to release a combination of super-hot gases. Either
way, that steam or superheated mixture is piped into giant
turbines that catch the gases and rotate at extreme speeds.
Those turbines turn generators, which spin magnets within wire
coils to produce electricity. EIA, Electricity Explained
(Nov. 9, 2020), https://www.eia.gov/energyexplained/
electricity/how-electricity-is-generated.php (last visited Jan
11, 2021).
 24
2. Climate Change and the Federal Government

Electrical power has become virtually as indispensable to
modern life as air itself. But electricity generation has come
into conflict with air quality in ways that threaten human health
and well-being when power generated by burning fossil fuels
emits carbon dioxide and other polluting greenhouse gases into
the air.

Since the late 1970s, the federal government has focused
“serious attention” on the effects of carbon dioxide pollution
on the climate. Massachusetts v. EPA, 549 U.S. at 507. In
1978, Congress adopted the National Climate Program Act,
Pub. L. No. 95-367, 92 Stat. 601, which directed the President
to study and devise an appropriate response to “man-induced
climate processes and their implications[,]” id. § 3; see
Massachusetts v. EPA, 549 U.S. at 507–508. In response, the
National Academy of Sciences’ National Research Council
reported “no reason to doubt that climate changes will result”
if “carbon dioxide continues to increase,” and “[a] wait-and-
see policy may mean waiting until it is too late.”
Massachusetts v. EPA, 549 U.S. at 508 (quoting CLIMATE
RESEARCH BOARD , CARBON D IOXIDE & CLIMATE:
A SCIENTIFIC ASSESSMENT, at viii (1979)).

In 1987, Congress passed the Global Climate Protection
Act, which found that “manmade pollution[,]” including “the
release of carbon dioxide, * * * may be producing a long-term
and substantial increase in the average temperature on Earth[.]”
Pub. L. No. 100-204, Title XI, §1102(1), 101 Stat. 1407, 1408
(codified at 15 U.S.C. § 2901 note). The Climate Protection
Act directed the EPA to formulate a “coordinated national
policy on global climate change.” Id. § 1103(b), 101 Stat. at
1408; see Massachusetts v. EPA, 549 U.S. at 508.
 25
It was not until the Supreme Court’s 2007 decision in
Massachusetts v. EPA, however, that the Court confirmed that
carbon dioxide and other greenhouse gas emissions constituted
“air pollutant[s]” covered by the Clean Air Act. See 549 U.S.
at 528. The Supreme Court explained that the Clean Air Act’s
“sweeping definition of ‘air pollutant’ includes ‘any air
pollution agent or combination of such agents, including any
physical, chemical . . . substance or matter which is emitted
into or otherwise enters the ambient air[.]’” Id. at 528–529
(quoting 42 U.S.C. § 7602(g)). The Act, the Supreme Court
held, “is unambiguous” in that regard. Id. at 529. “On its face,
the definition embraces all airborne compounds of whatever
stripe, and underscores that intent through the repeated use of
the word ‘any.’” Id. And “[c]arbon dioxide” and other
common greenhouse gases are “without a doubt” chemical
substances that are “emitted into . . . the ambient air.” Id.
(quoting 42 U.S.C. § 7602(g)).

Given that statutory command, the Supreme Court ruled
that the EPA “can avoid taking further action” to regulate such
pollution “only if it determines that greenhouse gases do not
contribute to climate change” or offers some reasonable
explanation for not resolving that question. Massachusetts v.
EPA, 549 U.S. at 533.

Taking up the mantle, the EPA in 2009 found
“compelling[]” evidence that emissions of greenhouse gases
are polluting the atmosphere and are endangering human health
and welfare by causing significant damage to the environment.
2009 Endangerment Finding, 74 Fed. Reg. at 66,497; see id.
(“[T]he Administrator finds that greenhouse gases in the
atmosphere may reasonably be anticipated both to endanger
public health and to endanger public welfare. * * * The
Administrator has determined that the body of scientific
evidence compellingly supports this finding.”); id. at 66,497–
 26
66,499. The EPA concluded that “‘compelling’ evidence
supported the ‘attribution of observed climate change to
anthropogenic’ [that is, human-influenced] emissions of
greenhouse gases[.]” AEP, 564 U.S. at 417 (quoting 74 Fed.
Reg. at 66,518). The “[c]onsequent dangers of greenhouse gas
emissions,” the EPA determined, include

increases in heat-related deaths; coastal inundation
and erosion caused by melting icecaps and rising sea
levels; more frequent and intense hurricanes, floods,
and other “extreme weather events” that cause death
and destroy infrastructure; drought due to reductions
in mountain snowpack and shifting precipitation
patterns; destruction of ecosystems supporting
animals and plants; and potentially “significant
disruptions” of food production.

Id. (quoting 74 Fed. Reg. at 66,524–66,535).

Not long thereafter, the Supreme Court ruled that the
significant greenhouse gas pollution caused by fossil-fuel-fired
power plants is subject to regulation under Section 7411 of the
Clean Air Act. AEP, 564 U.S. at 424 (holding that Section
7411 “speaks directly to emissions of carbon dioxide from
[fossil-fuel-fired] plants[]”) (internal quotation marks omitted).
The Court concluded that the EPA’s expertise made it “best
suited to serve as primary regulator of greenhouse gas
emissions.” Id. at 428.

In 2015, with the 2009 carbon dioxide endangerment
finding continuing in effect, the EPA reaffirmed that
greenhouse gases “endanger public health, now and in the
future.” New Source Rule, 80 Fed. Reg. at 64,518. The EPA
explained that, “[b]y raising average temperatures, climate
change increases the likelihood of heat waves, which are
associated with increased deaths and illnesses[,]” particularly
 27
among “[c]hildren, the elderly, and the poor[.]” Id. at 64,517.
In addition, the EPA found that “[c]limate change impacts
touch nearly every aspect of public welfare.” Id. Among the
“multiple threats caused by human emissions of [greenhouse
gases],” the EPA pointed to climate changes that “are expected
to place large areas of the country at serious risk of reduced
water supplies, increased water pollution, and increased
occurrence of extreme events such as floods and droughts.” Id.
The EPA “emphasize[d] the urgency of reducing [greenhouse
gas] emissions due to * * * projections that show [greenhouse
gas] concentrations climbing to ever-increasing levels in the
absence of mitigation[,]” citing independent assessments
finding that, “without a reduction in emissions, CO2
concentrations by the end of the century would increase to
levels that the Earth has not experienced for more than 30
million years.” Id. at 64,518.

The federal government’s consistent recognition of the
danger to public health and welfare caused by climate change,
and the signal contribution of greenhouse gas emissions from
power plants to global warming, continues to the present. In
2018, President Trump’s administration concluded that
“Earth’s climate is now changing faster than at any point in the
history of modern civilization, primarily as a result of human
activities.” U.S. GLOBAL CHANGE RESEARCH PROGRAM,
FOURTH NATIONAL CLIMATE A SSESSMENT, VOLUME II:
IMPACTS, RISKS, AND ADAPTATION IN THE UNITED STATES
(REPORT- IN-B RIEF) 24 (2018). The administration added that
“the evidence of human-caused climate change is
overwhelming and continues to strengthen,” and “the impacts
of climate change are intensifying across the country[.]” Id. at
26 (emphasis omitted). “Climate-related changes in weather
patterns and associated changes in air, water, food, and the
environment are affecting the health and well-being of the
American people, causing injuries, illnesses, and death.” Id. at
 28
102. The administration’s report concluded that urgent action
is needed to mitigate these dangers because “[f]uture risks from
climate change depend primarily on decisions made today.” Id.
at 13.

In preparing the ACE Rule, the EPA expressly
acknowledged its continued adherence to the 2015
endangerment finding. 84 Fed. Reg. at 32,533 (The 2015 New
Source Rule “continues to provide the requisite predicate for
applicability of [Clean Air Act] section 111(d).”); id. at 32,557
n.250; see also Emission Guidelines for Greenhouse Gas
Emissions from Existing Electric Utility Generating Units;
Revisions to Emission Guideline Implementing Regulations;
Revisions to New Source Review Program: Proposed Rule, 83
Fed. Reg. 44,746, 44,751 (Aug. 31, 2018) (confirming that the
2015 New Source Rule “remains on the books[]”); EPA
Br. 217.

That endangerment finding provided the essential factual
foundation—and triggered a statutory mandate—for the EPA
to regulate greenhouse gas emissions from both new and
existing power plants. See New Source Rule, 80 Fed. Reg. at
64,527, 64,529–64,532; Clean Power Plan, 80 Fed. Reg. at
64,683–64,690; see also 42 U.S.C. §§ 7411(b)(1)(A)–(B) (duty
to regulate new stationary sources that contribute significantly
to dangerous pollution identified in endangerment finding),
7411(d)(1)(A)(ii) (duty to regulate existing stationary sources
that would be regulated under § 7411(b) if they were new
stationary sources). Recall, Section 7411(b)(1)(A) provides
that the EPA Administrator “shall” regulate any category of
sources that, “in his judgment * * * causes, or contributes
significantly to, air pollution which may reasonably be
anticipated to endanger public health or welfare.” The EPA
endangerment findings reflect such well-established risks.
 29
C. T HE C LEAN POWER PLAN

In the last decade, the EPA has heavily focused its
regulation of greenhouse gases on the power sector because
“power plants are far and away the largest stationary-category
source of greenhouse gases[,]” and “power plants’
contributions to CO2 pollution * * * dwarf[] other
categories[.]” EPA Br. 169.

In October 2015, the EPA issued greenhouse gas emission
standards for new and modified power plants. See New Source
Rule, 80 Fed. Reg. at 64,510. In so doing, the EPA found that,
“[a]ll told, these fossil fuel-fired [power plants] emit almost
one-third of all U.S. [greenhouse gas] emissions, and are
responsible for almost three times as much as the emissions
from the next ten stationary source categories combined.” Id.
at 64,531. That rule and finding remain in effect and are not
challenged in this litigation.

The EPA then turned to the regulation of existing power
plants. The EPA began, as the Clean Air Act requires, by
determining the best system of emission reduction that has
been adequately demonstrated for existing fossil-fuel-fired
power plants. See 42 U.S.C. § 7411(a)(1); Clean Power Plan,
80 Fed. Reg. at 64,718. In identifying that system, the EPA
chose to build on the established grid system and methods of
operation already adopted by and familiar to the power sector.
See 80 Fed. Reg. at 64,725, 64,727–64,728. The regulations
and standards that the EPA formulated came to be known as
the Clean Power Plan. Id. at 64,663.

In the Clean Power Plan, the EPA determined that a
combination of three existing methods of emission reduction—
which the Plan referred to as building blocks, 80 Fed. Reg. at
64,667—formed the “best system of emission reduction,” 42
U.S.C. § 7411(a)(1).
 30
First, the system incorporated heat-rate improvements—
that is, technological measures that improve efficiency at coal-
fired steam power plants and, in that way, reduce the amount
of coal that must be burned to produce each watt of electricity
to the grid. 80 Fed. Reg. at 64,667.

Second, the system added the “substitut[ion of] increased
generation from lower-emitting existing natural gas combined
cycle units for generation from higher-emitting affected steam
generating” power plants, which are mostly coal-fired. 80 Fed.
Reg. at 64,667.

Third, the system prioritized the use of electricity
generated from zero-emitting renewable-energy sources over
electricity from the heavily greenhouse-gas-polluting fossil-
fuel-fired power plants. 80 Fed. Reg. at 64,667.

Those second and third methods of emission control are
often referred to as “generation shifting” because the
reductions occur when the source of power generation shifts
from higher-emission power plants to less-polluting sources of
energy. See Clean Power Plan, 80 Fed. Reg. at 64,728–64,729.
As the EPA observed, such shifts in generation already occur
all the time as a matter of grid mechanics. That is, within the
grid’s “Constrained Least-Cost Dispatch” system, production
from “generators with the lowest variable costs” will be
dispatched “first, as system operational limits allow, until all
demand is satisfied.” Grid Experts Amicus Br. 12.
“[R]enewable energy generators typically receive dispatch
priority because they have lower variable costs than fossil-fuel-
fired generators, which must purchase fuel.” Id. at 13 (citing
80 Fed. Reg. at 64,693). The EPA found that most electricity
is generated by diversified utilities that could achieve most or
all of the shift to lower- or no-emission generation by
 31
reassessing the dispatch priority of their own assets. See 80
Fed. Reg. at 64,796, 64,804.

As required by Section 7411(a)(1), the EPA then
quantified the degree of emission reduction achievable under
that three-tier best system for the relevant fossil-fuel-fired
power plants and translated it into state-specific emissions
goals for 2030. Clean Power Plan, 80 Fed. Reg. at 64,824–
64,825. To permit additional flexibility, the Plan actually
provided two alternative types of targets: rate-based goals,
reflecting the rate of emission per certain amount of generation,
and mass-based goals, reflecting the total emission from a
State’s sources. Id. at 64,820, 64,824–64,825 Tables 12, 13.
The alternative metrics were an added source of flexibility for
States in choosing how they would meet the federal limits.

Under the Clean Air Act, States could then propose plans
that set standards of performance for their existing power
plants that would meet those emission goals. Clean Power
Plan, 80 Fed. Reg. at 64,664. In doing so, the States and their
power plants were under no obligation to use the three specific
methods that the EPA had identified in determining the best
system of emission reduction. Rather, consistent with Section
7411(d)’s cooperative federalism approach, States were free to
choose any measures, approaches, or technologies that they
deemed appropriate to meet the federal guidelines. For
example, they could adopt technological controls already in use
by some power plants like carbon capture and sequestration (by
which carbon dioxide is captured from the plant’s flue gas
before it is emitted and then securely stored so it cannot reach
the atmosphere) or co-firing (where fuels that release less
carbon dioxide are burned alongside fuels that release more to
reduce the amount of the latter used). See id. at 64,883. The
EPA also suggested that States might rely on emissions-trading
 32
programs (often referred to as cap-and-trade) and other
potential compliance strategies. Id. at 64,887.

The EPA found that its proposed approach was “consistent
with, and in some ways mirrors, the interconnected,
interdependent and highly regulated nature of the utility power
sector[]” and its grid, as well as “the daily operation of affected
[power plants] within this framework, and the critical role of
utilities in providing reliable, affordable electricity at all times
and in all places within this complex, regulated system.” Clean
Power Plan, 80 Fed. Reg. at 64,678.

The Clean Power Plan was challenged in this court. West
Virginia v. EPA, No. 15-1363 (and consolidated cases) (D.C.
Cir. Oct. 23, 2015). After we heard argument en banc, but
before we issued a decision, that litigation was held in
abeyance and ultimately dismissed as the EPA reassessed its
position. No. 15-1363, Docs. 1673071, 1806952.

D. T HE ACE RULE

In 2019, the EPA issued a new rule that repealed and
replaced the Clean Power Plan: The Affordable Clean Energy
(ACE) Rule. See Repeal of the Clean Power Plan; Emission
Guidelines for Greenhouse Gas Emissions from Existing
Electric Utility Generating Units; Revisions to Emission
Guidelines Implementing Regulations, 84 Fed. Reg. 32,520
(July 8, 2019). That Rule is the subject of this litigation.

1. Repeal of the Clean Power Plan

At the outset, the ACE Rule repealed the Clean Power
Plan. The EPA explained that it felt itself statutorily compelled
to do so because, in its view, “the plain meaning” of Section
7411(d) “unambiguously” limits the best system of emission
reduction to only those measures “that can be put into operation
 33
at a building, structure, facility, or installation.” ACE Rule, 84
Fed. Reg. at 32,523–32,524. Because the Clean Power Plan’s
best system was determined by using some emission control
measures that the EPA characterized as physically operating
off the site of coal-fired power plants—such as some forms of
generation shifting and emissions trading—the EPA concluded
that it had no choice but to repeal the Plan. Id. The EPA
emphasized “that [its] action is based on the only permissible
reading of the statute and [it] would reach that conclusion even
without consideration of the major question doctrine,” while
adding that application of that latter doctrine “confirms the
unambiguously expressed intent” of Section 7411. Id. at
32,529.

2. Best System of Emission Reduction

Considering its authority under Section 7411 to be
confined to physical changes to the power plants themselves,
the EPA’s ACE Rule determined a new best system of emission
reduction for coal-fired power plants only. The EPA left
unaddressed in this rulemaking (or elsewhere) greenhouse gas
emissions from other types of fossil-fuel-fired power plants,
such as those fired by natural gas or oil. ACE Rule, 84 Fed.
Reg. at 32,533.

The EPA’s proposed system relied solely on heat-rate
improvement technologies and practices that could be applied
at and to existing coal-fired power plants. ACE Rule, 84 Fed.
Reg. at 32,525, 32,537. The EPA selected only seven heat-rate
improvement techniques as components of its best system. Id.
at 32,537. Six of those measures were new-to-the-plant
technologies or “equipment upgrades.” Id. at 32,536–32,537
(naming as part of the best system (1) adding or upgrading
neural networks and intelligent sootblowers; (2) upgrading
boiler feed pumps; (3) replacing or upgrading air heater and
 34
duct leakage control devices; (4) adding variable frequency
drives in feed pumps and induced-draft fans; (5) blade path
upgrades; and (6) redesigning or replacing economizers). The
seventh measure was the use of “best operating and
maintenance practices” implementing heat-rate improvement
techniques. Id. at 32,537, 32,540. The EPA limited itself to
techniques that could be “applied broadly” to the Nation’s coal-
fired plants, which primarily amounted to upgrades to existing
equipment. Id. at 32,536.

The EPA explained that only five of the seven listed
techniques directly reduce the heat rate of power plants. See
ACE Rule, 84 Fed. Reg. at 32,538–32,540. The other two
techniques—replacing or upgrading the boiler feed pump and
installing variable frequency drives—serve to reduce the
amount of energy that a power plant must use to run its own
general operations. Id. at 32,538–32,539.3 So those two
techniques do not make a power plant more efficient in turning
coal into power, but instead allow power plants to dispatch
more of the power they produce to the grid rather than using it
internally. Id.

3
The boiler feed pump is a device that is used to pump water
into the boiler. 84 Fed. Reg. at 32,538. It consumes a “large
fraction” of the power used to run the plant. Id. Because the boiler
feed pump requires so much energy, the EPA suggested that
“maintenance on these pumps should be rigorous to ensure both
reliability and high-efficiency operation.” Id. Variable frequency
drives “enable[] very precise and accurate speed control” of both
boiler feed pumps and “induced draft (ID) fans,” which “maintain
proper flue gas flow through downstream air pollutant control
equipment[.]” Id. at 32,539. This precise control would reduce the
excess use of fans and pumps, requiring less energy. See id.
 35
The EPA identified two of its other chosen techniques—
blade path and economizer upgrades—as the measures that, of
all the considered technologies, were “expected to offer some
of the largest [heat-rate] improvements.” ACE Rule, 84 Fed.
Reg. at 32,537 (showing table predicting highest heat-rate
improvement range in economizer redesign or replacements
and blade path upgrades).4

But the EPA then stated that it expected some power plants
would not adopt those two technologies because their use could
trigger additional regulation that the companies would find
burdensome. 84 Fed. Reg. at 32,537 (“[B]ased on public
comments * * *, [blade path upgrades and economizer redesign
or replacement] are [heat-rate improvement] technologies that
have the most potential to trigger [New Source Review]
requirements.”). In fact, the EPA did not model those two
techniques in its regulatory impact analysis precisely because
it was unlikely that they would be adopted. J.A. 1656–1657.

Finally, the EPA acknowledged that the proposed
technologies could create a “rebound effect.” ACE Rule, 84
Fed. Reg. at 32,542. A rebound effect means that net carbon
dioxide emissions actually increase as a result of the efficiency
improvements made by power plants. Id. This happens
because, as the efficiency upgrades make coal-based energy
cheaper to produce, coal-fired power plants will have an
incentive to run more often, thereby increasing their overall
emissions. Id. The EPA found that risk of increased emissions
irrelevant because its best system of emission reduction “is
aimed at improving a source’s emissions rate performance at

4
“Blade path upgrades” consist of upgrades to the steam
turbine. Economizers are heat-exchange devices that “capture waste
heat from boiler flue gas” and use that captured heat to help heat the
boiler feedwater. Id. at 32,540.
 36
the unit-level,” rather than reducing the overall volume of
emissions by individual sources. Id. at 32,543.

In choosing its seven proposed power-plant-based heat-
rate improvement technologies, the EPA excluded from its best
system several other suggested methods of reducing emissions,
including (1) natural gas co-firing, repowering, and refueling;
(2) biomass co-firing; and (3) carbon capture and storage
technologies. ACE Rule, 84 Fed. Reg. at 32,543–32,547. The
EPA rejected biomass co-firing primarily because “any
potential net reductions in emissions from biomass use occur
outside of the regulated source,” and so do not fall within the
EPA’s reading of Section 7411(d) as confined to emission
limits imposed at and to individual plants. Id. at 32,546. The
EPA excluded natural gas co-firing and carbon capture and
storage from its own best system, citing cost, geographical, and
operational concerns. Id. at 32,544–32,545, 32,547–32,548.
The EPA provided that sources could choose to use natural gas
co-firing or carbon capture—but not biomass co-firing—to
meet state-established standards of performance. Id. at 32,555.

3. Degree of Emission Limitation Achievable

Having determined its best system of emission reduction,
the EPA then purported to prescribe the “degree of emission
limitation achievable,” which States could use to create their
own standards of performance. 42 U.S.C. § 7411(a)(1). What
the EPA produced as its emission guidelines was a chart that
prescribed heat-rate improvement “ranges” for each of the
EPA’s chosen heat-rate improvement technologies, organized
by power plants of differing sizes. ACE Rule, 84 Fed. Reg. at
32,537. The ranges show how much heat-rate improvement
can be “expected” from use of each of the identified
technologies. Id.
 37
The EPA was explicit, though, that the “potential” range
of heat-rate reduction was only illustrative and that the actual
reduction for each of the EPA’s chosen technologies would be
“unit-specific” and would “depend upon a range of unit-
specific factors.” ACE Rule, 84 Fed. Reg. at 32,537–32,538.
In that way, the ACE Rule made States responsible for
evaluating “[heat-rate improvement] potential, technical
feasibility, and applicability for each of the [best system of
emission reduction] candidate technologies” on a power-plant–
by–power-plant basis. Id. at 32,538. The ACE Rule expressly
left States free to establish their own standards of performance
for their power plants that “reflect a value of [heat-rate
improvement] that falls outside” the ranges provided in the
EPA’s chart. Id. (emphasis added). In other words, the
minimums listed in the EPA’s emission-reduction chart were
only suggestions.

The EPA explained that its non-mandatory ranges of
efficiency reduction were valid because the applicability of the
heat-rate improvement techniques to different plants and the
effectiveness of each power plant’s existing technology may
vary. See ACE Rule, 84 Fed. Reg. at 32,538 (stating that “not
all” of the technologies would be “applicable or warranted at
the level of a particular facility due to source-specific factors
such as the site-specific operational and maintenance history,
the design and configuration, [or] the expected operating
plans”).

The EPA predicted that its ACE Rule would reduce carbon
dioxide emissions by less than 1% from baseline emission
projections by 2035. J.A. 1651. That calculation did not
reflect emission increases that could result from the rebound
effect.
 38
4. Implementing Regulations

The ACE Rule included some new regulations under
Section 7411(d). ACE Rule, 84 Fed. Reg. at 32,575-32,584
(codified at 40 C.F.R. pt. 60, subpart Ba). As relevant here, the
regulations significantly extend the States’ deadlines for the
development and submittal of their plans for emission
reduction from nine months to three years. See 40 C.F.R.
§ 60.23a(a)(1). Similarly, the new regulations extend the
EPA’s deadline to act on those plans from four months to one
year. 40 C.F.R. § 60.27a(b). The new regulations also extend
the EPA’s deadline to substitute its own plan for a non-
compliant State’s plan from six months after the submission
deadline to two years after a finding that the plan was
incomplete, disapproved, or unsubmitted. See 40 C.F.R.
§ 60.27a(c). Finally, the requirement that States demonstrate
compliance progress is now triggered only where a State’s
compliance schedule stretches more than two years from when
its plan was originally due, as opposed to the one-year period
in the prior regulations. See 40 C.F.R. § 60.24a(d).

E. PETITIONS FOR REVIEW

Twelve petitions for review of the ACE Rule were timely
filed in this court and consolidated in this case. Nos. 19‑1140
(lead case), 19‑1165, 19‑1166, 19‑1173, 19‑1175, 19‑1176,
19‑1177, 19‑1179, 19‑1185, 19‑1186, 19‑1187, 19‑1188. The
petitioners fall into three groups.

The first grouping consists of petitioners who seek review
of the ACE Rule’s conclusion that Section 7411 only permits
emission reduction measures that can be implemented at and
applied to the source. Those petitioners include (i) a coalition
of State and municipal governments; (ii) power utilities;
 39
(iii) trade associations from the renewable energy industry; and
(iv) several public health and environmental advocacy groups.5

The second grouping is petitioners who challenge the ACE
Rule’s imposition of any emission limits as unlawful because,
in their view, (i) the EPA failed to make a specific
endangerment finding for carbon dioxide emitted from existing
power plants; (ii) the EPA’s regulation of mercury emissions
from coal-fired power plants under Section 7412 precludes the
regulation of greenhouse gas emissions under Section 7411;
and (iii) the EPA should have regulated carbon dioxide from
stationary sources, including power plants, under the NAAQS
program, 42 U.S.C. §§ 7408–7410.

The third petitioner group is the Biogenic CO2 Coalition.
They object only to the ACE Rule’s determination that States
may not count biomass co-firing as a method of complying
with numerical emission limits.

F. JURISDICTION AND STANDARD OF REVIEW

This court has jurisdiction to review these petitions under
the Clean Air Act. 42 U.S.C. § 7607(b)(1); see also Sierra
Club v. EPA, 955 F.3d 56, 61 (D.C. Cir. 2020).

We may set aside the ACE Rule if it is “arbitrary,
capricious, an abuse of discretion, or otherwise not in
accordance with law.” 42 U.S.C. § 7607(d)(1)(C),
(d)(9)(A); see also Maryland v. EPA, 958 F.3d 1185, 1196
(D.C. Cir. 2020) (“[W]e apply the same standard of review
under the Clean Air Act as we do under the Administrative

5
The public health and environmental advocacy groups also
challenge the third prong of the ACE Rule—the new implementing
regulations—as arbitrary and capricious.
 40
Procedure Act.”) (quoting Allied Local & Reg’l Mfrs. Caucus
v. EPA, 215 F.3d 61, 68 (D.C. Cir. 2000)).

II. SECTION 7411

A. STATUTORY CONTEXT

In enacting the Clean Air Act, “Congress delegated to EPA
the decision whether and how to regulate carbon-dioxide
emissions from powerplants.” American Elec. Power Co. v.
Connecticut (AEP), 564 U.S. 410, 426 (2011). As the Supreme
Court has observed, 42 U.S.C. § 7411 “speaks directly to” and
outlines the framework for that regulation. Id. at 424 (internal
quotation marks omitted). Specifically, Section 7411 marks
out a pair of distinct regulatory tracks for stationary sources of
air pollutants. See 42 U.S.C. § 7411(a)(2), (6). The first track
applies to new sources, id. § 7411(b), and the second to existing
sources, id. § 7411(d). The statute calls for federal-state
cooperation in regulating existing sources, affording distinct
roles to the federal and state agencies in arriving at what
Section 7411 calls “standards of performance” for the emission
of air pollutants. Id. § 7411(a)(1), (c), (d)(1).

The regulatory regimes for new and existing sources differ
in the process by which such standards are established—and
the roles played by the respective regulatory actors. The Act
assigns the EPA the main regulatory role in specifying the new-
source pollution controls: After the EPA determines that a
particular “category of sources * * * causes, or contributes
significantly to, air pollution which may reasonably be
anticipated to endanger public health or welfare,” it publishes
regulations establishing standards of performance for new
sources in that category. Id. § 7411(b)(1).

The process for regulating existing sources—which raise
distinct concerns about sunk costs and the health and
 41
environmental effects of older processes—involves more
actors and steps. Regulation of a given category of existing
sources is triggered by the same EPA air-pollution
determination as for new sources. But for existing sources the
Act adopts a cooperative-federalism approach that leaves the
States discretion in determining how their State and industry
can best meet quantitative emissions guidelines established by
the EPA. See AEP, 564 U.S. at 424. Under Section 7411(d),
the EPA and the States thus have distinct but complementary
roles subject to different procedures and limitations. See 42
U.S.C. § 7411 (a)(1), (d)(1). This case concerns the mechanics
of that cooperative framework for existing sources and,
specifically, restrictions the Agency now claims the statute
imposes on regulation of the air pollutants those sources emit.

Two provisions of Section 7411 shape the existing-source
framework. Subsection (a)(1) defines a standard of
performance, by reference to the “degree of emission
limitation” that the EPA determines is “achievable,” as:

a standard for emissions of air pollutants which
reflects 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 environmental impact and energy
requirements) the Administrator determines has been
adequately demonstrated.

42 U.S.C. § 7411(a)(1).

Subsection (d)(1), in turn, requires the Secretary to set up
a system by which willing States can submit to the EPA “a plan
which [] establishes standards of performance for any existing
source.” Id. § 7411(d)(1). Only “where [a] State fails to submit
a satisfactory plan” may the EPA step in and directly
 42
promulgate standards of performance for existing sources. Id.
§ 7411(d)(2).

Putting these two provisions together results in what are
best understood as three distinct steps involving three sets of
actors, each exercising a degree of leeway in choice of control
measures. See ACE Rule, 84 Fed. Reg. at 32,533, 32,549–
32,550; Clean Power Plan, 80 Fed. Reg. at 64,665–64,666.

First, under subsection (a)(1), the EPA determines the
“best system of emission reduction” that is “adequately
demonstrated,” taking into consideration certain enumerated
statutory criteria: cost, any nonair quality health and
environmental impacts, and energy requirements. 42 U.S.C.
§ 7411(a)(1). The Agency then issues emission guidelines that
quantify the “degree of emission limitation achievable through
the application of the best system” it has identified. Id.; 40
C.F.R. § 60.22a; see AEP, 564 U.S. at 424; EPA Br. 21–22;
ACE Rule, 84 Fed. Reg. at 32,523, 32,551.

Second, under subsection (d)(1), States issue standards of
performance for existing sources that comply with the EPA’s
emission guidelines and “reflect” the achievable degree of
emission limitation set in those guidelines. AEP, 564 U.S. at
424; 42 U.S.C. § 7411(d)(1); 40 C.F.R. § 60.23a; see also
Clean Power Plan, 80 Fed. Reg. at 64,666. That the standards
must “reflect” the emission guidelines does not mean that they
must embody the methods EPA contemplated in identifying the
best system; rather, the States have flexibility in determining
the specifics of the standards they issue so long as they
accomplish the “degree of emission limitation” the EPA
calculated based on its “best system.”

Third, the operators of regulated stationary sources
implement measures to ensure they will in practice comply
with the standards of performance their state agency has
 43
established for them. See ACE Rule, 84 Fed. Reg. at 32,555.
States often grant regulated entities some discretion in how
they meet those standards. See, e.g., N.Y. COMP. CODES R. &
REGS. Tit. 6 § 201-6.4(f) (2013) (describing the “operational
flexibility” afforded to Title V facility owners in New York
State to “propose a range of operating conditions that will allow
flexibility [for a facility] to operate under more than one
operating scenario”).

The issue before us arises at the first step—the EPA’s
determination of the best system of emission reduction. In the
Clean Power Plan, the Agency determined that the best system
was one that both improved the heat rate at power plants and
prioritized generation from lower-emitting plants ahead of
high-emitting plants. Clean Power Plan, 80 Fed. Reg. at
64,707. The EPA then calculated specific emission reductions
achievable through application of that best system that it
published as emission guidelines for States. Id. Had the Clean
Power Plan gone into effect, States would then have submitted
to the EPA plans based on the Agency’s guidelines that
established standards of performance for sources in their
jurisdictions, as provided for in subsection (d)(1). The Clean
Power Plan left States flexibility in the measures they included
in their plans, so long as they achieved a reduction in emissions
at least as great as that achieved by EPA-established
quantitative guidelines. See, e.g., id. at 64,665, 64,756–64,757,
64,734–64,737, 64,832–64,837. And it further allowed States,
at their option, to give leeway to sources to select alternate
compliance measures to make the requisite reductions. See id.
at 64,834–64,835.

Based on what it now perceives to be an express and
unambiguous textual limitation in Section 7411 that it says the
Clean Power Plan overlooked, the EPA repealed that Plan and
replaced it with the ACE Rule. The EPA’s new reading of the
 44
statute requires the Agency, in modeling its “best system of
emission reduction,” to consider only emission-reduction
measures that “can be applied at and to a stationary source.”
ACE Rule, 84 Fed. Reg. at 32,534; see also id. at 32,526–
32,532.

We address below the EPA’s arguments regarding how the
text and structure of Section 7411 purportedly support this
limitation. That discussion is necessarily somewhat abstract
and technical. So, for starters, it is worth bringing the matter
more concretely into view.

Consider the effect the EPA’s new statutory interpretation
had on its resulting Rule. First, because generation shifting is
not, in the EPA’s view, a measure that can be applied “at and
to” any one individual source, the ACE Rule limits the best
system of emission reduction to heat-rate improvements alone.
84 Fed. Reg. at 32,534–32,535. Then, instead of publishing
emission guidelines quantifying emission reductions
achievable through application of the best system, the ACE
Rule identifies what the Agency has determined are the most
effective heat-rate technologies available and a potential range
of heat-rate improvements achievable through application of
each of those technologies. Id. 32,535–32,537.

As under the Clean Power Plan, the ACE Rule grants
States flexibility in establishing standards of performance for
sources pursuant to the Agency’s emission guidelines. Unlike
the Clean Power Plan, however, the ACE Rule does not require
that the States reach any specified minimum emission
reduction. Instead, States must merely “evaluate the
applicability of each of the candidate technologies” to sources
within their jurisdiction and report their conclusions back to the
Agency. ACE Rule, 84 Fed. Reg. at 32,550, 32,538–32,561.

The Rule recites that regulated entities have “broad
 45
discretion” in meeting state-established standards, ACE Rule,
84 Fed. Reg. at 32,555, yet at the same time the Rule deems
impermissible any compliance measure that cannot be applied
at and to the source, id. The ACE Rule thereby disqualifies
compliance by, for example, burning biofuel, id. at 32,557–
32,558, which emits recently captured carbon dioxide, in
contrast to fossil fuels’ release of carbon dioxide stored away
millions of years ago. See generally Center for Biological
Diversity v. EPA, 722 F.3d 401, 405–406 (D.C. Cir. 2013).

The question here is a relatively discrete one. We are not
called upon to decide whether the approach of the ACE Rule is
a permissible reading of the statute as a matter of agency
discretion. Instead, the sole ground on which the EPA defends
its abandonment of the Clean Power Plan in favor of the ACE
Rule is that the text of Section 7411 is clear and unambiguous
in constraining the EPA to use only improvements at and to
existing sources in its best system of emission reduction.

The EPA contends that its current interpretation is “the
only permissible interpretation of the scope of the EPA’s
authority.” ACE Rule, 84 Fed. Reg. at 32,535. Our task is to
assess whether Section 7411 in fact compels the EPA’s new
interpretation. And because “deference to an agency’s
interpretation of a statute is not appropriate when the agency
wrongly believes that interpretation is compelled by
Congress,” Peter Pan Bus Lines, Inc. v. Fed. Motor Carrier
Safety Admin., 471 F.3d 1350, 1354 (D.C. Cir. 2006) (quoting
PDK Labs., Inc. v. DEA, 362 F.3d 786, 798 (D.C. Cir. 2004)
(internal quotation marks omitted)), we may not defer to the
EPA’s reading if it is but one of several permissible
interpretations of the statutory language, see Negusie v. Holder,
555 U.S. 511, 521 (2009). That is, the “regulation must be
declared invalid, even though the agency might be able to adopt
the regulation in the exercise of its discretion, if it ‘was not
 46
based on the agency’s own judgment but rather on the
unjustified assumption that it was Congress’ judgment that
such a regulation is desirable” or required. Prill v. NLRB, 755
F.2d 941, 948 (D.C. Cir. 1985) (quoting FCC v. RCA
Commc’ns, 346 U.S. 86, 96 (1953) (formatting modified));
accord Arizona v. Thompson, 281 F.3d 248, 259 (D.C. Cir.
2002) (quoting Prill, 755 F.2d at 948).

For the reasons explained below, Section 7411 does not,
as the EPA claims, constrain the Agency to identifying a best
system of emission reduction consisting only of controls “that
can be applied at and to a stationary source.” ACE Rule, 84
Fed. Reg. at 32,534. The EPA here “failed to rely on its own
judgment and expertise, and instead based its decision on an
erroneous view of the law.” Prill, 755 F.2d at 956. We
accordingly must vacate and remand to the Agency “to
interpret the statutory language anew.” Peter Pan Bus Lines,
471 F.3d at 1354.

1. Text

As just noted, Section 7411 contemplates distinct roles for
the EPA and the States in regulating existing stationary
sources. See 42 U.S.C. § 7411(a)(1) (granting authority to the
EPA to designate the best system and determine achievable
degree of emissions reduction); id. § 7411(d)(1) (outlining the
States’ role in setting standards of performance for their
sources). Nevertheless, the EPA now contends that language
in Section 7411(a)(1) and (d)(1) “unambiguously limits the
[best system of emission reduction] to those systems that can
be put into operation at a building, structure, facility, or
installation.” ACE Rule, 84 Fed. Reg. at 32,524 (emphasis in
original); see id. at 32,528; EPA Br. 70.

In the Agency’s current view, the only pollution-control
methods the Administrator can consider in selecting the “best
 47
system of emission reduction” within the meaning of Section
7411(a) are add-ons or retrofits confined to the level of the
individual fossil-fuel-fired power plant. ACE Rule, 84 Fed.
Reg. at 32,524. That is so even though the record before the
EPA shows that generation shifting to prioritize use of the
cleanest sources of power is one of the most cost-effective
means of reducing emissions that plants have already adopted
and that have been demonstrated to work, and that generation
shifting is capable of achieving far more emission reduction
than controls physically confined to the source. See, e.g., Clean
Power Plan, 80 Fed. Reg. at 64,693, 64,728–64,729; 2 J.A. 598;
Grid Experts Amicus Br. 13–16. In other words, the EPA reads
the statute to require the Agency to turn its back on major
elements of the systems that the power sector is actually and
successfully using to efficiently and cost-effectively achieve
the greatest emission reductions. See Grid Experts Amicus Br.
22 (observing that the ACE Rule “imposes greater abatement
costs on industry than other approaches would to achieve the
same effect”).

The Clean Power Plan could not stand, the EPA now
concludes, because its consideration of generation shifting
exceeded the Agency’s narrow authority under Section 7411’s
plain text. ACE Rule, 84 Fed. Reg. at 32,526–32,527. In
promulgating the Clean Power Plan, the EPA read “system of
emission reduction” to mean “a set of measures that work
together to reduce emissions and that are implementable by the
sources themselves.” Clean Power Plan, 80 Fed. Reg. at
64,762. And it concluded that both heat-rate improvements
and generation shifting “are components of a best system of
emission reduction for the affected [electricity generating
units] because they entail actions that the affected [units] may
themselves undertake that have the effect of reducing their
emissions.” Id. at 64,709 (internal quotation marks omitted).
 48
All of that is wrong, the EPA has since decided. “[T]he
Agency now recognizes that Congress ‘spoke to the precise
question’ of the scope of [42 U.S.C. § 7411](a)(1) and clearly
precluded the unsupportable reading of that provision asserted
in the [Clean Power Plan].” ACE Rule, 84 Fed. Reg. at 32,527.
The EPA insists that its current reading is mandated by the
statutory text.

It is the EPA’s current position that is wrong. Nothing in
Section 7411(a)(1) itself dictates the “at and to the source”
constraint on permissible ingredients of a “best system” that the
Agency now endorses. For the EPA to prevail, its reading must
be required by the statutory text. Peter Pan Bus Lines, 471
F.3d at 1354. It fails for at least three reasons, any of which is
alone fatal.

First, the plain language of Section 7411(a)(1), the root of
the EPA’s authority to determine the best system, announces
its own limitations. Those limitations simply do not include
the source-specific caveat that the EPA now interposes and
casts as unambiguous.

Second, there is no basis—grammatical, contextual, or
otherwise—for the EPA’s assertion that the source-specific
language of subsection (d)(1) must be read upstream into
subsection (a)(1) to equate the EPA’s “application of the best
system” with the controls States eventually will apply “at and
to” an individual source. As the EPA at times acknowledges,
the two subsections address distinct steps in the regulatory
process, one focused on the EPA’s role and the other focused
on the States’. Any question as to which limitations pertain to
each regulatory actor cannot reasonably be said to have been
resolved by Congress in favor of the unambiguous meaning the
EPA now advocates.

Third, even if subsections (a)(1) and (d)(1) were read
 49
together in the way the EPA proposes, they would not confine
the EPA to designating a best system consisting of at-the-
source controls. The EPA’s entire theory hinges on the
Agency’s unexplained replacement of the preposition “for” in
“standards of performance for any existing source” with the
prepositions “at” and “to.” Yet the statutory text calls for
standards of performance “for” existing sources. Emission-
reduction measures “for” sources may readily be understood to
go beyond those that apply physically “at” and “to” the
individual source. Emissions trading, for example, might be a
way “for” a source to meet a standard of performance.

The shortcomings of its statutory interpretation are more
than enough to doom the Agency’s claim that Section 7411
announces an unambiguous limit on the best system of
emission reduction. The issue is not whether the EPA’s
counterarguments to each of these points might show its
interpretation to be permissible as an exercise of discretion.
Again, the EPA has not claimed to be exercising any such
discretion here. It insists instead that the unambiguous terms
of the statute tie its hands.

After reviewing what Section 7411 clearly says about the
nature and limits of the “best system of emission reduction”
that Congress called on the EPA to determine, we take up each
of the EPA’s arguments to show why Section 7411 does not
unambiguously support its at-the-source restriction.

a. Section 7411(a) Defines the Best System

The EPA acknowledges, as it must, that Section 7411(a) is
the source of the EPA’s authority and responsibility to
determine the best system of emission reduction for existing
sources and set corresponding emission guidelines. See, e.g.,
ACE Rule, 84 Fed Reg. at 32,534. Indeed, that is the only
subsection in which the term “best system of emission
 50
reduction” appears. But the EPA offers no reading of
subsection (a)(1) itself.

Section 7411(a)(1) expresses Congress’ expectation that
the EPA will study all “adequately demonstrated” means of
emission reduction. And it directs the EPA to draw on
“adequately demonstrated” methods to determine the “best”
system to reduce emissions. Congress imposed no limits on
the types of measures the EPA may consider beyond three
additional criteria: cost, any nonair quality health and
environmental impacts, and energy requirements. 42 U.S.C.
§ 7411(a)(1). Congress largely called on the expert judgment
of the EPA to determine for a particular source category and
pollutant which already-demonstrated methods compose the
“best system.”

Because it did not set out separate definitions for either
“system” or “best,” those words take their ordinary meanings.
See Sandifer v. United States Steel Corp., 571 U.S. 220, 227
(2014). Webster’s Dictionary offers a representative definition
of “system” contemporaneous with the Act’s adoption: “[A]
complex unity formed of many often diverse parts subject to a
common plan or serving a common purpose.” System,
WEBSTER’S THIRD NEW INTERNATIONAL D ICTIONARY OF THE
ENGLISH LANGUAGE UNABRIDGED 2322 (2d ed. 1968). The
superlative “best” as applied to a “system of emission
reduction” plainly places a high priority on efficiently and
effectively reducing emissions. See Best, Merriam-Webster,
https://www.merriam-webster.com/dictionary/best (last visited
Jan. 11, 2021) (“excelling all others,” “offering or producing
the greatest advantage, utility, or satisfaction”).

The ordinary meanings of these terms “reflect[] an
intentional effort to confer the flexibility necessary” for
effective regulation appropriate to the context. Massachusetts
 51
v. EPA, 549 U.S. 497, 532 (2007). As the Supreme Court has
acknowledged, “the degree of agency discretion that is
acceptable varies according to the scope of the power
congressionally conferred.” Whitman v. American Trucking
Ass’n, 531 U.S. 457, 475 (2001); see Gaughf Props., L.P. v.
Commissioner, 738 F.3d 415, 424–425 (D.C. Cir. 2013);
Sabre, Inc. v. Department of Transp., 429 F.3d 1113, 1122,
1124–1125 (D.C. Cir. 2005). Congress in Section 7411
deliberately charged the EPA with identifying the best system
of emission reduction to keep pace with escalating threats to air
quality, and, within expressed limits, empowered it to make the
judgments how best to do so.

The Agency simply ignores how the statutory text defines
the “best system of emission reduction,” asserting instead that
definitional language does not confer regulatory authority. See,
e.g., EPA Br. 58–59 (“[I]t is not Section 7411(a) (‘Definitions’)
that grants the agency authority to act.”). Section 7411(a)(1)’s
designation as a definitional provision deprives it of standalone
meaning, the EPA contends. The EPA instead reads it as
“subsidiary” to Section 7411(d), regarding state standards of
performance for existing sources. EPA Br. 58. But Congress
does indeed use definitional provisions to confer regulatory
authority. See, e.g., Weinberger v. Bentex Pharm, Inc., 412
U.S. 645, 652–653 (1973) (holding that the statutory definition
of “new drug” confers authority upon the FDA). That is
precisely what it did in Section 7411(a)(1). See Sierra Club v.
Costle, 657 F.2d 298, 321 (D.C. Cir. 1981) (describing Section
7411(a)(1) as authorizing the EPA to determine the best system
of emission reduction and regulate accordingly); 40 C.F.R.
60.22a.

The EPA offers no support—apart from its own new-
found version of “statutory interpretation 101,” EPA Br. 65—
for ignoring how the Act itself defines and limits the “best
 52
system” determination. Nor does it offer any sound
justification for importing language from a different provision
governing States’ “standards of performance.” The EPA’s “at
and to the source” limitation on “best system” finds no footing
in the text of Section 7411(a)(1).

b. Section 7411(d)(1) Does Not Change the Definition

Even taking the EPA’s argument on its own terms does not
work because Section 7411(d)(1)’s text and statutory context
get it no further. To support its narrow reading of the EPA’s
authority to determine the “best system,” the Agency focuses
on the phrase “through the application of” in Section
7411(a)(1). That provision defines a “standard of
performance” as an emission standard that “reflects the degree
of emission limitation achievable through the application of the
best system of emission reduction[.]” The EPA says the
“application” phrase “requires both a direct object and an
indirect object.” ACE Rule, 84 Fed. Reg. at 32,524; accord
EPA Br. 66–68. And, it continues, Congress cannot have
meant to leave its indirect object undefined. The EPA says that,
grammatically speaking, someone must apply something (the
direct object) to something else (the indirect object). EPA Br.
115–116, 118–119. It then picks its preferred, narrow indirect
object from a different statutory subsection and casts that
object as the only statutorily permissible choice. See 84 Fed
Reg. at 32,524.

The EPA locates an indirect object in Section 7411(d).
Unlike subsection (a)(1), subsection (d)—entitled “Standards
of performance for existing sources”—explicates an indirect
object. 42 U.S.C. § 7411(d). Borrowing from subsection (d),
then, the EPA imports into subsection (a)(1) a limitation of the
“best system of emission reduction” to measures that can be
applied “to and at an individual existing source—i.e., any
 53
building or facility subject to regulation.” EPA Br. 58
(emphasis added); see also ACE Rule, 84 Fed. Reg. at 32,534.

But the language to which the EPA points supplies the
indirect object only of “standards of performance” adopted by
States pursuant to Section 7411(d)(1), not of the EPA’s “best
system of emission reduction” determined pursuant to Section
7411(a)(1). The latter phrase does not even appear in Section
7411(d)(1). To reach its preferred result, the Agency invokes
surmise rather than statutory text. It insists that the limitations
on States’ standards of performance in Section 7411(d)(1)—
the second step in the regulatory process—must be read
upstream to limit the EPA’s “best system of emission
reduction” in subsection (a)(1). Nothing in the statute so
requires.

In the text, States’ standards of performance need only
“reflect” the emission guidelines (or “degree of emission
limitation achievable”) the EPA calculates based on the “best
system of emission reduction” it determines. As laid out in the
statute and explained above, those state-developed “standards
of performance” follow on but are legally and functionally
distinct from the “best system” that the EPA develops. The
EPA is simply wrong that the statute clearly and
unambiguously requires that the unstated indirect object of
“application of the best system of emission reduction” under
Section 7411(a)(1) must be the same as the indirect object of
States’ standards of performance as stated in Section
7411(d)(1).

Neither does the grammatical rule the EPA invokes to
bridge the gap between these subsections hold up. The crux of
the EPA’s textual argument is that “the verb ‘to apply,’ requires
both a direct object and an indirect object.” ACE Rule, 84 Fed.
Reg. at 32,524; EPA Br. 66–68. The first obvious problem is
 54
that, in the relevant passage of Section 7411(a)(1), Congress
did not use the verb “apply,” but rather the noun “application.”
The EPA acknowledges this distinction in passing in the ACE
Rule, but dismisses it without discussion, offering only that
“‘application’ is derived from the verb ‘to apply[.]’” 84 Fed.
Reg. at 32,524. That is, of course, true, as far as it goes. The
phrase “application of the best system of emission reduction”
is what is called a nominalization, a “result of forming a noun
or noun phrase from a clause or a verb.” Nominalization,
Merriam-Webster Dictionary https://www.merriam-
webster.com/dictionary/nominalization (last visited Jan. 11,
2021). Grammar assigns direct or indirect objects only to
verbs—not nouns. No objects are needed to grammatically
complete the actual statutory phrase. So much for the
grammatical imperative.

Even if we were to take the EPA’s leap to the verb “apply”
from the noun “application” that actually appears in the statute,
the Agency comes up short. The EPA is incorrect to insist that
the verb “apply” requires an indirect object. There is nothing
ungrammatical about the sentence “In its effort to reduce
emissions, the EPA applied the best system of emission
reduction.” The verb “apply,” like its nominalization, may
properly be used in a sentence with or without an explicit
indirect object. See Apply, THOMAS HERBST ET AL., A
VALENCY DICTIONARY OF ENGLISH 41–42 (Ian F. Roe et al.
eds., 2004) (listing examples of grammatically correct uses
with and without direct and indirect objects).6

6
Take, for instance, the following sentences: “It appears to
violate GATT regulations, but the rules for applying the regulations
are vague and the Netherlands has so far escaped censure”; “This
information may not apply in Scotland, which has a different legal
system.” Apply, THOMAS HERBST ET AL., A VALENCY DICTIONARY
 55
The EPA’s shift from nominalization to verb does not, in
any event, accomplish much. Either way, the lack of an explicit
indirect object in Section 7411(a)(1) does not require that one
be borrowed from Section 7411(d)(1). Equally logical indirect
objects include, for example, the entire category of stationary
sources, or the air pollutant to be limited. In any event, the best
system cannot reasonably be said to be unambiguously
applicable only to the indirect object the EPA suggests.

The EPA faults the Clean Power Plan for reading
“application of” to be functionally equivalent to
“implementation of,” because “implement” “does not require
an indirect object.” EPA Br. 73. But neither does
“application.” So “application” textually supports adoption of
the Clean Power Plan just as well as “implementation.” Again,
so much for grammar mandating the EPA’s result.

The argument fails either way, but the fact is that Congress
used the nominalization “application of” the best system of
emission reduction. A nominalization enables the drafter to
leave certain information unspecified—namely, who is acting
and where their action is directed. See, e.g., George D. Gopen,
Who Done It? Controlling Agency in Legal Writing, Part II,
39 LITIG . 12, 12–13 (Spring 2013) (describing how
nominalizations create ambiguity). Legal writers, including
Congress, employ nominalizations all the time. And they do
so with the full awareness that their use preserves flexibility.

Congress reasonably built in leeway for the EPA to

OF ENGLISH 41–42 (examples from sections D1 and D5). Additional
examples abound. See, e.g., Apply, OXFORD ENGLISH DICTIONARY
(3d ed. 2008) (def. I.9) (“Crest bought the firm[,] and, by applying
its marketing and distribution muscle, has turned it into a $200
million category killer.”).
 56
exercise technical expertise in applying Section 7411, given the
variety of pollution problems that it covers and the importance
of allowing States maneuvering room under the cooperative
federalism scheme. Congress may avoid specifying subjects,
objects, or other grammatical information because a degree of
adaptability suits the statutory role and purpose. One way
Congress can denote that it has delegated to an agency’s
judgment the task of filling in the on-the-ground details of a
statutorily defined program is by declining to dictate
grammatically optional information, see Lehrfeld v.
Richardson, 132 F.3d 1463, 1465–1466 (D.C. Cir. 1998);
Appalachian Power Co. v. EPA, 135 F.3d 791, 808–810 (D.C.
Cir. 1998), including an indirect object that the rules of
grammar do not require be explicitly stated, see, e.g., Peter Pan
Bus Lines, 471 F.3d at 1353–1354.

Even if an implicit indirect object can be surmised, there
is more than one plausible candidate here, and the statute does
not unambiguously dictate the object. There certainly is no
rule—grammatical or otherwise—that the specific indirect
object must be the one to which the EPA now points. At the
least, other contextually appropriate indirect objects of the
“best system” include the source category or the emissions.
The EPA has failed to establish that the sole and unambiguous
indirect object must be the individual source. The EPA, of
course, “may fill the gap[s] the Congress left,” and any such
“regulation is entitled to deference.” Gaughf Props., 738 F.3d
at 424; see also Appalachian Power, 135 F.3d at 811–812. But
in the ACE Rule and in its briefing here, the EPA has
assiduously denied the existence of any gap at all. That was
error.

c. EPA’s Reading Itself Falls Short

The third and equally fatal flaw in the EPA’s textual
 57
analysis is its unexplained substitution of the prepositions “at”
and “to” where the text it would have us borrow from
subsection (d)(1) actually says “for” in referencing “standards
of performance for any existing source.” See, e.g., ACE Rule,
84 Fed. Reg. at 32,534. As we do with any words enacted by
Congress, we must give effect to the preposition it chose.
Cf. Telecommunications Res. & Action Ctr. v. FCC, 801 F.2d
501, 517–518 (D.C. Cir. 1986) (finding decisive Congress’ use
of the preposition “under” instead of “by”). The word
Congress actually used—“for” the source—lacks the site-
specific connotation on which the EPA’s case depends.

In its brief, the EPA presents the compound construction
it says inexorably follows from reading text from subsection
(a)(1) together with text from subsection (d)(1), and says it is
restricted to determining a “best system of emission reduction
for any building, structure, facility, or installation.” EPA
Br. 56 (formatting modified) (quoting 42 U.S.C. § 7411(a)(1),
(a)(3), (a)(6), (d)(1)). The Agency then asserts that “the natural
reading” of its proffered construction is that “the methods
planned would be ‘for’ and act at the level of the singular,
individual source.” Id. at 62 (emphasis added).

In the preamble to the ACE Rule, the EPA went further,
fully substituting the prepositions “at” and “to” in place of the
preposition “for” that actually appears in the text the Agency
says must be borrowed from subsection (d)(1). ACE Rule, 84
Fed. Reg. at 32,534. It relies on that further substitution to
insist that the best system of emission reduction designated by
the EPA must be limited to controls “that can be applied at and
to,” not “for,” “a stationary source.” Id.; see also id. at 32,524
(“at”); id. at 32,532, 32,534, 32,556 (“at and to”); id. at 32,555,
32,529 (“to and at”); id. at 32,543 (“at or to”); id. at 32,526 n.65
(“to or at”); EPA Br. 4, 58, 74. But nowhere in the ACE Rule
does the EPA explain this swap of one preposition for two
 58
meaningfully more restrictive ones. See, e.g., 84 Fed. Reg. at
32,523–32,524, 32,534–32,535.

The EPA rewrites rather than reads the plain statutory text.
Section 7411(a)(1), even if cross-referenced to subsection
(d)(1) in the way the EPA says it must be, calls for the Agency
to determine “the degree of emission limitation achievable
through the application of the best system of emission
reduction for any existing source”—not the application of the
best system “at” and “to” such a source. And the word “for”
lacks the physical on-site connotation that is so critical to the
EPA’s reading of the statutory text. Indeed, a standard of
performance or system of emission reduction “for” a source
just means that the system is “with regard or respect to” or
“concerning” the source. See For, OXFORD ENGLISH
DICTIONARY (2d ed. 1989) (def. 26). In contrast, “at” and “to”
tend to connote direct physical proximity or contact. See At,
OXFORD ENGLISH D ICTIONARY (3d ed. 2008) (def. 1.a)
(“usually determining a point or object with which a thing or
attribute is practically in contact”); To, OXFORD ENGLISH
DICTIONARY (3d. ed 2008) (def. 5.a) (“Into (or in) contact with;
on, against”). A best system “for” a source thus might entail a
broader array of controls that concern but are not immediately
physically proximate to the source—such as, for instance,
generation shifting.

***

In sum, the straitened vision of the EPA’s best system that
the Agency espies in Section 7411 is simply not supported by
the text, let alone plainly and unambiguously required by it.
The Act calls on the EPA to determine the degree of emission
limitation achievable through “application of the best system
of emission reduction” without specifying the system’s indirect
object, and uses the preposition “for” when it calls on the States
 59
to develop “standards of performance for existing sources.” 42
U.S.C. § 7411(a), (d). It simply does not unambiguously bar a
system of emission reduction that includes generation shifting.

The EPA’s position depends critically on words that are
not there. It erroneously treats a nominalization of a verb as
requiring an indirect object, collapses two separate functions
and provisions of the Act in order to supply a borrowed indirect
object, does so without any evidence that the borrowed indirect
object was what Congress necessarily intended, and narrowly
focuses the Agency’s authority on that indirect object by using
a different preposition from the one that actually appears in the
borrowed text. Each of those interpretive moves was a misstep.
Read faithfully, Section 7411(a)(1) lacks the straitjacket that
the EPA imposes.

Policy priorities may change from one administration to
the next, but statutory text changes only when it is amended.
The EPA’s tortured series of misreadings of Section 7411
cannot unambiguously foreclose the authority Congress
conferred. The EPA has ample discretion in carrying out its
mandate. But it may not shirk its responsibility by imagining
new limitations that the plain language of the statute does not
clearly require.

2. Statutory History, Structure, and Purpose

Even looking beyond the text does nothing to substantiate
the EPA’s proposed reading of Section 7411. See Kiewit
Power Constructors Co. v. Secretary of Labor, 959 F.3d 381,
395 (D.C. Cir. 2020) (Henderson, J.) (“To discern the
Congress’s intent, we generally examine the statutory text,
structure, purpose and its legislative history.”) (quoting
Lindeen v. SEC, 825 F.3d 646, 653 (D.C. Cir. 2016)). These
other tools of statutory interpretation underscore the flexibility
 60
of Section 7411(a)’s text, not the cabined reading the EPA
proposes.

We begin by acknowledging Section 7411’s role within
the Clean Air Act. It is a catch-all, intended to ensure that the
Act achieves comprehensive pollution control by guaranteeing
that there are “no gaps in control activities pertaining to
stationary source emissions that pose any significant danger to
public health or welfare.” S. REP. NO. 91-1196, at 20 (1970).
In other words, Section 7411 is intended to reach pollutants that
do not fit squarely within the ambit of the Act’s other
regulatory provisions. It authorizes regulation of pollutants not
controlled by the other programs under the Act. The EPA does
not contest that greenhouse gases emitted by powerplants fit
that description.

The Agency points to statutory structure and history for
evidence that Congress restricted the “best system of emission
reduction” under Section 7411(a) to physical controls that are
applied “at and to” an existing source. But the history and
structure only confirm what the text shows: Nothing the EPA
has identified suggests that Congress in Section 7411 meant to
so constrict what might be part of a “best system of emission
reduction.”

The Congress that enacted Section 7411 was well aware
that what a “best system” might comprise is necessarily
dynamic and evolving. Congress’ main limitation was that the
“best system” selected by the EPA must be “adequately
demonstrated.” 42 U.S.C. § 7411(a)(1). And it stated three
other key criteria—cost, nonair quality health and
environmental impact, and energy requirements—as factors
the EPA must take into account. See id. With those parameters
in place, Congress largely left the identification of the best
 61
system of emission reduction to the Agency’s expert scientific
judgment.

Consider cues from the Clean Air Act as a whole. In
contrast to other systemic benchmarks in the Act,
Section 7411(a)(1)’s prescription of the “best system of
emission reduction” is striking for its paucity of restrictive
language. References to more specific categories of emission-
reduction tools appear elsewhere in the Act. A provision
governing the Nitrogen Oxides Emissions Reduction Program,
for example, directs the Administrator to establish limits based
on the “degree of reduction achievable through the retrofit
application of the best system of continuous emission
reduction, taking into account available technology[.]” 42
U.S.C. § 7651f(b)(2) (emphasis added). The Act’s regional
haze program is likewise specific in its call for use of the “best
available retrofit technology.” Id. § 7491(b)(2)(A), (g)(2). The
specificity of those other provisions highlights the comparative
generality of Section 7411(a)’s reference to the “best system of
emission reduction.”

The sole provision the EPA highlights to shore up its at-
the-source theory only further undermines it. The EPA points
to the Act’s Prevention of Significant Deterioration (PSD)
program, 42 U.S.C. § 7475, and its requirement of controls at
least as stringent as limits set under Section 7411, see
id. § 7479(3), to argue that that “the interrelationship between
the two types of standards”—the best system of emission
reduction and the best available control technology—“is only
intelligible if the standards are in pari materia.” EPA Br. 85.
But the distinct roles of the two provisions make clear that the
limits in Section 7475 have no place in Section 7411(a)(1).

To qualify for a permit under the PSD program before a
source may be built or modified, an applicant must affirm that
 62
it will apply to each source the “best available control
technology,” or BACT, to limit its emissions. 42 U.S.C.
§ 7475(a)(4). The statute defines BACT as the degree of
control that the permitting agency “determines is achievable for
such [major emitting] facility through application of
production processes and available methods, systems, and
techniques, including fuel cleaning, clean fuels, or treatment or
innovative fuel combustion techniques[.]” Id. § 7479(3). The
statute further provides that BACT cannot “result in emissions
of any pollutants which will exceed the emissions allowed by
any applicable standard established pursuant to [S]ection 7411
or 7412 of this title.” Id. § 7479(3). The listed BACT options,
EPA observes, are all physically applicable to the source unit.
EPA Br. 85.

But the EPA ignores a critical detail: The BACT
requirement applies only to newly constructed or modified
sources. See Alaska Dep’t of Env’t Conservation v. EPA, 540
U.S. 461, 472 (2004) (describing 42 U.S.C. § 7475). Any
standard established under Section 7411 and also “applicable,”
per the statutory cross-reference, to a facility regulated for
prevention of significant deterioration under Section 7475
would be a standard for new or modified sources established
pursuant to Section 7411(b). The BACT requirement does not
apply to the existing sources covered by the provision at issue
here, Section 7411(d). See New York v. EPA, 413 F.3d 3, 13
(D.C. Cir. 2005). Even if Section 7475 tracks Section 7411(b),
there is simply no conflict between, on one hand, requiring new
source construction to employ the newest and best at-the-
source control technologies and, on the other, empowering the
EPA to look to a wider range of ways to reduce emissions when
it regulates older, existing sources.

The anomaly of looking to Section 7475(a)(4) to confine
Section 7411 is highlighted by the fact that BACT permits are
 63
required only in so-called “non-attainment” areas of the
country. See 42 U.S.C. §§ 7407, 7472, 7474. We are
unpersuaded that Congress buried a limit on the EPA’s
Section 7411 authority to address pollution from existing
sources throughout the Nation by making reference to a floor
for certain new facilities in certain parts of the country.

The statutory history of the BACT requirement further
demonstrates that Congress did not intend that it weaken
Section 7411(d). Sections 7475 and 7479 were enacted in the
1977 Clean Air Amendments, Pub. L. No. 95-95, §§ 165, 169,
91 Stat. 685, 735–742 (Aug. 7, 1977). In the very same
legislation, Congress restricted the best system of emission
reduction for new sources to technological methods while
explicitly allowing the best system for existing sources to
include non-technological methods. § 109(c)(1)(A), 91 Stat.
at 700. If Congress wanted to confine Section 7411 to at-the-
source technologies, it would have done so directly rather than
hiding such a substantial limitation in an implicit inference
from a more remote statutory provision.

The Clean Air Act’s legislative history, including the
history of the 1970 enactment of Section 7411 and the 1977
and 1990 amendments, further shows that Congress never
imposed on the “best system of emissions reduction” the
constraints the EPA now advocates. Before Congress settled
on the best-system language it enacted in 1970, the Senate bill
proposed to authorize the EPA to set standards for stationary
sources “reflect[ing] the greatest degree of emission control”
achievable through “the latest available control technology,
processes, operating methods, or other alternatives.” S. 4358,
91st Cong. § 6 (1970). The phrase “other alternatives” was
understood to encompass “[t]he maximum use of available
means of preventing and controlling air pollution”—without
limitation to technological or at-the-source means. S. REP. NO.
 64
91-1196, at 16. The Senate believed that was “essential” to
limit emissions from both new and existing sources. Id. The
House, for its part, proposed an initial version of Section 7411
that would have “require[d] new sources to ‘prevent and
control [their] emissions to the fullest extent compatible with
the available technology and economic feasibility,’” H.R.
17255, 91st Cong. § 5 (1970), but included no provision
regarding the regulation of existing sources.

As enacted, Section 7411 simply requires that the EPA
identify as its benchmark for existing sources the “best system
of emission reduction.” 42 U.S.C. § 7411(a)(1). Nothing that
the EPA identifies or that we discern in the relevant history
shows the enacting Congress myopically “focused on steps that
can be taken at and by individual sources to reduce emissions.”
EPA Br. 69. And of course, even if Congress at that time was
only thinking of at-the-source controls, the EPA was well
aware that environmental problems and their solutions rapidly
evolve. At the end of the day, it is the statutory text that
governs. See Bostock v. Clayton County, 140 S. Ct. 1731, 1738
(2020).

Congress has consistently relied on the EPA’s expert
judgment in identifying the “best system” for existing sources.
Its action in making, and then undoing, a limiting amendment
to Section 7411’s “best system of emission reduction” just for
new and modified sources—not existing sources—underscores
the point. First, Congress in 1977 amended the standard for
new sources to require use of “the best technological system of
continuous emission reduction,” but did not make any parallel
change to the standard for existing sources to add those
“technological” and “continuous” limitations. Clean Air
Amendments Act of 1977, Pub. L. No. 95-95, § 109(c)(1)(A),
91 Stat. 685; see also id. at 700 (adding Section 7411(a)(1)(C)).
Then, in 1990, Congress again amended Section 7411, this time
 65
to remove those additional limitations, reverting for new
sources to the “best system of emission reduction” that had
applied all along to existing sources. Clean Air Act
Amendments of 1990, Pub. L. No. 101-549, § 403(a), 104 Stat.
2399, 2631 (1990).

The amendment and re-amendment of the new-source
“best system” language emphasizes that Congress consistently
avoided imposing any such technological, at-the-source
limitation on the measures that EPA might include in the “best
system” for reducing emissions from existing-source
categories. And it shows that Congress had always understood
the existing-source “best system” language to go beyond the
technological restrictions that it briefly imposed on the parallel
new source provision.

The ACE Rule is the first EPA rule to read the statute as
so strictly boxing in the Agency. Although agency practice
cannot directly show whether Congress had a specific intent on
the matter in question, it is notable that the regulators closest to
the issue never before saw what the EPA now insists is obvious
on the face of Section 7411.

Over the last half century, no prior Administrator read the
Act to foreclose from consideration in the “best system” all but
at-the-source means of emission control. Rather, the EPA has
exercised latitude to consider any adequately demonstrated
approach to reducing harmful pollutants from existing source
categories that it believed met the cost, grid-reliability and
other statutory criteria. 42 U.S.C. § 7411(a)(1). Where the
characteristics of the source category and the pollutant at issue
point to emissions trading programs or production shifts from
higher- to lower-emitting sources as components of the “best
system,” the EPA has in the past consistently concluded that it
had the authority to consider them.
 66
During the administration of President George W. Bush,
for example, the EPA adopted the Clean Air Mercury Rule, 70
Fed. Reg. 28,606 (May 18, 2005), which included a mercury
cap-and-trade program as a component of its best system of
emissions reduction for existing coal-fired power plants, see id.
at 28,619–28,620; id. at 28,617 (“EPA has determined that a
cap-and-trade program based on control technology available
in the relevant timeframe is the best system for reducing
[mercury] emissions from existing coal-fired Utility Units.”). 7

The EPA’s Clinton-era regulation of nitrogen oxide
emissions from municipal solid waste combustors likewise
relied on Section 7411(d), together with the EPA’s waste-
management authority under Section 7429, to authorize States
to include emissions-trading programs in their State Plans. 40
C.F.R. § 60.33b(d)(2). Under state standards of performance
designed to meet guidelines the EPA derived from its “best
system,” regulated entities were permitted to average the
emission rates of multiple units within a single plant as well as
trade emission credits with other plants. Municipal Waste
Combustors Rule, 60 Fed. Reg. 65,387, 65,402 (Dec. 19,
1995).

The EPA’s efforts to distinguish those other
Section 7411(d)(1) programs do not work. The EPA claims
that the Mercury Rule did not primarily rely on a cap-and-trade

7
We vacated the Mercury Rule for unlawfully delisting
mercury-emitting electric utility steam generating units from the
Section 7412 Hazardous Air Pollutants list. See New Jersey v. EPA,
517 F.3d 574, 582–584 (D.C. Cir. 2008). Because we held those
mercury sources must be listed, and because Section 7411 cannot be
used to regulate air pollutants listed under Section 7412, the existing-
source rule the EPA had adopted under Section 7411(d) to control
those same mercury emissions from power plants failed as well.
 67
or dispatch shifting program, but rather that the best system
rested on a “combination of a cap-and-trade mechanism
and * * * the technology needed to achieve the chosen cap
level.” EPA Br. 72 n.20 (quoting ACE Rule, 84 Fed. Reg. at
32,526). To be clear, that sort of hybrid best system, involving
both on-site and system-wide elements, is precisely what the
EPA now insists is unprecedented and expressly barred by the
statute’s text.

Lest there be any doubt that the Mercury Rule’s best
system rested in significant part on the cap-and-trade
mechanism, we note that the EPA in fact approved state
implementation plans that adopted none of the on-site controls
included in the best system and instead relied entirely on
implementation of the best system’s cap-and-trade program.
See, e.g., Notice of Intent, 32 La. Reg. 869, 870 (May 20, 2006)
(proposing an implementation plan solely reliant on cap-and-
trade); Approval and Promulgation of State Plan for
Designated Facilities and Pollutants: Louisiana, 72 Fed. Reg.
46,188, 46,188 (Aug. 17, 2007) (approving Louisiana’s
proposal on the basis that it “would meet [Clean Air Mercury
Rule] requirements by participating in the EPA administered
cap-and-trade program addressing [mercury] emissions”).
Contrary to the EPA’s assertions, e.g. EPA Br. 4, the Agency
plainly has previously embraced beyond-the-source measures
of emission reduction as authorized by the statutory text.

The EPA’s invocation of its own past practice under
Section 7411 falls wide of the mark. It errs in insisting that
“the more than seventy Section 7411 rules” promulgated for
“roughly forty-five years” somehow reflect a consistent
adherence to the Agency’s new view. EPA Br. 4, 88; see id. at
37–38, 88–89; ACE Rule, 84 Fed. Reg. at 32,526. Almost all
of the rules to which it refers are irrelevant to the issue at hand.
They were for new sources, subject to Section 7411(b), not
 68
existing sources under Section 7411(d). See 84 Fed. Reg. at
32,526.

Older facilities that may be capable only of outdated, more
polluting methods of generation present different regulatory
challenges than new sources. As discussed above in
connection with the EPA’s reference to BACT requirements
for new-source permitting under the PSD program, a
requirement that owners and operators constructing new
facilities apply state-of-the-art, lowest-emitting equipment and
methods “at and to the source” might well be the best available
means of reducing emissions for that source category. The
same cannot be said for existing sources. A central error of the
ACE Rule is that it fails to appreciate that difference. It
identifies a handful of measures applicable to and at the source
that the EPA suggests may achieve slight reductions. But
industry practice demonstrates that better, lower-emitting,
reliable, and cost-effective systems for reducing emissions
from existing power plants typically also shift generation away
from higher-emitting, fossil-fuel-fired capacity when
renewable or lower- or zero-emitting generation is an available
substitute.

Because the best, most efficient and effective systems for
controlling emissions from existing sources ordinarily differ
from the best systems for new sources, they are regulated via a
distinct statutory track. Only the Section 7411(d) rules are
relevant to the EPA’s prior understanding of its authority to
regulate existing sources. Those prior EPA rules contradict the
EPA’s position here. Before its about-face in the ACE Rule,
all three of the Agency’s most recent Section 7411(d) rules
included emissions trading or generation shifting to lower-
emitting sources. See Clean Power Plan, 80 Fed. Reg. at
64,755–64,756; Clean Air Mercury Rule, 70 Fed. Reg. at
 69
28,606, 28,617, 28,619–28,620; Municipal Waste Combustors
Rule, 60 Fed. Reg. 65,387, 65,402 (Dec. 19, 1995).

To put the EPA’s mistaken reading of Section 7411 in
perspective, consider how it effectively relegates federal
regulators back to the sidelines where they stood before
Congress overhauled the Clean Air Act in 1970. The federal
government had until then done little more than provide
information and guidance to cheer on States’ air-quality
regulators. See Train v. NRDC, 421 U.S. 60, 64 (1975) (noting
that the States’ response to earlier iterations of the Act focused
on information and incentives had been “disappointing”).

With the 1970 amendments, a virtually unanimous
Congress dramatically strengthened the federal government’s
hand in combatting air pollution. See Train, 421 U.S. at 64
(“These Amendments sharply increased federal authority and
responsibility. * * * The difference * * * was that the States
were no longer given any choice as to whether they would meet
th[eir statutory] responsibility.”); cf. EPA v. EME Homer City
Generation, LP, 572 U.S. 489, 497 (2014) (noting this
progression toward “increasing[ly] rigor[ous]” federal
regulation of interstate air pollution). Congress did so “to
protect and enhance the quality of the Nation’s air resources so
as to promote the public health and welfare and the productive
capacity of its population[.]” 42 U.S.C. § 7401(b)(1). The
EPA’s newly enhanced authority was “designed to provide the
basis” for “a massive attack on air pollution.” S. REP. NO . 91-
1196, at 1. Section 7411(d) ensured that there would be “no
gaps in control activities pertaining to stationary source
emissions that pose any significant danger to public health or
welfare.” Id. at 20.

Describing the Act shortly before its passage, Republican
Senator John Cooper explained that the “philosophy of the bill
 70
abandons the old assumption of requiring the use of only
whatever technology is already proven and at hand” and
instead “set[s] out what is to be achieved.” 116 CONG. REC.
32,919 (1970). To that end, the Act did not finely detail
specific approaches to enumerated sources or types of air
pollution. See 116 CONG. REC. 32,901–32,902 (1970)
(statement of Sen. Muskie). Congress chose instead to entrust
the EPA with flexible powers to craft effective solutions. Only
by doing so could air quality regulation hope to reflect
developing understandings of escalating problems and bring to
bear as-yet-unseen solutions.

American air quality is the proof of that approach. The
EPA has worked closely with industry, States, and the public
to develop the world’s most nimble, responsive, and effective
regime of air pollution regulation. For example, in the half-
century since the 1970 Act, “the combined emissions
of * * * six key pollutants regulated under the National
Ambient Air Quality Standards dropped by 73 percent”
between 1970 and 2017. EPA Releases 2018 Power Plant
Emissions Demonstrating Continued Progress, EPA (Feb. 20,
2019), https://www.epa.gov/newsreleases/epa-releases-2018-
power-plant-emissions-demonstrating-continued-progress
(last visited Jan. 11, 2021).

The EPA’s new reading of Section 7411 would atrophy the
muscle that Congress deliberately built up. The EPA asserts it
lacks authority to curb a pollutant that the Agency itself has
repeatedly deemed a grave danger to health and welfare but that
eludes effective control under other provisions of the Act. We
do not believe that Congress drafted such an enfeebled gap-
filling authority in Section 7411.

***
 71
In sum, traditional tools of statutory interpretation reveal
nothing in the text, structure, history, or purpose of
Section 7411 that compels the reading the EPA adopted in the
ACE Rule.

3. Compliance Measures

In the ACE Rule, the EPA also limited the measures that
sources may use to comply with the States’ standards of
performance set under Section 7411(d). Recognizing that
sources generally have “broad discretion” in how they comply
with state standards, 84 Fed. Reg. at 32,555, the EPA
nonetheless categorically excluded two specific measures from
the States’ consideration: averaging and trading, and biomass
co-firing. It did so on the ground that these measures do not
meet two criteria it determined were required of compliance
measures: that they be (1) “capable of being applied to and at
the source” and (2) “measurable at the source using data,
emissions monitoring equipment or other methods to
demonstrate compliance[.]” Id. The EPA identified these
criteria on account of “both legal and practical concerns[.]” Id.

The Agency’s legal concern was that non-source-specific
compliance measures “would be inconsistent with the EPA’s
interpretation of the” best system of emission reduction as itself
plant-specific. ACE Rule, 84 Fed. Reg. at 32,555–32,556. In
that way, the EPA extended to States’ compliance measures the
same incorrect textual interpretation of the Clean Air Act that
underlay its determination of what best systems may include—
namely, that the system must be one that can be applied to and
at the individual source. The EPA reasoned that
“implementation and enforcement of such standards should
correspond with the approach used to set the standard in the
first place.” Id. at 32,556.
 72
The Agency’s practical concern was that compliance
measures that are not source-specific could result in
“asymmetrical regulation[,]” meaning the stringency of
standards could vary across sources. ACE Rule, 84 Fed. Reg.
at 32,556. It argues here that such regulation “could have
significant localized adverse consequences” in the case of
many pollutants regulated under Section 7411(d). EPA Br.
240.

Because we hold that the EPA erred in concluding Section
7411 unambiguously requires that the best system of emission
reduction be source specific, we necessarily reject the ACE
Rule’s exclusion from Section 7411(d) of compliance
measures it characterizes as non-source-specific. The Agency
tied that exclusion to its flawed interpretation of the statute as
unambiguously confined to measures taken “at” individual
plants, so it falls with that decision. ACE Rule, 84 Fed. Reg.
at 32,555–32,556.

The statute says nothing about the measures that sources
may use to comply with the standards States establish under
Section 7411(d), and the EPA cites no separate authority that
would require compliance measures to be source-specific, or
that Congress meant to so hogtie the States in devising
standards of performance. Regardless of any policy-based
reasons the EPA offers for limiting compliance measures, then,
its decision to exclude averaging and trading and biomass
co-firing is foreclosed by its legally erroneous starting point.

Neither can the EPA’s policy-based reasons sustain its
decision to exclude its disfavored non-source-specific
compliance measures in the context of carbon dioxide
emissions. Apart from its statutory interpretation, the EPA’s
only ground for excluding those compliance measures is the
Agency’s stated concern to avoid asymmetrical regulation.
 73
ACE Rule, 84 Fed. Reg. at 32,556. It argues that asymmetrical
regulation “could have significant localized adverse
consequences for public health and the environment.” EPA Br.
240. The Agency points to the case of fluoride—another
pollutant regulated under Section 7411(d)—to note that
allowing sources to meet state standards of performance by
averaging emissions across units or between facilities “could
cause serious environmental impacts on local communities
where pollution was under-controlled, causing localized
damage.” Id. In light of such considerations, the EPA worried
that a system of averaging and trading “would undermine the
EPA’s determination” of the best system of emission reduction,
leading to the sort of localized consequences the system is
designed to guard against. ACE Rule, 84 Fed. Reg. at 32,557.

But that point does not support the EPA’s categorical rule,
let alone prove that the statute unambiguously compels the
Agency’s reading. Unlike pollutants such as fluoride, carbon
dioxide emissions do not pose localized concerns at the site of
emission. Whereas the EPA might determine that the best
system for reducing fluoride emissions is one that can be
applied to and at the source, and it would be reasonable for the
EPA in turn to limit compliance measures to correspond with
such a “best system,” the same cannot be said of carbon
dioxide. Indeed, the EPA recognizes that “CO2 is a global
pollutant with global effects[,]” meaning “there may be few
direct and area public health consequences from asymmetrical
regulation of carbon dioxide within a State.” EPA Br. 239.

The Agency defends its concern about asymmetrical
regulation in the context of carbon dioxide emissions with the
unsupported contention that an interpretation of Section
7411(d) that allowed non-source-specific compliance measures
“would not be limited to carbon dioxide alone.” EPA Br. 240.
But there is no reason to conclude, and petitioners do not argue,
 74
that the statute requires the EPA to permit non-source-specific
compliance measures for every pollutant it regulates under
Section 7411. The statute is not so rigid as EPA supposes. In
fact, Section 7411 itself does not textually restrict the States’
choice of compliance measures for their sources at all. See also
Power Cos. Pet’rs Br. 25–26; Biogenic Pet’r Br. 16–17. Even
if the EPA might reasonably limit compliance measures in
specific situations based on its determination of the best system
for reducing particular types of emissions with localized
consequences, the statute imposes no requirement that such
limitations be uniform across the regulation of different
pollutants.

In sum, the EPA’s conclusion on compliance by sources
rises and falls with its legally flawed interpretation of the
statute. The Agency’s practical concern about asymmetrical
regulation could not, in any event, support the exclusion of
biomass co-firing or averaging and trading in the particular
context of carbon dioxide emission regulation.

B. T HE MAJOR QUESTIONS D OCTRINE

The EPA also references the so-called “major questions”
doctrine in defense of its statutory interpretation and the ACE
Rule. 84 Fed. Reg. at 32,529. But that doctrine does not
confine the EPA to adopting solely emission standards that can
be implemented physically to and at the individual plant.

The Supreme Court has said in a few cases that sometimes
an agency’s exercise of regulatory authority can be of such
“extraordinary” significance that a court should hesitate before
concluding that Congress intended to house such sweeping
authority in an ambiguous statutory provision. See King v.
Burwell, 576 U.S. 473, 485–486 (2015); Gonzales v. Oregon,
546 U.S. 243, 262, 266–267 (2006); FDA v. Brown &
Williamson Tobacco Corp., 529 U.S. 120, 159 (2000); accord
 75
Utility Air Regulatory Group v. EPA (UARG), 573 U.S. 302,
324 (2014); see also MCI Telecommc’ns v. AT&T, 512 U.S.
218, 231 (1994). Where there are special reasons for doubt, the
doctrine asks whether it is implausible in light of the statute and
subject matter in question that Congress authorized such
unusual agency action. See, e.g., UARG, 573 U.S. at 324
(considering whether the challenged rule would “bring about
an enormous and transformative expansion in EPA’s
regulatory authority without clear congressional
authorization”); Brown & Williamson, 529 U.S. at 161 (holding
that the FDA could not regulate tobacco because it was “plain
that Congress ha[d] not given the FDA the authority that it
s[ought] to exercise”).

In the ACE Rule, the EPA stated that, while its
interpretation of Section 7411 did not depend on the “major
question[s] doctrine[,]” the Agency believed that “that doctrine
should apply here[.]” 84 Fed. Reg. at 32,529. The Agency
reasoned that the Clean Power Plan would have had “billions
of dollars of impact on regulated parties and the economy,”
would have “affected every electricity customer[,]” was
“subject to litigation involving almost every State,” and would
have upset the balance of regulatory authority between federal
agencies and the States. Id. For those reasons, the Agency
concluded that the “interpretive question raised”—whether the
“best system of emission reduction” can include measures
other than improvements to and at the physical source—“must
be supported by a clear[]statement from Congress.” Id. That
was incorrect.

1. The EPA’s Regulatory Mandate

Unlike cases that have triggered the major questions
doctrine, each critical element of the Agency’s regulatory
 76
authority on this very subject has long been recognized by
Congress and judicial precedent.

Most importantly, there is no question that the regulation
of greenhouse gas emissions by power plants across the Nation
falls squarely within the EPA’s wheelhouse. The Supreme
Court has ruled specifically that greenhouse gases are “air
pollutants” covered by the Clean Air Act. Massachusetts v.
EPA, 549 U.S. at 532. More to the point, the Court has told the
EPA directly that it is the Agency’s job to regulate power
plants’ emissions of greenhouse gases under Section 7411.
“Congress delegated to EPA the decision whether and how to
regulate carbon-dioxide emissions from powerplants” through
a “§ 7411 rulemaking[.]” AEP, 564 U.S. at 426–427. The
separate opinion agrees. See Separate Op. at 14 (“Does the
Clean Air Act direct the EPA to make our air cleaner? Clearly
yes. Does it require at least some carbon reduction? According
to Massachusetts v. EPA, again yes.”).

On top of that, the issuance of regulations addressing
greenhouse gas pollution is mandatory under the statute
because of longstanding endangerment findings. In
Massachusetts v. EPA, the Supreme Court directed the EPA
either to make an endangerment finding under the statute for
greenhouse gas pollution, or to explain why it would not do so.
549 U.S. at 532–535. The EPA complied. For now more than
a decade—from 2009 to the present day in the ACE Rule
itself—the EPA has consistently and repeatedly recognized the
serious danger that greenhouse gas pollution poses to human
health and welfare. See ACE Rule, 84 Fed. Reg. at 32,533;
New Source Rule, 80 Fed. Reg. at 64,530–64,531; 2009
Endangerment Finding, 74 Fed. Reg. at 66,496–66,497. By
statute, that finding triggers a mandatory duty on the EPA to
regulate greenhouse gas pollution. 42 U.S.C. § 7521(a)(1)
(motor vehicle emissions); 42 U.S.C. § 7411(b) (stationary
 77
sources that contribute significantly to such dangerous
pollution).8

So the EPA has not just the authority, but a statutory duty,
to regulate greenhouse gas pollution, including specifically
from power plants.

In that way, the pollution measures in the Clean Power
Plan do not fit the major-question mold of prior cases. For
example, in Brown & Williamson, the major question was
whether the agency had authority to regulate tobacco at all.
There, the Supreme Court ruled that there was “reason to
hesitate” before concluding that the provisions of the Food,
Drug, and Cosmetic Act covering restricted devices, Brown &
Williamson, 529 U.S. at 134 (citing 21 U.S.C. § 360j(e)), gave
the Food and Drug Administration the authority to regulate

8
As discussed below with respect to the challenge brought by
the Coal Petitioners (infra at III.A.1), the legal basis for the EPA’s
regulation of greenhouse gas emissions from existing power plants
in both the Clean Power Plan and the ACE Rule was the Agency’s
prior 2015 decision to issue standards of performance for carbon
dioxide emitted from new power plants. That decision, in turn, was
based on the Agency’s recognition (since the 1970s) that fossil-fuel-
fired power plants contribute significantly to air pollution, which
“may reasonably be anticipated to endanger the public health or
welfare.” 42 U.S.C. § 7411(b)(1)(A); see Air Pollution Prevention
and Control: List of Categories of Stationary Sources, 36 Fed. Reg.
5931, 5931 (March 31, 1971); Air Pollution Prevention and Control:
Addition to the List of Categories of Stationary Sources, 42 Fed. Reg.
53,657, 53,657 (Oct. 3, 1977). The EPA also determined in 2015
that power plants contribute significantly to greenhouse gas pollution
in particular. See New Source Rule, 80 Fed. Reg. at 64,531. That
determination, combined with the determination that greenhouse
gases are dangerous to public health and welfare, triggers a
mandatory duty to regulate under Section 7411(b)(1)(A).
 78
tobacco given its “unique political history” and its role as a
“significant portion of the American economy.” Id. at 159.
The Court reasoned based on the overall drug-regulatory
scheme, as well as Congress having “created a distinct
regulatory scheme for tobacco products,” that Congress “could
not have intended to delegate a decision of such economic and
political significance to an agency in so cryptic a fashion.” Id.
at 159–160.

That question of agency authority to regulate the matter in
question was absent for the Clean Power Plan. In fact, the
Supreme Court in Massachusetts v. EPA rejected the analogy
between regulation of greenhouse gases as a pollutant under the
Clean Air Act and regulation of tobacco as a drug under the
Food, Drug, and Cosmetic Act. 549 U.S. at 530–531. Treating
tobacco as a drug would have been wholly novel, requiring the
agency to ban virtually all tobacco products—a result the Court
suspected Congress did not intend. Id. at 531; Brown &
Williamson, 529 U.S. at 143. By contrast, the Supreme Court
explained, greenhouse gases are air pollutants that fall squarely
within the Clean Air Act’s coverage, and the Act would subject
such pollutants, if the agency makes the necessary findings,
only to regulation, not prohibition. Massachusetts v. EPA, 549
U.S. at 531.

The Clean Air Act also contains its own limits on
regulation, like mandating that the EPA take into account such
factors as available technology and the cost of compliance. Id.
(citing 42 U.S.C. § 7521(a)(2)); see also 42 U.S.C.
§ 7411(a)(1) (requiring consideration of health and
environmental impacts, energy requirements, and cost). In that
way, Congress designed the Clean Air Act’s processes for
regulating air pollution to adapt to “changing circumstances
and scientific developments” without imposing unreasonable
technological or financial burdens on industry. Massachusetts
 79
v. EPA, 549 U.S. at 532. So, unlike the major question of
tobacco regulation in Brown & Williamson, there is “nothing
counterintuitive” about the EPA’s reasonable regulation of
dangerous airborne substances like greenhouse gases. Id. at
531–532.

Similarly, the major question in UARG was whom the
EPA was attempting to regulate. In that case, the Supreme
Court held that the EPA’s statutory permitting authority for the
construction and modification of stationary sources was
“designed to apply to, and cannot rationally be extended
beyond, a relative handful of large sources capable of
shouldering heavy substantive and procedural burdens”—
sources like power plants. 573 U.S. at 322. The Court held
that, without clear statutory grounding, the EPA’s effort to
extend permitting requirements to literally millions of small
sources of greenhouse gas pollution but of no other regulated
pollutants—sources like schools, hospitals, churches, and
shopping malls—overshot its statutory authority. Id. at 324,
328.

The Clean Power Plan, by contrast, regulated the very
entities the EPA was told by the Supreme Court in AEP and
UARG to regulate—fossil-fuel-fired power plants. And it
employed statutory tools that were “suitable” for application to
the long-regulated power industry. See UARG, 573 U.S. at
323, 324 n.7. American Electric Power pointed the Agency to
regulation under Section 7411 specifically, explaining that
“Congress delegated to EPA the decision whether and how to
regulate carbon-dioxide emissions from [new, modified, and
existing] powerplants” using the regulatory tools laid out in
Section 7411. 564 U.S. at 424–426.

That is no doubt a significant task for the EPA. But that is
not because of any agency overreach. It is the product of
 80
Congress’ charge that the EPA regulate air pollution
nationwide. And with respect to regulating greenhouse gas
pollution in particular, it reflects the fact that fossil-fuel-fired
power plants predominate the power industry and are spread
across the Nation. See United States Energy Information
Administration (EIA), Frequently Asked Questions: What is
U.S. Electricity Generation by Source? (Nov. 2, 2020),
https://www.eia.gov/tools/faqs/faq.php?id=427&t=2 (last
visited Jan. 11, 2021); EIA, U.S. Energy Mapping System,
https://www.eia.gov/state/maps.php (last visited Jan. 11,
2021). So much so that they “are by far” the greatest stationary
contributor to greenhouse gas pollution and the significant
dangers it causes for the public health and welfare. New
Source Rule, 80 Fed. Reg. at 64,522.

2. Best System of Emission Reduction

So what the EPA may regulate (greenhouse gas pollution),
and whom it may target (power plants), and how (under Section
7411) have all been resolved and so do not trigger the major
questions doctrine.

That leaves the EPA no place to house its major-question
objection other than in the interpretation of the statutory term
“best system of emission reduction,” 42 U.S.C. § 7411(a)(1).
More specifically, the EPA says the use of any emission-
control measures that do not operate at the individual physical
plant level requires an express statement from Congress, and
that federal standards that might encourage generation-shifting
are therefore categorically forbidden under Section 7411.

But the major questions doctrine does not apply there
either for a number of reasons.
 81
a. Statutory Design

For starters, the “best system of emission reduction” plays
a cabined role in the statutory scheme. The determination of
the best system of emission reduction is entirely internal to the
EPA. The EPA itself evaluates relevant scientific,
technological, and economic evidence to identify, in its
judgment, the “best system of emission reduction” available,
and the “degree of emission limitation achievable” through it.
42 U.S.C. § 7411(a)(1).

In making that determination, the statute significantly
reins in the EPA’s judgment by requiring the Agency to
(1) “tak[e] into account the cost of achieving such reduction,”
(2) factor in “any nonair quality health and environmental
impact,” (3) balance the effect on “energy requirements,” and
(4) ensure that the system has been “adequately
demonstrated[.]” 42 U.S.C. § 7411(a)(1). To be “adequately
demonstrated[,]” we have explained, the system must be shown
to be reasonably “reliable,” “efficient,” and “expected to serve
the interests of pollution control without becoming exorbitantly
costly[.]” Essex Chem. Corp. v. Ruckelshaus, 486 F.2d 427,
433 (D.C. Cir. 1973), cert. denied, 416 U.S. 969 (1974); see
also Portland Cement Ass’n v. Ruckelshaus, 486 F.2d 375, 391
(D.C. Cir. 1973) (whether a system is adequately demonstrated
“cannot be based on ‘crystal ball’ inquiry”).9

9
In addition to these statutory constraints, the EPA has tied its
own hands by requiring that the best system include only actions
touching three bases: (i) they reduce emissions (rather than, for
example, capturing emissions after they are released into the air by
planting trees), (ii) sources themselves can implement them, and
(iii) they target supply-side activities. See Clean Power Plan, 80 Fed.
Reg. at 64,776, 64,778–64,779.
 82
Once the EPA identifies a best system that meets those
requirements and calculates the degree of emission limitation
it allows, the Clean Air Act leaves it to the States to set their
own standards of performance for their existing pollution
sources. 42 U.S.C. § 7411(d). The cooperative-federalism
design of Section 7411(d) gives the States broad discretion in
achieving those emission limitations. See AEP, 564 U.S. at 428
(“The Act envisions extensive cooperation between federal and
state authorities, generally permitting each State to take the first
cut at determining how best to achieve EPA emissions
standards within its domain[.]”) (internal citations omitted). In
addition, Section 7411(d) expressly allows States, in setting
their emission standards, to “take into consideration, among
other factors, the remaining useful life” of its existing sources.
42 U.S.C. § 7411(d).

So the EPA’s scientific and technological identification of
the best system of emission reduction cannot bear the major-
question label. Determining the system is a task expressly and
indisputably assigned by Congress to the EPA and requiring
specialized agency expertise. That system serves only as the
basis for the EPA to set the emission-reduction targets in its
quantitative guidelines. The States retain the choice of how to
meet those guidelines through standards of performance
tailored to their various sources. Neither exercise entails
resolution of a major question.

The EPA argues that its own best-system process raised a
major question by “impos[ing] ‘generation shifting[.]’” EPA
Br. 99. But under Section 7411(d), the EPA does not impose
the “best system of emission reduction” on anyone. Instead,
each State decides for itself what measures to employ to meet
the emission limits, and in so doing may elect to consider the
“remaining useful life” of its plants and “other factors.” 42
U.S.C. § 7411(d). See Clean Power Plan, 80 Fed. Reg. at
 83
64,709–64,710, 64,783. The Clean Power Plan, in fact,
afforded States considerable flexibility in choosing how to
calculate and meet their emissions targets. See, e.g., id. at
64,665, 64,756–64,757, 64,834–64,837.10

Congress already focused on the issue and made the
decision to rope the EPA’s selection of a best system of
emission reduction about with all of those substantive and
structural limitations. So the major questions doctrine does not
provide any basis for concluding that the Clean Air Act
categorically forecloses the EPA’s consideration of even those
generation-shifting measures that are already widely in use by
States and power plants and have been demonstrated to be
reasonable, reliable, effective, and not unduly disruptive to the
regulated industry. See Clean Power Plan, 80 Fed. Reg. at
64,735, 64,769.

10
The Clean Power Plan expressly contemplated that States and
sources might choose to meet their emissions targets by using
measures other than the specific heat-rate improvements and
generation shifting that the EPA had identified in its best system. See
80 Fed. Reg. at 64,755–64,758. The EPA offered a list of alternative
available technologies that reduced power plants’ carbon dioxide
emissions per megawatt, including carbon capture and storage, heat-
rate improvements at non-coal plants, fuel switching to gas, fuel
switching to biomass, and waste heat-to-energy conversion. Id. at
64,756. In certain situations, for example, modifying coal-fired
plants to burn natural gas could “help achieve emission limits
consistent with the [best system].” Id. The Agency also identified a
list of alternative measures that States could implement to lower
overall emissions from fossil-fuel-fired plants. Those measures
included, for example, demand-side energy efficiency—a policy tool
that the EPA expected some States to use because “the potential
emission reductions from demand-side [energy efficiency] rival
those from [generation shifting] in magnitude[.]” Id.
 84
In that respect, the EPA’s argument sounds much like a
second argument rejected by the Supreme Court in UARG. In
addition to the scope question discussed above, the Court
addressed whether the EPA could require facilities that emit
conventional pollutants also to implement the “best available
control technology” for greenhouse gases. UARG, 573 U.S. at
329–333 (citing 42 U.S.C. § 7475(a)(4)). Like the EPA here,
the industry petitioners argued that the “best available control
technology” standard was “fundamentally unsuited” to
greenhouse gas emissions because it had “traditionally”
focused on “end-of-stack controls.” Id. at 329–330.
“[A]pplying it to greenhouse gases,” the industry petitioners
insisted, would make the “best available control technology”
standard “more about regulating energy use, which will enable
regulators to control every aspect of a facility’s operation and
design[.]” Id. at 330 (internal quotation marks omitted).

The Supreme Court rejected that challenge. The Court
explained that the EPA’s guidance contemplated both “end-of-
stack”–type controls and energy efficiency measures. UARG,
573 U.S. at 330. And, critically, the Court emphasized that the
statute and regulations already imposed “important limitations
on [best available control technology] that may work to
mitigate petitioners’ concerns about ‘unbounded’ regulatory
authority.” Id. at 331. Among those limitations was the EPA’s
longstanding statutory interpretation that the best available
control technology was required “only for pollutants that the
source itself emits,” and the EPA’s existing guidance that
permitting authorities should “consider whether a proposed
regulatory burden outweighs any reduction in emissions to be
achieved.” Id. The statute also required the EPA to determine
the best available control technology with reference to “energy,
environmental, and economic impacts and other costs.” 42
U.S.C. § 7479(3); see also UARG, 573 U.S. at 333 n.9.
 85
So too here: The numerous substantial and explicit
constraints on the EPA’s selection of a best system of emission
reduction foreclose using the major questions doctrine to write
additional, extratextual, and inflexibly categorical limitations
into a statute whose “broad language * * * reflects an
intentional effort to confer the flexibility necessary to
forestall * * * obsolescence.” Massachusetts v. EPA, 549 U.S.
at 532; see also Transmission Access Policy Study Group v.
FERC, 225 F.3d 667, 711 (D.C. Cir. 2000) (where Congress
has spoken, court upholds as within agency authority an order
that “fundamentally change[d] the regulatory environment in
which utilities operate” and “introduc[ed] meaningful
competition into an industry that since its inception has been
highly regulated and affecting all utilities in a similar way”),
aff’d sub nom. New York v. FERC, 535 U.S. 1 (2002).

The EPA points to the Supreme Court’s statement in
UARG that “[w]hen an agency claims to discover in a long-
extant statute an unheralded power to regulate ‘a significant
portion of the American economy,’ we typically greet its
announcement with a measure of skepticism.” 573 U.S. at 324
(quoting Brown & Williamson, 529 U.S. at 159).

True. But, as already explained, the EPA made no new
discovery of regulatory power with the Clean Power Plan.
While power plants are significant players in the American
economy, they have been subject to regulation under Section
7411 for nearly half a century. See, e.g., Costle, 657 F.2d at
318; Oljato Chapter of Navajo Tribe v. Train, 515 F.2d 654,
656–57 (D.C. Cir. 1975). Their emission of massive amounts
of carbon dioxide has long been known. And the source of the
EPA’s duty to regulate that greenhouse gas pollution from
power plants was the plain statutory text and Supreme Court
precedent, not something the EPA pulled out of a hat. See AEP,
564 U.S. at 425; Massachusetts v. EPA, 549 U.S. at 532.
 86
In sum, the Clean Air Act expressly confers regulatory
authority on the EPA to set standards for reducing greenhouse
gas emissions from fossil-fuel-fired power plants nationwide.
Congress knew both the scope and importance of what it was
doing. And it cabined the EPA’s authority with concrete and
judicially enforceable statutory limitations. The major
questions doctrine is meant to discern, not override, such
statutory judgments. Doubly so when the regulatory authority
and its reach have been affirmed and enforced by the Supreme
Court.

b. Regulatory Consequences

The problems with the EPA’s approach to the major-
question analysis do not stop there. The Agency also conflates
the significance of greenhouse gas regulation of power plants
generally with any significance attributable solely to the EPA’s
choice of a “best system of emission reduction”—the statutory
provision where the EPA tried to anchor its major-question
objection. Remember, the EPA concluded that the major
questions doctrine was triggered centrally by (i) the Clean
Power Plan’s “billions of dollars of impact” on the economy;
(ii) its effect on “every electricity customer”; (iii) the number
of litigation challenges it spawned, “involving almost every
State”; and (iv) its perceived shifting of regulatory authority
between federal agencies and the States. ACE Rule, 84 Fed.
Reg. at 32,529.

Taking the characterizations as true, those consequences
are a product of the greenhouse gas problem, not of the best-
system’s role in the solution. Given the number and dispersion
of fossil-fuel-fired power plants, any nationwide regulation of
their greenhouse gas pollution that meaningfully addresses
emissions will necessarily affect a broad swath of the Nation’s
electricity customers. Under the EPA’s grave endangerment
 87
finding, so too would a failure to regulate those greenhouse gas
emissions. See 2009 Endangerment Finding, 74 Fed. Reg. at
66,496.

As for the “billions of dollars of impact[,]” the EPA has
offered no evidence tying that cost to generation shifting rather
than physical plant adjustments or a variety of other means
States might choose for complying with emission limits. As
the EPA itself previously acknowledged, generation shifting
can be cheaper than other demonstrated methods of reducing
greenhouse gas emissions, like carbon capture and storage, that
take place “at” the source (and thus fall within the EPA’s
current statutory vision). See Clean Power Plan, 80 Fed. Reg.
at 64,727.11 Moreover, the Clean Power Plan’s significant
projected economic impact was not atypical for Clean Air Act
rulemakings by the EPA. See, e.g., Costle, 657 F.2d at 314
(upholding 1979 new source performance standards governing
emission control by coal-burning power plants that imposed
“tens of billions of dollars” of costs on the power sector).

Even assuming that the EPA’s federalism concerns could
trigger the major questions doctrine (rather than the federalism
clear-statement canon), they carry no material weight here.
That is because the statutory role of the best system of emission
reduction under Section 7411(d) textually preserves and
enforces the States’ independent role in choosing from among
the broadest range of options to set standards of performance
appropriate to sources within their jurisdiction. In fact, it is the

11
The EPA now takes the position that natural gas co-firing is
not adequately demonstrated and that neither co-firing nor carbon
capture and storage is part of the best system of emission reduction.
See ACE Rule, 84 Fed. Reg. at 32,544–32,545, 32,549. But those
methods are amenable to implementation “to” and “at” the source, in
keeping with the EPA’s statutory view.
 88
ACE Rule’s unreasoned barriers to certain compliance
measures, like generation shifting and biomass co-firing, that
hamstring the States. See supra Part II.A.3 (analyzing ACE
Rule, 84 Fed. Reg. at 32,555–32,556).

Finally, it seems doubtful that the volume of litigation
aimed at a regulation can reasonably bear on its major-question
status. The Supreme Court has certainly never embraced that
idea. For good reason. A doctrine at the mercy of litigation
stratagems, or the mere existence of disagreements over which
parties find advantage in filing suit, cannot be an elucidating or
even logically relevant tool of statutory interpretation.

In any event, the EPA offers no basis for concluding that
the best-system determination is what lit the litigation fire.
After all, the ACE Rule too has been “subject to litigation”
involving 43 States and all manner of other interested parties,
despite the Rule’s jettisoning of generation shifting as part of
the best system of emission reduction. See Opinion Caption,
supra.

c. Regulating in the Electricity Sector

The ACE Rule’s last attempt to wrap the best-system
determination in the major-question mantle asserts that
including generation shifting as part of the best system of
emission reduction lacks a “valid limiting principle,” and that,
by “shifting focus to the entire grid[,]” it would “empower” the
Agency “to order the wholesale restructuring of any industrial
sector[.]” ACE Rule, 84 Fed. Reg. at 32,529. But that is
entirely wrong. The Clean Power Plan was aimed not at
regulating the grid, but squarely and solely at controlling air
pollution—a task at the heart of the EPA’s mandate. Indeed,
the EPA’s reasoning in the ACE Rule defeats its own
argument.
 89
The EPA suggests that counting generation shifting among
the tools for emission reduction risks expanding the Agency’s
regulatory sights too far, because “any action affecting a
generator’s operating costs could impact its order of dispatch
and lead to generation shifting.” ACE Rule, 84 Fed. Reg. at
32,529 (emphasis added). That is exactly right: Any regulation
of power plants—even the most conventional, at-the-source
controls—may cause a relative increase in the cost of doing
business for particular plants but not others, with some
generation-shifting effect. That is how pollution regulation in
the electricity sector has always worked. Regulators—
including, for example, Congress in the Clean Air Act’s acid
rain cap-and-trade program, 42 U.S.C. §§ 7651–7651o—have
long facilitated those generation-shifting effects to serve the
goal of pollution reduction. See Grid Experts Amicus Br. 13–
15.

So the EPA’s contention that it cannot consider measures
resulting in generation shifting as part of its best system proves
far too much: If that were so, the EPA would be limited to
considering only measures that power plants could adopt at
zero cost, so as to maintain their relative-dispatch position.
That is, of course, incompatible with Congress’ instruction that
the best system take cost into account as only one factor among
several, see 42 U.S.C. § 7411(a)(1), and contrary to the very
nature of environmental law, which requires the regulation of
polluters and material changes in their pollution emissions.

The EPA’s argument also ignores, again, the critical
statutory limitations that the Clean Air Act imposes on the
selection of a best system of emission reduction and its function
in state plans. Under Section 7411(d), the EPA lacks the
authority to “order the wholesale restructuring” of anything.
All it can do is identify the best system of emission reduction
that has been adequately demonstrated within the cost, energy-
 90
requirement, and other substantive constraints set by Congress,
and then calculate achievable emission goals by reference to
that system. 42 U.S.C. § 7411(a)(1). States, in turn, set
standards of performance only “for” any “existing source[,]”
and need not implement any aspect of the EPA’s “best
system[.]” Id. § 7411(d)(1) (emphasis added). And the EPA’s
determination about how best to combat air pollution is, of
course, subject to judicial review, including on questions like
whether a system has been adequately demonstrated and
whether the Agency adequately considered costs. 42 U.S.C.
§ 7607(b); cf. AEP, 564 U.S. at 427; UARG, 573 U.S. at 333
n.9. Congress’ carefully calibrated system—involving
scientific and technological evidence-gathering, close study of
existing industry practice, constrained discretion, divided
regulatory authority, collaboration with States, and judicial
review—leaves no room for the unauthorized agency
overreach that the EPA fears.

A group of States and industry groups intervened with
other major-question challenges, but their salvos all fall short.
They argue that the major questions doctrine is implicated
because the EPA has “‘no expertise’ in electricity generation,
transmission, and reliability.” State & Industry Intervenors
Repeal Br. 30 (quoting King, 576 U.S. at 474); see also
Gonzales, 546 U.S. at 267 (rejecting interpretive rule of the
Attorney General that was “both beyond his expertise and
incongruous with the statutory purposes and design”). But
Section 7411 not only foresees, but demands that the EPA
consider “energy requirements” when assessing the best
system of emission reduction. 42 U.S.C. § 7411(a)(1). The
Supreme Court in AEP recognized the EPA’s signal role in
regulating greenhouse gases under Section 7411
notwithstanding that the EPA must consider energy
requirements and ensure a reliable energy supply when it does
so. 564 U.S. at 427. The Court explained that, when the EPA
 91
is formulating greenhouse gas regulations, it must consider not
only “the environmental benefit potentially achievable,” but
also “our Nation’s energy needs and the possibility of
economic disruption[.]” Id. The Clean Air Act “entrusts such
complex balancing to EPA in the first instance, in combination
with state regulators.” Id.; see 42 U.S.C. § 7411(a)(1). That
definitive reading of the statute by the Supreme Court cannot
suddenly become a forbidden major question when the EPA
regulates as it was told to do.

The statutory scheme simply gives no quarter to the
proposition that, in following Congress’ directive to regulate
electricity-producing power plants, the EPA is categorically
forbidden to consider emission-reduction measures that take
into account the nature of the electricity grid in which those
power plants operate day in and day out. Nor is it sensible to
categorically put off-limits the generation-shifting measures
that power plants are already actually using to meet emission
requirements. See Clean Power Plan, 80 Fed. Reg. at 64,784–
64,785.

The State and Industry Intervenors also overlook that the
EPA developed the Clean Power Plan with input from other
agencies with relevant expertise. See Clean Power Plan, 80
Fed. Reg. at 64,672–64,673 (explaining that “[i]nput and
assistance from FERC [the Federal Energy Regulatory
Commission] and DOE [the Department of Energy] have been
particularly important in shaping” aspects of the Clean Power
Plan); id. at 64,671 (noting “extensive consultation with key
agencies responsible for [electric system] reliability[,]” as well
as reliance on the “EPA’s longstanding principles in setting
emission standards for the utility power sector”). Contrast
Delaware Dep’t of Nat. Res. & Env’t Control v. EPA, 785 F.3d
1, 14, 18 (D.C. Cir. 2015) (invalidating rule in part because the
EPA had failed to consult with other expert agencies on grid
 92
reliability issues). EPA could hardly do its job without
substantively engaging with the on-the-ground facts about the
electricity system that power plants support. Quite the
opposite: An agency’s wooden refusal to factor in reality and
such on-point considerations would ordinarily render its
decisionmaking arbitrary and capricious. See Motor Vehicle
Mfrs. Ass’n v. State Farm Mut. Auto Ins. Co., 463 U.S. 29, 43
(1983).

All told, the EPA’s consideration of already-in-use
generation shifting as part of the “best system of emission
reduction” does nothing to enlarge the Agency’s regulatory
domain. “We are not talking about extending EPA jurisdiction
over millions of previously unregulated entities,” but about a
familiar process of cooperative federalism applied to “entities
already subject to * * * regulation” to address a recognized
form of air pollution that repeatedly has been found to endanger
public health and welfare. See UARG, 573 U.S. at 332. The
major questions doctrine cannot rescue the ACE Rule’s
mistaken interpretation of Section 7411(d) as categorically
confining the best system of emission reduction to physical
adjustments made only “at” and “to” the power plant.

C. FEDERALISM

The federalism canon lends no support to the ACE Rule’s
decision to confine the best system of emission reduction to
measures that apply exclusively at and to the source. That
canon recognizes that “the States retain substantial sovereign
powers under our constitutional scheme, powers with which
Congress does not readily interfere.” Gregory v. Ashcroft, 501
U.S. 452, 460–461 (1991). So as a matter of constitutional
avoidance, courts require Congress to “enact exceedingly clear
language if it wishes to significantly alter the balance between
federal and state power.” United States Forest Serv. v.
 93
Cowpasture River Pres. Ass’n, 140 S. Ct. 1837, 1849–1850
(2020).

The federalism clear-statement rule prevents direct federal
intrusion into areas of traditional state responsibility unless
Congress has made its intent to cross that line explicit. For
example, courts will not assume that Congress meant to
preempt a State’s mandatory retirement age for state judges
through the passage of a generic age discrimination law, unless
it has clearly expressed its intent to police the qualifications of
such high-level state officials. See Gregory, 501 U.S. at 463–
464. Nor will courts lightly assume that Congress intended to
claim state-owned land as part of the National Park System, see
Cowpasture River, 140 S. Ct. at 1849–1850, to transform
simple state-law assaults into breaches of international
chemical weapons compacts, see Bond v. United States, 572
U.S. 844, 862–863 (2014), or to displace the States’ traditional
authority to regulate the practice of law, see American Bar
Ass’n v. FTC, 430 F.3d 457, 466, 471–472 (D.C. Cir. 2005).
Only when such conflicts between federal and state regulatory
domains are plainly joined by Congress itself will the court
confront the sensitive constitutional implications of such
measures.

That doctrine does not support the EPA’s cramped reading
of Section 7411. Interstate air pollution is not an area of
traditional state regulation. And federalism concerns do not bar
the United States government from addressing areas of federal
concern just because its actions have incidental effects on areas
of state power. Cf. FERC v. Electric Power Supply Ass’n, 136
S. Ct. 760, 775–778 (2016) (federal regulation of wholesale
electricity market did not intrude on traditional state authority
over the retail electricity market, even though wholesale
market regulation created an incentive for retail consumers to
change their behavior in state-regulated markets).
 94
What is more, the Supreme Court has suggested that the
federalism clear-statement rule is of limited applicability when
a federal regulatory regime is enforced through a statutory
cooperative-federalism framework, as Section 7411(d) is. See
AT&T Corp. v. Iowa Util. Board, 525 U.S. 366, 378 n.6 (1999)
(noting appeals to States’ rights as “most peculiar” in the
context of “a federal program administered by 50 independent
state agencies”); see also Alaska Dep’t of Env’t Conservation
v. EPA, 540 U.S. 461 (2004) (declining to adopt dissent’s
proposed clear-statement rule for federal constraints on state
implementation decisions in cooperative-federalism program).
See generally Abbe Gluck, Intrastatutory Federalism and
Statutory Interpretation: State Implementation of Federal Law
in Health Reform and Beyond, 121 YALE L.J. 534, 555–556
(2011).

In any case, the Clean Power Plan’s incorporation of
generation shifting into its best system of emission reduction
fell squarely within an area of the federal government’s
constitutional competence. The EPA does not dispute the
government’s authority or its statutory mandate to reduce the
emission of pollutants that endanger public health and welfare.
42 U.S.C. § 7411(b)(1)(A), (d)(1). The EPA also agrees that
greenhouse gases are among the pollutants properly regulated
by the federal government. See AEP, 564 U.S. at 416–417,
424; see also supra Part I.B.2.

The Clean Power Plan directly regulated only the amount
of greenhouse gas pollutants that may be emitted into the
atmosphere. 80 Fed. Reg. at 64,663–64,664. That is an area
of unique federal concern. After all, “[a]ir pollution is
transient, heedless of state boundaries,” EME Homer City
Generation, 572 U.S. at 496, particularly where the pollutants
are greenhouse gases, which have little if any localized effect
but great cumulative impact. The inability of individual States
 95
to redress the problem of interstate air pollution, in fact, was
among the very reasons for the enactment of the Clean Air Act.
See 42 U.S.C. § 7401(a)(1), (4); S. REP. NO. 88-638, at 3 (1963)
(“Polluted air is not contained in a specific area but is carried
from one political jurisdiction to another. It does not know
State lines or city limits. Providing air of good quality * * * is
a challenge and an obligation for Government operations on all
levels.”); id. at 5 (“The nationwide character of the air
pollution problem requires an adequate Federal program to
lend assistance, support, and stimulus to State and community
programs.”).

To be sure, the federal government’s regulation of such an
interstate problem can have indirect effects on State energy
production and utility regulation decisions. But even when
those effects are the fully anticipated “natural consequences”
of an agency’s policy choice, Electric Power Supply Ass’n, 136
S. Ct. at 776, that does not transform a fundamentally federal
action in a core federal area of concern into a restriction on state
action that triggers the federalism canon.12

The EPA protests that the Clean Power Plan breached that
divide because it expressly considered generation shifting to

12
In the ACE Rule, the EPA suggested that the Clean Power
Plan’s best system of emission reduction was also impermissible as
an encroachment on “measures and subjects exclusively left to
FERC[.]” 84 Fed. Reg at 32,530. The EPA has not pressed that
argument here. For good reason. The effects of environmental
regulations on the power grid do not amount to power regulation
statutorily reserved to FERC. And, in any event, the constitutional
concerns that require us to patrol the boundaries between federal and
state authority with vigilance do not support any similar clear-
statement requirement regarding turf battles between federal
agencies.
 96
determine the best system of emission reduction and, in so
doing, stepped on the States’ power to regulate electrical
utilities’ mix of electricity generation. Reg.

But that argument has nothing to do with the narrow
construction of Section 7411 that the EPA adopted. After all,
the EPA could have set the same emission guidelines
predicated on a best system of emission reduction that
exclusively employed technological controls applicable at and
to the source, like carbon capture and sequestration. And the
EPA must agree that the federalism canon would play no role
in determining the appropriateness of that system, since on the
Agency’s own reading, measures applicable at and to the
source are precisely what Section 7411 allows.13

Nowhere does the EPA explain why reference to a
different mechanism—generation shifting—in its calculation
of the best system would raise materially different federalism

13
While the EPA did not select carbon capture and sequestration
as the best system of emission reduction in the ACE Rule, it excluded
that process because of cost and feasibility concerns, not federalism
interests. See 84 Fed. Reg. at 32,547–32,549. That exclusion was a
change of position from the Clean Power Plan, where the EPA found
that the process was “technically feasible and within price ranges that
the EPA has found to be cost effective[.]” 80 Fed. Reg. at 64,727.
Carbon capture and sequestration ultimately was not selected as the
best system of emission reduction in the Clean Power Plan solely
because generation shifting was even more cost-effective. Id. at
64,727–64,728. What matters here is that the EPA did not express
any concern in either the ACE Rule or the Clean Power Plan that
such a system would intrude upon traditional areas of State authority.
In the ACE Rule, the EPA permits the use of such technological
controls to meet its emission standards, 84 Fed. Reg. at 32,549,
32,555, as it did in the Clean Power Plan, 80 Fed. Reg. at 64,883–
64,884.
 97
concerns. Under either system, the only direct obligation
imposed on States is the same: a federally set emissions
guideline. In both scenarios, the States remain equally free to
choose the compliance measures that best fit the needs of their
State and industry. And as a practical matter, many if not most
States would likely opt for generation shifting over carbon
capture and sequestration under either rule because the former
is cheaper for existing plants. See Clean Power Plan, 80 Fed.
Reg. at 64,727–64,728; ACE Rule, 84 Fed. Reg. at 32,532
(“Market-based forces have already led to significant
generation shifting in the power sector.”).

The EPA also suggests that the clear-statement rule
operates with particular force here because the Plan imposed
uneven regulatory burdens weighted toward States with more
high-emitting power plants. But that argument tries to twist
principles of federalism into a command of regulatory
homogenization that defies on-the-ground reality. Regulations
under the Clean Air Act or any environmental law will
commonly affect States differently depending on the States’
activities. The regulation of pollutants associated with
automotive manufacturing affects States with production
facilities more than those without. See, e.g., General Motors
Corp. v. United States, 496 U.S. 530, 534–535 (1990). The
regulation of mining-related pollutants imposes greater costs
on States with more plentiful mineral resources. See, e.g.,
Alaska Dep’t of Env’t Conservation, 540 U.S. at 469–470, 474;
Hodel v. Virginia Surface Mining & Reclamation Ass’n, 452
U.S. 264, 289–290 (1981). The same point applies to industries
like petroleum refining, which are concentrated near navigable
waters. See generally EIA, U.S. Energy Mapping System
https://www.eia.gov/state/maps.php (last visited Jan. 11,
2021). Indeed, some regulations impose additional regulatory
burdens based literally on the direction the wind blows. See
EME Homer City Generation, 572 U.S. at 520. Likewise,
 98
States with more navigable water necessarily carry more
burdens under the Clean Water Act than those with less.

Affected States, of course, could raise statutory challenges
to enforce the Clean Air Act’s express constraints, such as
required consideration of cost, non-air quality health and
environmental impact, or energy requirements under Section
7411(a). And they could always challenge any unreasoned or
unwarranted distinctions in regulatory coverage as arbitrary or
capricious. But in the absence of such an objection, it does not
offend—or even implicate—principles of federalism to
observe that States whose industries pollute the Nation’s air
and so harm the public’s health more will, in turn, be affected
more by emission controls.

For all of those reasons, nothing in the federalism canon
supports the EPA’s effort to categorically constrict the best
system of emission reduction to measures physically applied at
and to the individual plant.

III. THE EPA’ S AUTHORITY TO REGULATE
CARBON DIOXIDE EMISSIONS UNDER SECTION 7411

A. THE COAL PETITIONERS’ CHALLENGES

The North American Coal Corporation and Westmoreland
Mining Holdings LLC, both coal mine operators (the Coal
Petitioners), bring two challenges to the ACE Rule. Both
question the EPA’s legal authority to enact the rule. First, the
Coal Petitioners argue that the EPA failed to make the required
endangerment finding––that carbon dioxide emissions from
power plants cause or contribute significantly to air pollution
that may reasonably be anticipated to endanger public health or
welfare––before regulating those emissions. See 42 U.S.C.
§ 7411(b)(1)(A). Second, they claim that the EPA’s previous
regulation of a different air pollutant (mercury) from power
 99
plants under the Hazardous Air Pollutants provision, 42 U.S.C.
§ 7412, precludes the EPA from now regulating power plants’
emission of greenhouse gases under Section 7411(d).

Both arguments fail. The EPA made the requisite
endangerment finding in 2015, and the ACE Rule expressly
retained that finding. As for the Section 7412 challenge, the
EPA has correctly and consistently read the statute to allow the
regulation both of a source’s emission of hazardous substances
under Section 7412 and of other pollutants emitted by the same
source under Section 7411(d). The Coal Petitioners’ argument
rests not on the enacted statutory language, but instead on their
own favored reading of one statutory amendment inserted by
codifiers. Reading the statutory text as a whole—that is, all of
the relevant language enacted by Congress, including two duly
enacted amendments—the Clean Air Act authorizes the EPA
to regulate both power plants’ emissions of greenhouse gases
under Section 7411(d) and hazardous air pollutants under
Section 7412. That reading is reinforced by the statutory
structure, purpose, and history.

1. Endangerment Finding

a. The Record of Endangerment

The Coal Petitioners argue that the ACE Rule was
unlawful right out of the box because the EPA failed to make a
statutorily required finding that greenhouse gas emissions from
power plants cause air pollution that endangers the public
health and welfare. That is wrong.

As a reminder, before the EPA can regulate a category of
stationary sources like electricity-generating power plants
under Section 7411, the EPA Administrator must first find that
the source category “in his judgment * * * causes, or
contributes, significantly to, air pollution which may
 100
reasonably be anticipated to endanger the public health or
welfare.” 42 U.S.C. § 7411(b)(1)(A). A formal
pronouncement meeting those criteria is known as an
“endangerment finding.” New Source Rule, 80 Fed. Reg. at
64,529. And once it is made, the EPA is not just empowered,
but obligated, to regulate. See 42 U.S.C. § 7411(b)(1)(A); see
also supra note 8.

After the Administrator makes an endangerment finding,
the source category is added to the EPA’s Section 7411 list, 42
U.S.C. § 7411(b)(1)(A), and the Administrator must
promulgate emissions standards (called “standards of
performance”) for new sources in the category, id.
§ 7411(b)(1)(B). As relevant here, unless those dangerous
emissions are regulated under another relevant provision of the
Clean Air Act, the Administrator must also set an achievable
emission guideline based on the “best system of emission
reduction” and provide a process for States to submit a plan
setting out standards of performance for existing stationary
sources in that same category. Id. § 7411(d)(1)(A)(ii).

The EPA has for decades been regulating emissions other
than carbon dioxide from electricity-generating power plants.
In 1971, the EPA listed fossil-fuel-fired electricity-generating
units with steam-generating boilers as a new source category
under Section 7411(b) and promptly established standards of
performance for them. See Air Pollution Prevention and
Control: List of Categories of Stationary Sources, 36 Fed. Reg.
5931 (March 31, 1971); Standards of Performance for New
Stationary Sources, 36 Fed. Reg. 24,876, 24,878–24,880 (Dec.
23, 1971). Then, in 1977, the EPA listed fossil-fuel-fired
combustion turbines as a new source category under
Section 7411 and set performance standards for them. See Air
Pollution Prevention and Control: Addition to the List of
Categories of Stationary Sources, 42 Fed. Reg. 53,657 (Oct. 3,
 101
1977); New Stationary Sources Performance Standards;
Electric Utility Steam Generating Units, 44 Fed. Reg. 33,580
(June 11, 1979). These categories cover the power plants at
issue today. See New Source Rule, 80 Fed. Reg. at 64,531.

Through the 2015 New Source Rule, the EPA began
regulating carbon dioxide emissions from electricity-
generating power plants. See New Source Rule, 80 Fed. Reg.
64,510. Because power plants had already been listed as a
regulated source category, the New Source Rule did not need
to take any action to add those plants to the Section 7411 list of
regulated sources. It just issued, for the first time, standards of
performance for carbon dioxide emitted from new power
plants. In so doing, the New Source Rule provided the
statutory predicate and corresponding duty for the EPA to
establish carbon dioxide emission standards for existing power
plants as well. Clean Power Plan, 80 Fed. Reg. at 64,715; see
42 U.S.C. § 7411(d)(1). The New Source Rule now serves that
same function for the ACE Rule, 84 Fed. Reg. at 32,533.

Because the New Source Rule did not add a new category
of pollution sources to the Section 7411 list, the EPA
concluded that no new endangerment finding was needed.
New Source Rule, 80 Fed. Reg. at 64,529–64,530. The EPA
nevertheless went on to explain that it chose to regulate carbon
dioxide emissions from electricity-generating plants
specifically because greenhouse gas pollution endangers public
health and welfare and contributes significantly to air pollution.
See id. at 64,530–64,531. The EPA found in particular that
increased atmospheric levels of greenhouse gases, including
carbon dioxide, could lead to, among other things, more
frequent extreme weather events and wildfires; threats to
mental and physical health, especially for children and the
elderly; reduced access to food and safe water; and mass
 102
migrations and displacements as a result of rising sea levels.
Id. at 64,517–64,520.

b. Timeliness

At the outset, the EPA argues that we must disregard the
Coal Petitioners’ challenge concerning the endangerment
finding because it was not timely filed. This is a close question,
but we ultimately conclude that the petition is timely.

The Clean Air Act requires that petitions for review
challenging an EPA regulation––including any Section 7411
standard of performance––generally must be filed within 60
days of the regulation’s publication in the Federal Register. 42
U.S.C. § 7607(b)(1). The Clean Air Act’s timeliness bar is
“jurisdictional in nature[.]” Motor & Equip. Mfrs. Ass’n v.
Nichols, 142 F.3d 449, 460 (D.C. Cir. 1998) (quoting Edison
Elec. Inst. v. EPA, 996 F.2d 326, 331 (D.C. Cir. 1993)).

Importantly, Congress carved out an exception to that 60-
day time limit if the petition “is based solely on grounds arising
after [the] sixtieth day[.]” 42 U.S.C. § 7607(b)(1). In that
situation, the clock resets, and the petitioner must file within 60
days of the occurrence of the new event that “ripens [the]
claim” and thereby triggers the basis for a challenge. Coalition
for Responsible Regulation, Inc. v. EPA, 684 F.3d 102, 129
(D.C. Cir. 2012), rev’d in part on other grounds sub nom.
Utility Air Regulatory Group v. EPA (UARG), 573 U.S. 302
(2014); see also Alon Refining Krotz Springs, Inc. v. EPA, 936
F.3d 628, 646 (D.C. Cir. 2019); Honeywell Int’l, Inc. v. EPA,
705 F.3d 470, 472–473 (D.C. Cir. 2013); Sierra Club de Puerto
Rico v. EPA, 815 F.3d 22, 26 (D.C. Cir. 2016). A claim
“ripens” for purposes of the Clean Air Act when “subsequent
factual or legal development creat[es] new legal consequences”
for the party seeking review. Sierra Club de Puerto Rico, 815
 103
F.3d at 28. This type of delayed challenge is commonly
referred to as an “after arising” claim.

We agree with the Coal Petitioners that the ACE Rule is
an after-arising event that ripened their challenge to the New
Source Rule’s endangerment finding.

When the EPA promulgated the New Source Rule in 2015,
the Coal Petitioners did not challenge that rule’s endangerment
finding.14 That is because they did not plan “to build any new
facilities affected by the New Source Rule,” and so were not
directly affected by it. Coal Pet’rs Reply Br. 3. But when the
ACE Rule used the New Source Rule as the predicate for
regulating existing coal-fired power plants, ACE Rule, 84 Fed.
Reg. at 32,533, the Coal Petitioners became concretely
aggrieved by the finding.

Under those circumstances, the Coal Petitioners’
challenge to the New Source Rule as an insufficient predicate
for the ACE Rule is timely. If the Coal Petitioners had filed
suit when the New Source Rule was first promulgated in 2015,
their standing would have been in doubt because they did not
have any, or intend to build any, new power plants. An asserted
injury arising from how the New Source Rule might come to

14
The Coal Petitioners claim that there is no timeliness problem
because two trade associations with which the Coal Petitioners are
affiliated––the National Mining Association and the United States
Chamber of Commerce––challenged the New Source Rule. Coal
Pet’rs Reply Br. 3 & n.2. There is no evidence or declaration
regarding that relationship in the record, aside from counsel’s
representation at oral argument. Oral Argument Tr. 131:13–17.
Because we hold that the after-arising exception makes the Coal
Petitioners’ own challenge timely, we do not address the relevance,
if any, of a prior trade association challenge.
 104
affect the regulation of their existing plants in the future might
well have been too speculative to support judicial review. See
Coalition for Responsible Regulation, 684 F.3d at 115–116,
129–131 (challenge to preexisting regulations was timely,
where regulations first affected petitioners due to the recent
promulgation of rule targeting motor vehicle emissions); see
also Sierra Club de Puerto Rico, 815 F.3d at 27; Honeywell,
Int’l, 705 F.3d at 473. That is why “this court has assured
petitioners with unripe claims that ‘they will not be foreclosed
from judicial review when the appropriate time comes,’ * * *
and that they ‘need not fear preclusion by reason of the 60-day
stipulation barring judicial review,’” as long as they file a
petition within 60 days of the injury that ripened their claim.
Coalition for Responsible Regulation, 684 F.3d at 131
(formatting modified).

The EPA urges that the Coal Petitioners could have
pressed a challenge to the New Source Rule in 2015 at the
latest, as other coal-related entities did, once the EPA
promulgated the Clean Power Plan in reliance on the New
Source Rule’s endangerment finding. See North Dakota v.
EPA, No. 15-1381 (and consolidated cases).

Perhaps. See North American Coal Corp. v. EPA, No. 15-
1451 (D.C. Cir.) (consolidated with West Virginia v. EPA, No.
15-1363 (D.C. Cir.)). But that would argue over spilled milk.
The Clean Power Plan litigation came to a halt when the EPA
reconsidered that rule, and the case was ultimately dismissed
as moot after the ACE Rule withdrew the Clean Power Plan.
Per Curiam Order, West Virginia v. EPA, No. 15-1363 (D.C.
Cir. Sept. 17, 2019), ECF No. 1806952. The Coal Petitioners
have raised their claim in the ACE Rule litigation, and it would
 105
seem perverse to say they instead should have litigated the
matter in a case that will never be decided.15

c. Adequacy of the Endangerment Finding

On the merits, the Coal Petitioners press a two-fold
challenge to the EPA’s compliance with the endangerment-
finding requirement. First, they argue that Section 7411(b)
requires the EPA to make a pollutant-specific endangerment
finding for each stationary source category newly regulated
under that provision. In their view, even though the EPA had
already found that carbon dioxide emissions significantly cause
or contribute to greenhouse gas air pollution that endanger the
public health or welfare, the EPA also separately had to find
that carbon dioxide specifically from coal-fired power plants is
a significant source of that danger. 2009 Endangerment
Finding, 74 Fed. Reg. at 66,499, 66,542 (for motor vehicles).
Second, the Coal Petitioners claim that the EPA did not make
such a finding, leaving it without authority to enact the ACE
Rule.

We need not address the Coal Petitioners’ first argument.
Even assuming that Section 7411(b) requires a source-specific

15
There is a second exception to the timeliness bar known as
the “reopening rule.” See, e.g., Environmental Def. v. EPA, 467 F.3d
1329, 1333 (D.C. Cir. 2006). The gist of that rule is that the 60-day
jurisdictional review window restarts when an agency, either
explicitly or implicitly, reconsiders its former action. See National
Ass’n of Reversionary Prop. Owners v. Surface Transp. Bd., 158
F.3d 135, 141 (D.C. Cir. 1998); National Mining Ass’n v. United
States Dep’t of Interior, 70 F.3d 1345, 1351 (D.C. Cir. 1995).
Because the after-arising ripeness exception preserves the Coal
Petitioners’ claim, we need not address the reopening doctrine.
 106
endangerment finding for each pollutant, the EPA made a
sufficient finding in the New Source Rule.

i. The New Source Rule

Before making the New Source Rule’s endangerment
finding keyed to carbon dioxide from new fossil-fuel-fired
power plants, the EPA explained its “rational basis” for
regulating those sources’ emissions of that pollutant under
Section 7411. New Source Rule, 80 Fed. Reg. at 64,530. The
EPA first outlined why greenhouse gas emissions pose a
danger to the public health and welfare, and then explained
why it should regulate those emissions from power plants
specifically.

For evidence of the harms posed by greenhouse gas air
pollution, the EPA first pointed to its 2009 Endangerment
Finding, made in connection with the motor vehicle emissions
regulation at issue in Coalition for Responsible Regulation.
New Source Rule, 80 Fed. Reg. at 64,530. There, this court
upheld as reasonable the EPA’s finding that greenhouse gas
emissions threaten public health and welfare. Id.; see also
Coalition for Responsible Regulation, 684 F.3d at 119–126.

In the 2015 New Source Rule, the Agency reviewed
substantial scientific evidence, including contemporary studies
from the National Research Council, the Intergovernmental
Panel on Climate Change, and others that post-dated the record
from the 2009 motor vehicle emissions regulation. 80 Fed.
Reg. at 64,530–64,531; see also id. at 64,517–64,520 (detailing
updated developments in scientific evidence). The EPA found
that the new studies “len[t] further credence to the validity of
the [2009] Endangerment Finding.” Id. at 64,530. The EPA
added that “[n]o information that commentators have presented
or that the EPA has reviewed provides a basis for reaching a
different conclusion,” and that the science at the time had
 107
reaffirmed its understanding of the effects of greenhouse gases
on the public health and welfare. Id. “The facts,” the EPA
concluded, “unfortunately, have only grown stronger and the
potential adverse consequences to public health and the
environment more dire in the interim.” Id. at 64,531.

The EPA next explained its reasons for regulating
greenhouse gases from fossil-fuel-fired power plants
specifically, pointing to the exceptionally high levels of
emissions from those power plants. See New Source Rule, 80
Fed. Reg. at 64,522–64,523, 64,530. To that end, the EPA
found that fossil-fuel-fired power plants are the largest
stationary sources of greenhouse gas emissions in the United
States, accounting for nearly one-third of the United States’
greenhouse gas emissions and as much as three times the
emissions from the next ten categories of stationary sources
combined. Id. at 64,530. Coal-fired power plants in particular,
the EPA added, are the largest of those large emitters, with just
one coal-fired power plant emitting potentially millions of tons
of carbon dioxide annually. Id. at 64,531. In that way, power
plant emissions “far exceed[ed] in magnitude the emissions
from motor vehicles,” which had been the subject of the
endangerment finding upheld in Coalition for Responsible
Regulation. Id.

ii. All Required Findings Were Made

The Coal Petitioners acknowledge the EPA’s findings, but
argue that Section 7411 requires a two-part endangerment
finding––that carbon dioxide from fossil-fuel-fired power
plants (1) endangers the public health and welfare, and
(2) causes or contributes significantly to greenhouse gas air
pollution. See 42 U.S.C. § 7411(b)(1)(A) (findings must be for
the “category of sources”). The Coal Petitioners do not contest
 108
that carbon dioxide endangers the public health and welfare.
See Oral Argument Tr. 129:21–22.

Instead, they train their arguments on the second prong,
arguing that the New Source Rule did not properly make a
finding that fossil-fuel-fired power plants “contribute[]
significantly” to greenhouse gas pollution. First, they fault the
EPA for relying on the New Source Rule, which provided a
rational basis for regulation to support a significant-
contribution finding.16 Second, they argue that the EPA
arbitrarily and capriciously failed to define the threshold
measure of a “significant” contribution.

To survive those challenges, the EPA needed only to
“articulate a satisfactory explanation” for the New Source
Rule’s endangerment finding, making a “rational connection
between the facts found and the choice made.” Motor Vehicle
Mfrs. Ass’n v. State Farm Mut. Auto. Ins. Co. (State Farm), 463
U.S. 29, 43 (1983) (quoting Burlington Truck Lines v. United
States, 371 U.S. 156, 168 (1962)). For an endangerment

16
The Coal Petitioners also argue that the EPA was wrong to
rely on the 2009 Endangerment Finding because it used the lower
“more than a de minimis or trivial” contribution standard. Coal
Pet’rs Br. (quoting 2009 Endangerment Finding, 74 Fed. Reg.
at 66,542). But the New Source Rule relies on the 2009
Endangerment Finding only for part one of the endangerment finding
test––that greenhouse gas pollution may reasonably be anticipated to
endanger the public health and welfare––which the Coal Petitioners
do not contest. See New Source Rule, 80 Fed. Reg. at 64,530–
64,531. The EPA separately considered the volume of greenhouse
gas emissions that motor vehicles contribute to the problem and
found it significant. See 2009 Endangerment Finding, 74 Fed. Reg.
at 66,499, 66,543; Coalition for Responsible Regulation, 684 F.3d at
128.
 109
finding, that choice need not include a “precise numerical
value” that defines the threshold at which air pollution
endangers the public health and welfare. Coalition for
Responsible Regulation, 684 F.3d at 326. Instead, a “‘more
qualitative’ approach,” employing reasoned predictions based
on “empirical data and scientific evidence,” may suffice. Id. at
327 (quoting Ethyl Corp. v. EPA, 541 F.2d 1, 56 (D.C. Cir.
1976)). Such an approach “is a function of the precautionary
thrust of the [Clean Air Act] and the multivariate and
sometimes uncertain nature of climate science, not a sign of
arbitrary or capricious decision-making.” Id. By that measure,
both of the Coal Petitioners’ objections fail.

For starters, it is perfectly permissible, and commendably
efficient, for an agency to re-confirm and build consistently
upon such formally made factual determinations. It makes
eminent sense, for example, for the EPA to take what it learned
in regulating automobiles’ greenhouse gas emissions and apply
that in evaluating the need for regulation of another source of
the same pollutant––fossil-fuel-fired power plants. What
matters here is that the EPA did not simply conclude that power
plants’ greenhouse gas emissions significantly contribute to air
pollution and stop there. Instead, the EPA went on to explain
why that significant-contribution finding was warranted. See
New Source Rule, 80 Fed. Reg. at 64,530–64,531 (explaining
that power plants are the largest stationary sources of domestic
greenhouse gas emissions and that each coal-fired plant may
emit millions of tons of carbon dioxide per year).

The Coal Petitioners’ argument that the EPA failed to
articulate a specific threshold measurement for significance
fares no better. While the failure to identify the trigger point
for significance might prove problematic in cases at the
margins, the EPA sensibly found that this one is not even close.
Because of their substantial contribution of greenhouse gases,
 110
“under any reasonable threshold or definition,” carbon dioxide
from fossil-fuel-fired power plants represents “a significant
contribution” to air pollution. New Source Rule, 80 Fed. Reg.
at 64,531; cf. Massachusetts v. EPA, 549 U.S. at 525 (While
domestic automobile emissions accounted for less than one-
third of the United States’ domestic emissions, “[j]udged by
any standard, U.S. motor-vehicle emissions make a meaningful
contribution to greenhouse gas concentrations and * * * to
global warming.”).

In that regard, we have already held that nothing in the
Clean Air Act “require[s] that [the] EPA set a precise numerical
value as part of” a contribution endangerment finding.
Coalition for Responsible Regulation, 684 F.3d at 122
(applying Section 7521(a)(1) of the Clean Air Act). So the
“EPA need not establish a minimum threshold of risk or harm
before determining whether an air pollutant endangers.” Id. at
123.

Nevertheless, the Coal Petitioners insist that, before
finding significance, the EPA had to decide whether its inquiry
would (1) address domestic or global emissions, (2) be
measured by a “simple percentage criterion” or another metric,
(3) factor in historical trends and/or future projections, and
(4) involve a different process for greenhouse gases than other
pollutants. See Coal Pet’rs Br. 17. Whether the EPA could
reasonably decide to factor in such considerations is not before
us. What matters here is that nothing in the Clean Air Act or
precedent mandates determinations on each of those factors––
at least not in a case in which there is no showing that any of
them would have made any difference. Given that the United
States, at the time of the endangerment finding, was the second-
largest emitter of greenhouse gases in the world, see 2009
Endangerment Finding, 74 Fed. Reg. at 66,538, it was not
arbitrary or capricious for the EPA to conclude that the source
 111
of close to one-third of those emissions is a significant
contributor to air pollution by any measure. The global nature
of the air pollution problem means that “[a] country or a source
may be a large contributor, in comparison to other countries or
sources, even though its percentage contribution may appear
relatively small” in the context of total emissions worldwide.
Id. Looking just at the Coal Petitioners’ calculations, power
plants contributed a hefty 4.5 percent to global greenhouse gas
emissions in 2013. See Coal Pet’rs Br. 18. More to the point,
a holding that greenhouse gas emissions by fossil-fuel-fired
power plants are not significant would make it nigh impossible
for any source of greenhouse gas pollution to cross that
statutory threshold.17

For those reasons, we hold that the New Source Rule’s
endangerment finding provided a sufficient basis for the EPA’s
promulgation of the ACE Rule.

2. Section 7411 and Section 7412’s Parallel Operation

a. Background on the 1990 Amendments

The Coal Petitioners next argue that the Clean Air Act
expressly and unambiguously prohibits the EPA from

17
The EPA recently solicited public comment through a
proposed rule on the appropriateness of considering such factors
when making a significant-contribution finding. See Oil and Natural
Gas Sector: Emission Standards for New, Reconstructed, and
Modified Sources Review, 84 Fed. Reg. 50,244, 50,269 (Sept. 24,
2019). But the EPA explained that the comments on the proposed
rule are meant only “to inform the EPA’s actions in future rules,” id.
at 50,267, and explicitly declined to consider the merits of the
comments or adopt any of the factors in that final rule, see Oil and
Natural Gas Sector: Emission Standards for New, Reconstructed,
 112
regulating coal-fired power plants’ carbon dioxide emissions
under Section 7411(d) because those same power plants’
mercury emissions are regulated under Section 7412’s
Hazardous Air Pollutants provision. The relevant statutory text
says otherwise.

To set the stage, as relevant here, the Clean Air Act
regulates pollutants emitted by stationary sources like power
plants under three distinct programs: (1) the National Ambient
Air Quality Standards (NAAQS) program that applies to
emissions of six common air pollutants, 42 U.S.C. §§ 7408–
7410; (2) the regulation of certain specified pollutants under
the Hazardous Air Pollutants program, 42 U.S.C. § 7412; and
(3) the regulation of all other dangerous pollutants from new
and existing sources under Section 7411.

Congress designed the existing source provision in
Section 7411(d) to ensure that there were “no gaps in control
activities pertaining to stationary source emissions that pose
any significant danger to public health or welfare.” S. REP.
NO. 91-1196, at 20 (1970). So Section 7411(d), in its gap-
filling capacity, covers all dangerous pollutants except those
already regulated by NAAQS or the Hazardous Air Pollutants
provision. See Clean Air Act Amendments of 1990 (“1990
Amendments”), Pub. L. No. 101-549, § 108(g), § 302(a), 104
Stat. 2399, 2467, 2574.

From the passage of the Clean Air Act until its amendment
in 1990, Congress had left substantially to the EPA the task of
building a program to effectively identify and regulate
hazardous air pollutants under Section 7412. Specifically,
Section 7412(b)(1)(A)––Section 112(b)(1)(A) of the 1970

and Modified Sources Review, 84 Fed. Reg. 57,018, 57,058
(Sept. 14, 2020).
 113
Public Law––had instructed the EPA to publish a list of
hazardous air pollutants that it would then regulate under
Section 7412’s terms. See Clean Air Act Amendments of 1970
(“1970 Amendments”), Pub. L. No. 91-604, sec. 4(a),
§ 112(b)(1)(A), 84 Stat. 1676, 1685. Section 7411(d), for its
part, covered “any air pollutant * * * for which air quality
criteria have not been issued or which is not included on a list
published under section * * * 112(b)(1)(A)” by the EPA. Id.,
sec. 4(a), § 111(d)(1)(A), 84 Stat. at 1684.

After two decades, Congress found that Section 7412 had
“worked poorly” in that the EPA had regulated only eight
hazardous pollutants under Section 7412. S. REP. NO. 102-228,
at 128 (1989); see id. at 131. Through the 1990 Amendments
to Section 7412, Congress forced the EPA’s hand by statutorily
designating 191 hazardous pollutants that Congress required
the EPA to regulate. See 1990 Amendments, sec. 301,
§ 112(b)(1), 104 Stat. at 2532–2535 (codified at 42 U.S.C.
§ 7412(b)(1)); see also S. REP. NO. 102-228, at 133. Congress
also called on the EPA to add to the list. 1990 Amendments,
sec. 301, § 112(b)(2)–(3), 104 Stat. at 2535–2537 (codified at
42 U.S.C. § 7412(b)(2)–(3)). Neither greenhouse gases in
general nor carbon dioxide in particular were on Congress’
statutory list. Nor have they ever been added by the EPA.

That change to Section 7412(b) necessitated a
corresponding technical change to Section 7411(d)’s carve-out
of pollutants already regulated under the Hazardous Air
Pollutants program, since the cross-referenced “list published
under section * * * 112(b)(1)(A)” no longer existed. Congress’
update of the statutory cross-reference is the root of the present
dispute. That is because each chamber of Congress articulated
 114
the technical correction differently, and yet both were enacted
into law.

The Senate––in a section entitled “Conforming
Amendments”––passed a straightforward amendment that
struck “112(b)(1)(A)” from the Section 7411(d) exclusion, and
replaced it with “112(b)”—which is the provision containing
the new statutory list of hazardous pollutants to which the EPA
could later add. 1990 Amendments, § 302(a), 104 Stat.
at 2574. Just as before the 1990 Amendments, under the
Senate Amendment, only hazardous pollutants on the
Section 7412 list were excluded from Section 7411(d)’s
regulation of existing sources’ emissions, while dangerous
pollutants not addressed by the Hazardous Air Pollutants or
NAAQS programs remained in Section 7411(d)’s domain.

The House, for its part, called its technical amendment of
the cross-reference “Miscellaneous Guidance,” and it similarly
deleted “112(a)(1)(B)[,]” and then excluded any air pollutant
that is “emitted from a source category which is regulated
under section 112.” 1990 Amendments, § 108, 108(g), 104
Stat. at 2465, 2467.

Both of those amendments made it into the Conference
Report, H.R. REP. NO. 101-952, at 73, 183 (1990) (Conf. Rep.),
and, after being passed by both chambers of Congress and
signed by the President, they both became part of the Public
Law.

Congress’ Office of the Legal Revision Counsel is tasked
with compiling and codifying the public law and publishing it
in the United States Code. The Counsel, of course, has no
authority to alter the substance of the Statutes at Large. See
Ganem v. Heckler, 746 F.2d 844, 851 (D.C. Cir. 1984) (“[T]he
changes made by the codifiers, whose ‘choice, made * * *
without approval of Congress * * * should be given no weight,’
 115
are of no substantive moment.”) (internal quotation marks
omitted) (quoting North Dakota v. United States, 460 U.S. 300,
310 n.13 (1983)); see also Positive Law Codification, OFFICE
OF THE LAW REVISION COUNSEL, https://uscode.house.gov/
codification/legislation.shtml (last visited Jan. 11, 2021) (For
non-positive law titles, such as Title 42, “there are certain
technical, although non-substantive, changes made to the text
for purposes of inclusion in the Code.”).

When faced with the Senate and House Amendments’
differing articulations of the cross-reference update, the
Counsel chose to publish only the House Amendment in the
United States Code.

b. Interpreting the House and Senate Amendments

The Coal Petitioners argue that the House Amendment’s
technical update of the cross-reference actually worked a major
substantive change in the law by categorically and
unambiguously excluding from Section 7411 not the hazardous
pollutants already regulated under Section 7412, but any
stationary sources of hazardous pollutants regulated under
Section 7412. In their view, once a source is subject to
regulation under Section 7412 for any single listed hazardous
pollutant, all of its other pollution emissions are off limits for
regulation under Section 7411(d). More specifically, the Coal
Petitioners’ position is that, because the EPA regulates one
hazardous air pollutant––mercury––emitted from coal-fired
power plants, the EPA is powerless to regulate under
Section 7411(d) every other non-“hazardous,” but still
 116
significantly dangerous, pollutant those same power plants
emit, including greenhouse gases.18

On the other hand, for thirty years—from the enactment of
the 1990 Amendments to the present day—the EPA has read
the House’s “Miscellaneous Guidance” as just that—a
miscellaneous technical amendment that, like the Senate
Amendment, simply updated the Section 7411(d) cross-
reference to exclude the regulation of a stationary source’s
emission of pollutants that are already regulated under
Section 7412.

For the Coal Petitioners’ challenge to succeed, we would
have to agree with their ambitious reading of the House
Amendment as precluding regulation under Section 7411 of
even those pollutants that are not covered by Section 7412. We
also would have to ignore the duly enacted Senate Amendment
entirely. And we would have to reject out of hand the EPA’s
three-decade-old harmonizing reading of the statutory
amendments, the text of Section 7411(d), and the statutory
structure. We decline the invitation because that is not how
statutory interpretation works.

At the outset, the EPA seeks deference under Chevron
U.S.A. Inc. v. Natural Resources Defense Council, Inc., 467
U.S. 837 (1984). If this were an ordinary EPA interpretation
of a Clean Air Act provision, we would apply exactly that
framework. See UARG, 573 U.S. at 315 (“We review EPA’s

18
See National Emission Standards for Hazardous Air
Pollutants from Coal- and Oil-Fired Electric Utility Steam
Generating Units, 77 Fed. Reg. 9304 (Feb. 16, 2012) (regulating
mercury).
 117
interpretations of the Clean Air Act using the standard set forth
in Chevron[.]”).

But this is no ordinary case. Here, the way in which the
codifiers assembled the U.S. Code version of Section 7411(d)
by omitting the Senate Amendment conflicts with the Statutes
at Large, which is the definitive legal evidence of what the law
is. 1 U.S.C. § 112; see id. § 204(a) (United States Code
provides only prima facie evidence of the federal law). So any
ambiguity arises from our duty to textually harmonize two duly
enacted but differently articulated statutory provisions. In
undertaking that task, we need not decide whether Chevron
supplies the appropriate framework for reconciling conflicting
statutory provisions. Compare Scialabba v. Cuellar de Osorio,
573 U.S. 41, 64 (2014) (Kagan, J.) (plurality opinion), with id.
at 76 (Roberts, C.J.) (concurring in the judgment). Instead, we
independently reach the same conclusion as the EPA,
harmonizing the House and Senate Amendments by giving
“full effect” to both. Id. at 64.

i. The Consistent Meaning of Both Amendments

In reconciling the Senate and House Amendments, we start
with what the mission of the amendments was. The plain
purpose of each amendment was to update Section 7411(d)’s
outdated cross-reference to a list created by the EPA under
Section 7412(b)(1)(A), in light of Congress’ publication of its
new statutory list under Section 7412(b). That is why the
Senate labeled its provision a “[c]onforming [a]mendment,”
and the House called its version “[m]iscellaneous [g]uidance.”
See 1990 Amendments, § 302(a), 104 Stat. at 2574 (Senate
Amendment); id. § 108(g), 104 Stat. at 2465, 2467 (House
Amendment). Neither amendment was meant to work a major
substantive change in the law.
 118
The Senate took the most direct textual path to updating
Section 7411(d)’s cross-reference. Using the Public Law
section number for Section 7412 (that is, Section 112), the
Senate Amendment simply substituted “section 112(b)” for the
outdated reference to “section 112(b)(1)(A).” See 1990
Amendments, § 302(a), 104 Stat. at 2574. That way, the Senate
Amendment maintains the parallelism of the two exclusions in
Section 7411(d) for already-regulated pollutants that are either
“included on a list published under section 108(a) [NAAQS] or
112(b) [the Hazardous Air Pollutants provision.]” Id. § 302(a),
104 Stat. at 2574 (incorporating Senate Amendment into the
preexisting 1970 text, see 1970 Amendments, sec. 4(a),
§ 111(d)(1)(A), 84 Stat. at 1684). Both exclusionary clauses
continue, as they had before the 1990 Amendments, to refer
directly to specific air pollutants listed for regulation under
other statutory provisions, and so to prevent duplicate
regulation of the same harmful emissions.

The House Amendment was less efficient, but ended up in
the same place. It substituted for “section 112(b)(1)(A)” the
phrase an air pollutant that is “emitted from a source category
which is regulated under section 112[.]” 1990 Amendments,
§ 108(g), 104 Stat. at 2467 (codified at 42 U.S.C.
§ 7411(d)(1)(A)(i)). So, with the House Amendment’s
phrasing, Section 7411(d)’s exclusion reads, as relevant here,
that each State shall

establish[] standards of performance for any existing
source for any air pollutant (i) for which air quality
criteria have not been issued or which is not included
on a list published under section 7408(a) of this title
[the NAAQS program] or emitted from a source
category which is regulated under section 7412 of this
title [the Hazardous Air Pollutant program] but (ii) to
which a standard of performance under this section
 119
would apply if such existing source were a new
source[.]

42 U.S.C. § 7411(d)(1)(A).

Reading the House Amendment within Section 7411(d)(1)
“in [its] context and with a view to [its] place in the overall
statutory scheme” shows that the House Amendment, like the
Senate Amendment, just updated the cross-reference to exclude
pollutant emissions already regulated for stationary sources
under the Hazardous Air Pollutant program. King v. Burwell,
576 U.S. 473, 486 (2015) (quoting FDA v. Brown &
Williamson Tobacco Corp., 529 U.S. 120, 132, 133 (2000)).

First, the entire point of the text that follows (i)––that is,
romanette one––is to modify the phrase “air pollutant.” “Air
pollutant” is, in fact, the last antecedent to which all of the
language in romanette one speaks. And grammatically, the
last-antecedent rule means that a limiting phrase is generally
read to “modify[] only the noun or phrase that it immediately
follows.” Lockhart v. United States, 136 S. Ct. 958, 962 (2016)
(quoting Barnhart v. Thomas, 540 U.S. 20, 26 (2003)). In other
words, the whole point of romanette one, including the House
Amendment language, is to define which “air pollutant[s]”
cannot be regulated under Section 7411(d) because those same
pollutants are already regulated under the NAAQS or
Hazardous Air Pollutants programs.

Second, reading the entirety of romanette one to modify
“air pollutant” gives the updated cross-reference to
Section 7412 full meaning.19 See UARG, 573 U.S. at 317 (The

19
Contrary to the separate opinion’s view, see Separate Op. 34,
use of the term “source category” (rather than “list”) leaves open
whether the EPA might regulate, in its Section 7411(d) gap-filling
 120
phrase “any air pollutant” in Section 7411 must be given “a
reasonable, context-appropriate meaning[.]”). The EPA has
regulated over 140 source categories under Section 7412. EPA
Br. 180. But it regulates only their emission of hazardous
pollutants. In other words, Section 7412’s regulatory scheme
operates not broadly on the source category, but only on its
emissions of the specified air pollutants. So Section 7412 does
not and cannot police a source category’s every emission, only
its emission of “hazardous” air pollutants. That is why it is
called the Hazardous Air Pollutants program, not the
Hazardous Sources program. Reading Section 7411(d) as
excluding only those air pollutants already governed by
Section 7412’s emissions regulations maps exactly onto
romanette one’s parallel exclusion of pollutants (not sources)
already regulated under NAAQS. See 42 U.S.C.
§ 7411(d)(1)(A)(i). And it fits with Section 7411’s gap-filling
purpose, which is to capture those dangerous air pollutants not
covered by NAAQS or the Hazardous Air Pollutants program.
See S. REP. NO. 91-1196, at 20.

Third, at the same time that Congress amended
Section 7411(d), it also added a savings clause,

capacity, the emission even of hazardous air pollutants listed under
Section 7412 when emitted by sources that Section 7412 does not
reach, but to which Section 7411 does apply, see 42 U.S.C.
§ 7412(c)(1), (3)–(6); see also Clean Power Plan, 80 Fed. Reg. at
64,714–64,715 (stating that “both the House and Senate amendments
should be read individually as having the same meaning in the
context presented in this rule,” but that “it is reasonable to interpret
the House amendment of the Section [7412] Exclusion as only
excluding the regulation of [hazardous air pollutant] emissions under
[Clean Air Act] section [7411(d)] and only when that source
category is regulated under [Clean Air Act] section [7412.]”)
(emphasis added).
 121
Section 7412(d)(7), to the Hazardous Air Pollutants provision.
That provision says that “[n]o emission standard or other
requirement promulgated under this section shall be
interpreted, construed, or applied to diminish or replace * * *
applicable requirements established pursuant to section [7411],
part C or D[.]” 1990 Amendments, sec. 301, § 112(d)(7), 104
Stat. at 2540–2541 (codified at 42 U.S.C. § 7412(d)(7)). That
language requires reading Section 7411(d)’s simultaneously
enacted cross-reference to regulation under Section 7412
narrowly and consistently with Section 7411(d)’s
complementary role in the statutory scheme. It certainly does
not allow courts to read the cross-reference as the major
amputation of authority to regulate that the Coal Petitioners
propose.

ii. The House Amendment Is Not a Trojan Horse

The Coal Petitioners and the separate opinion eschew
reading the House and Senate updates of the cross-reference
harmoniously. They prefer to pit the House Amendment
against the Senate Amendment and espy in the former a major
change in the law that—without a word of warning or
explanation—would have significantly curtailed the regulation
of air pollutants and broadly insulated stationary sources from
regulatory oversight for their non-hazardous but still-
dangerously polluting emissions.

There is a litany of problems with that approach.

For starters, recall that the House and Senate Amendments
were meant to address an outdated statutory cross-reference. It
is not the function of a single chamber’s miscellaneous
guidance or conforming amendment of a cross-reference to
materially overhaul or truncate a statutory provision’s
operative reach. Instead, reading both amendments together as
serving the same purpose of cross-referencing a new statutory
 122
list of air pollutants fits with their legislative purpose and text.
To be sure, the Clean Air Act “is far from a chef d’oeuvre of
legislative draftsmanship,” but “we, and EPA, must do our
best, bearing in mind the ‘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.’” UARG, 573 U.S. at 320 (quoting Brown &
Williamson, 529 U.S. at 133).

More to the point, neither the House nor Senate
Amendment said anything about changing the EPA’s
affirmative regulatory obligation under Section 7411(d) to
promulgate emissions guidelines for all air pollutants, except
those already regulated under the NAAQS or Section 7412.
Yet reading the House Amendment as abruptly withdrawing
from Section 7411(d)’s reach entire source categories and all
of the otherwise-unregulated emissions they spew would put
the House Amendment in direct conflict with not only the
unambiguous language of the Senate Amendment, but also
with the Clean Air Act’s gap-filling structure and purpose, as
well as with EPA’s overarching regulatory obligation. And it
would supposedly do all of that contrary to the statutory
history, in defiance of the technical and updating nature of the
two Amendments, and without a whisper of warning by a
single House or Senate member that the miscellaneous
guidance would cripple Section 7411’s correlative function in
the statutory scheme.

At best, the Coal Petitioners’ and separate opinion’s vision
of the House Amendment would have the EPA’s regulatory
authority under Section 7411(d) turn on a fluke of timing. The
Section 7412(d)(7) savings clause mentioned above, by its
terms, protects the operation of Section 7411 regulations
already in effect. So, too, does the House Amendment, which
only excises what already “is regulated” under Section 7412.
 123
Under the Coal Petitioners’ approach, then, the Clean Air Act
would allow the EPA to regulate sources under both
Section 7411(d) and Section 7412 if, and only if, the EPA
adopted its Section 7411(d) regulation before the Section 7412
regulation. No rational explanation is offered as to why
Congress would want the mere sequencing of regulations to
render them either lawful or invalid.

More to the point, the Coal Petitioners and the separate
opinion point to nothing in the legislative record even hinting
at a rationale for removing Section 7412 sources entirely from
Section 7411’s reach. Nothing suggests that Congress
intended to veer off in that substantive legislative direction.
The Senate certainly had no such intention.

The Coal Petitioners suggest that the EPA could instead
regulate carbon dioxide under Section 7412. But they do not
really mean it, as they say in the same breath that carbon
dioxide would be a “poor fit” for Section 7412. Coal Pet’rs
Br. 33 n.8. That is because Section 7412 strictly regulates all
sources that emit ten tons per year or more of hazardous
pollutants. 42 U.S.C. § 7412(a)(1). Adding carbon dioxide to
that list would lead to a massive regulatory expansion of EPA
authority to include everything from schools to hospitals and
apartment buildings. Cf. UARG, 573 U.S. at 328. It would
make no sense to conclude that Congress intended an
unheralded string of words in a “Miscellaneous Guidance”
amendment to hobble the gap-filling function of
Section 7411(d) and to disable the EPA from addressing the
source of one-third of this country’s greenhouse gas emissions.

Nor can the Coal Petitioners hang their hats on the
inclusion of the House Amendment in the codified version of
Section 7411(d). Putting aside that the two amendments
readily can, and so must, be read harmoniously as just updating
 124
the exclusion of already-regulated air pollutants, it is settled
that “the Code cannot prevail over the Statutes at Large when
the two are inconsistent.” Stephan v. United States, 319 U.S.
423, 426 (1943); see also Five Flags Pipe Line Co. v.
Department of Transp., 854 F.2d 1438, 1440 (D.C. Cir. 1988)
(“[W]here the language of the Statutes at Large conflicts with
the language in the United States Code that has not been
enacted into positive law, the language of the Statutes at Large
controls.”).

The Coal Petitioners’ and the separate opinion’s other
efforts to cast aside the Senate Amendment all fail.

First, the Coal Petitioners and the separate opinion point to
the Chafee-Baucus Statement of Senate Managers, in which
Senators Chafee and Baucus addressed the negotiations
surrounding the “Miscellaneous Guidance” in the 1990
Amendments. Using this statement, the Coal Petitioners and
the separate opinion try to brush off the duly enacted Senate
Amendment as a scrivener’s or drafter’s error. To that end,
they stress the Managers’ statement that, in the “Conference
agreement,” the “Senate recedes to the House except * * * with
respect to the requirement regarding judicial review of reports
* * * and with respect to transportation planning[.]” 136
CONG. REC. 36,007, 36,067 (Oct. 27, 1990).

That argument does not even get out of the starting gate.
It should go without saying that two Managers’ description of
what a report said does not override the Conference Report
itself. And it surely cannot erase the Senate Amendment text
that was enacted by both the House and the Senate, and signed
into law by the President.

In fact, the Managers were wrong about what the
Conference Report said. What the Conference Report actually
says is that “the Senate recede[s] from its disagreement to the
 125
amendment of the House to the text of the bill and agree to the
same with an amendment as follows.” H.R. REP. NO. 101-952,
at 1 (Conf. Rep.) (emphasis added). The “amendment [that]
follow[ed]” included the text of the Senate Amendment as well
as the House Amendment. See id. at 73, 183. So the agreement
retained the Senate Amendment language; the Senate plainly
did not withdraw it. The accompanying joint explanatory
statement of the Conference Committee confirms that the
Senate receded to the House subject to this amendment, “which
[was] a substitute for the Senate bill and the House
amendment” and contained both the House and Senate
Amendments at issue here. See id. at 335.

Beyond that, the Chafee-Baucus statement cannot bear the
weight the Coal Petitioners and the separate opinion need it to
carry. At most, as a “statement of managers,” it purports to
summarize the more than 800-page Conference Report. 136
CONG. REC. at 36,065. We generally do not view such
statements as persuasive evidence of congressional intent, let
alone an excuse for unceremoniously discarding unambiguous
statutory text as a “drafter’s error.” See Separate Op. at 25
cf. Weyerhaeuser Co. v. Costle, 590 F.2d 1011, 1052 n.67
(D.C. Cir. 1978). Not to mention that we have specifically
ruled that this very same floor statement carries little weight.
Environmental Def. Fund, Inc. v. EPA, 82 F.3d 451, 460 n.11
(D.C. Cir. 1996). Simply put, the statement’s purpose was to
explain the report, not to change the content of the law, to
resolve substantive conflicts, or to effect sweeping change in
the statute’s reach. See Glossary Term: Statement of
Managers, U.S. SENATE, https://www.senate.gov/reference/
glossary_term/statement_of_managers.htm (last visited
Jan. 11, 2021).

Second, the Coal Petitioners argue that we should
disregard the Senate Amendment because it is a “[c]onforming
 126
[a]mendment.” See 1990 Amendments, § 302, 104 Stat.
at 2574. A conforming amendment can serve to harmonize
statutory provisions, which is exactly what the Senate
Amendment did by updating the cross-reference. See Burgess
v. United States, 553 U.S. 124, 135 (2008).

That does not mean that the statutory provision can be
ignored. See Burgess, 553 U.S. at 135. The Senate
Amendment’s careful maintenance of the status quo through a
cross-reference update evidences a deliberate preservation of
the prior regulatory scope of Section 7411.

By the way, if labels were what matters, the House’s
“Miscellaneous Guidance” provides no platform for the major
legislative surgery on Section 7411 that the Coal Petitioners
and the separate opinion envision.

Third, the Coal Petitioners ask us to defer to the Office of
the Law Revision Counsel’s decision to codify the House
Amendment rather than the Senate Amendment. The separate
opinion reasons as well that the Office of Law Revision
Counsel is “the leading candidate” for deference. Separate Op.
23.

No such deference is due. While the Office of the Law
Revision Counsel has expertise in the technical aspects of the
codification process, it has no license, without Congress’
approval, to change the substantive meaning of enacted law or
to throw away an entire statutory provision. See Ganem, 746
F.2d at 851. That is why the Public Law prevails over the
United States Code in case of conflict. See 1 U.S.C. § 112;
Stephan, 319 U.S. at 426; United States v. Welden, 377 U.S.
95, 98 n.4 (1964).

Fourth, the Coal Petitioners point to Congress’ drafting
manuals, which suggest that a first-in-time amendment, such as
 127
the House Amendment, supersedes a later-in-the-legislative-
process amendment like the Senate Amendment. See U.S.
SENATE, OFFICE OF LEGISLATIVE COUNSEL , LEGISLATIVE
DRAFTING MANUAL (“SENATE MANUAL”) § 126(d) (1997) (“If,
after a first amendment to a provision is made * * * , the
provision is again amended, the assumption is that the earlier
(preceding) amendments have been executed.”); U.S. HOUSE
OF REPRESENTATIVES, OFFICE OF LEGISLATIVE COUNSEL,
HOUSE LEGISLATIVE COUNSEL’S MANUAL ON DRAFTING
STYLE (“HOUSE MANUAL”) § 332(d) (1995) (“The assumption
is that the earlier (preceding) amendments have been
executed.”).

One problem is that the Coal Petitioners provide no
evidence that those manuals or their provisions were in place
at the time of the 1990 Amendments.

A bigger problem is that it is doubtful that the cited manual
provisions even apply in this scenario. These provisions are
located in sections for “Cumulative Amendments,” in which an
amended provision is added onto by later provisions. See
SENATE MANUAL § 126(d); HOUSE MANUAL § 332(d). Both
manuals suggest that language should be added to such a
provision to “alert the reader” to the later amendments.
SENATE MANUAL § 126(d); see also, e.g., HOUSE MANUAL
§ 332(d)(1) (suggesting the following language for a
cumulative amendment: “Title XX is amended by adding after
section 123 (as added by section 802 of this Act) the following
new section:”). That alert did not happen here. The House
Amendment in Section 108 includes no reference to the Senate
Amendment in Section 302, and there is no evidence that
Congress believed it was adopting contradictory amendments
in the final law.
 128
The biggest problem of all is that nothing in the manuals
says that a later but duly enacted amendment that has been
signed into law can be cast aside as meaningless. Nor would it
make any sense to do so here, when Congress placed the Senate
Amendment in the logical statutory position to update a cross-
reference to Section 7412. That amendment is located in the
Public Law title addressing Hazardous Air Pollutants and is the
very first provision (in Section 302 of the Public Law) that
follows the many changes to Section 7412’s Hazardous Air
Pollutants program (in Section 301 of the Public Law). See
1990 Amendments, title III, sec. 301, § 112, 104 Stat. at 2531;
id. sec. § 302(s), 104 Stat. at 2574. The House Amendment, on
the other hand, appears as “[m]iscellaneous [g]uidance” in the
title of the Public Law pertaining to the NAAQS program, not
the Hazardous Air Pollutants program. See 1990 Amendments,
title I, § 108(g), 104 Stat. at 2467.

Finally, the Coal Petitioners and the separate opinion insist
that, by subsuming the Senate Amendment’s targeted focus
within their much broader reading of the House Amendment,
they are somehow giving effect to both. See Coal Pet’rs Br.
29–30; Separate Op. 28–30. The separate opinion sees it as no
different than if a father did not want to name a child after a
president from Virginia, and a mother did not want to name the
child after any president. There is no conflict there, as the
separate opinion sees it, because the mother’s sweeping
prohibition includes “every name excluded by the father (and
then some).” Separate Op. 29–30.

But, of course, it is the “and then some” that is the
problem. By vastly overshooting the technical task of
correcting a cross-reference, the separate opinion’s and Coal
Petitioners’ proposed reading of the House Amendment is not
“supplement[ing]” the Senate Amendment’s exclusion of
duplicate regulation. Separate Op. 30. It is supplanting it by
 129
destroying the Senate Amendment’s express preservation of
Section 7411(d)’s pre-existing regulatory directive. To borrow
the analogy, the separate opinion’s vision of parental harmony
is likely to be entirely lost on the father whose heart was set on
naming his child Abraham, Theodore, or Harry.

The Coal Petitioners’ and separate opinion’s fundamental
mistake in claiming to give effect to both Amendments is that
the statute cannot mean both what the Senate Amendment says
and what they think the House Amendment says:
Section 7411(d) as amended in the 1990 Act cannot have
simultaneously preserved and eliminated Section 7411(d)’s
preexisting reach. As this case shows, the difference is quite
material: It determines whether Section 7411(d) allows any
regulation of power plants’ greenhouse gas emissions or not.
Given that, it blinks reality to claim that absorbing the Senate
Amendment into the House Amendment in the manner the
Coal Petitioners and the separate opinion propose somehow
retains the Senate Amendment’s independent effect. A mouse
swallowed by a snake, while still present in some metaphysical
way, hardly feels equally preserved.

At bottom, when confronted with two competing and duly
enacted statutory provisions, a court’s job is not to pick a
winner and a loser. The judicial duty is to read statutory text
as a harmonized whole, not to foment irreconcilability. See
Bell Atl. Tel. Cos. v. FCC, 131 F.3d 1044, 1047 (D.C. Cir.
1997) (“Where, as here, we are charged with understanding the
relationship between two different provisions within the same
statute, we must analyze the language of each to make sense of
the whole.”). Reading both amendments consistently
“pursue[s] a middle course” that “vitiates neither provision but
implements to the fullest extent possible the directives of
each[.]” Citizens to Save Spencer Cnty. v. EPA, 600 F.2d 844,
871 (D.C. Cir. 1979). Said another way, the better and quite
 130
natural reading of all of the relevant enacted statutory text,
structure, context, purpose, and history is one that harmonizes
the House and Senate Amendments, avoids determining that
one chamber of Congress smuggled dramatic and unlikely
changes to the Agency’s regulatory authority into the Act
through miscellaneous “guidance,” and instead faithfully
accomplishes the legislative adjustment needed to respond to
the changes to Section 7412.

iii. The Harmonized Reading Stands the Test of Time

Reading the two provisions consistently as successfully
performing their “conforming” and “miscellaneous” task of
updating Section 7411(d)’s cross-reference to continue to
exclude air pollutants already regulated under Section 7412
also maps onto the EPA’s consistent interpretation of the
statute. And that reading has stood the test of time, without
congressional correction. The EPA first announced its
interpretation of Section 7411(d) as excluding Section 7412’s
hazardous pollutants, rather than source categories, in the
immediate wake of the 1990 Amendments. See Standards of
Performance for New Stationary Sources and Guidelines for
Control of Existing Sources: Municipal Solid Waste Landfills,
56 Fed. Reg. 24,468, 24,469 (May 30, 1991) (explaining that
Section 7411(d) requires States to submit plans for standards
of performance for pollutants that endanger the public health
or welfare but are “not ‘hazardous’ within the meaning of
section 112 of the CAA and [are] not controlled under
sections 108 through 110 of the CAA”). The EPA has not
deviated from that interpretation in the ensuing decades. Oral
Argument Tr. 174:19–22. The EPA’s view also gives effect to
Section 7411(d)’s gap-filling purpose, see S. REP. NO. 91-
1196, at 20, by allowing it to continue to regulate dangerous
pollutants that are not policed by Section 7412 or NAAQS.
 131
The EPA’s interpretation also dovetails with the
development of judicial precedent. The Supreme Court has
specifically addressed Section 7411(d)’s regulation of carbon
dioxide emissions from fossil-fuel-fired power plants. In
American Electric Power Co. v. Connecticut (AEP), the
Supreme Court held that the Clean Air Act foreclosed any
federal common law right to challenge the regulation (or lack
thereof) of carbon dioxide emissions from power plants. 564
U.S. 410, 424–425 (2011). In so ruling, the Supreme Court
relied on the displacing force of Section 7411, and specifically
Section 7411(d). Id. In ruling that “the Clean Air Act and the
EPA actions it authorizes displace any federal common-law
right to seek abatement of carbon-dioxide emissions from
fossil-fuel fired powerplants,” the Supreme Court pointed
directly to the Section 7411 regulatory scheme, including,
“most relevant here, § 7411(d).” Id. at 424. The Supreme
Court even noted that the “EPA is currently engaged in a
§ 7411 rulemaking to set standards for greenhouse gas
emissions from fossil-fuel fired powerplants.” Id. at 425. As
the Supreme Court explained, Section 7411 “‘speaks directly’
to emissions of carbon dioxide from * * * [power] plants.” Id.
at 424.

The Coal Petitioners and the separate opinion put all their
eggs in a footnote in AEP that notes Section 7411(d)’s
exclusions. The footnote states that the “EPA may not employ
§ 7411(d) if existing sources of the pollutant in question are
regulated under the national ambient air quality standard
program, §§ 7408–7410, or the ‘hazardous air pollutants’
program, § 7412.” AEP, 564 U.S. at 424 n.7. That footnote
comports with the EPA’s harmonized reading of the House and
Senate Amendments because it says that Section 7411(d) does
not apply when “the pollutant in question” is already regulated
under one of the other two programs. See EPA Br. 189
(pointing out that the footnote’s “use of the phrase ‘of the
 132
pollutant in question’ suggests that [the Court] understood the
regulatory bar to be pollutant-specific, consistent with EPA’s
interpretation”).

The footnote could not mean otherwise. At the time of
AEP, electricity-generating power plants as sources of different
pollutants were already regulated under the NAAQS
provisions. See, e.g., American Trucking Ass’ns, Inc. v. EPA,
283 F.3d 355, 359 (D.C. Cir. 2002) (considering NAAQS for
particulate matter and ozone). So if the footnote did anything
more than generally flag a statutory exclusion for already-
regulated emissions—if it instead embraced the Coal
Petitioners’ and separate opinion’s claim that Section 7411(d)
excludes sources, rather than already-regulated emissions—
then the Court could not have ruled as it did. Specifically, it
could not have relied on Section 7411(d) to hold that the Clean
Air Act displaced the common law by “speak[ing] directly” to
the EPA’s authority to regulate power plants’ emission of
greenhouse gases. See AEP, 564 U.S. at 424. The footnote
certainly did not purport to unravel the central rationale for
AEP’s holding.

***

For all of those reasons, we hold that Section 7411(d)
allows the EPA to regulate carbon dioxide emissions from
power plants, even though mercury emitted from those same
power plants is regulated as a hazardous air pollutant under
Section 7412.

B. T HE ROBINSON PETITIONERS’ CHALLENGES

Another group of petitioners—including the Texas Public
Policy Foundation, the Competitive Enterprise Institute, and
various businesses that petitioned jointly with a forest-services
firm named Robinson Enterprises, Inc. (together, the Robinson
 133
Petitioners)—challenge the ACE Rule as overstepping the
EPA’s authority. The Robinson Petitioners are the only parties
that claim that the ACE Rule impermissibly regulates carbon
dioxide emissions using Section 7411 of the Clean Air Act
rather than Sections 7408 through 7410, under which the EPA
sets NAAQS. Our ability to consider that claim fails due to the
Robinson Petitioners’ lack of standing.

The Robinson Petitioners assert the organizational
standing of the Texas Public Policy Foundation and the
Competitive Enterprise Institute, both nonprofit, nonpartisan
organizations. Because the Foundation and the Institute seek
the same relief on the same claim, only one needs to
demonstrate standing. See American Anti-Vivisection Soc’y v.
United States Dep’t of Agric., 946 F.3d 615, 619–620 (D.C.
Cir. 2020). The two organizations argue standing based on
harm to their own activities; neither appears to be a
membership organization, and they claim no associational, or
representational, standing based on harm to members.

To establish standing, an organization, like an individual,
must show an actual or imminent injury in fact that is fairly
traceable to the challenged action and likely to be redressed by
a favorable decision. See Havens Realty Corp. v. Coleman, 455
U.S. 363, 378–379 (1982); see also American Anti-Vivisection
Soc’y, 946 F.3d at 618. Because neither organization is directly
subject to the challenged rule, their “standing is ‘substantially
more difficult to establish[.]’” Public Citizen, Inc. v. National
Highway Traffic Safety Admin., 489 F.3d 1279, 1289 (D.C. Cir.
2007) (Kavanaugh, J.) (quoting Lujan v. Defenders of Wildlife,
504 U.S. 555, 562 (1992)).

Each organization proffers a distinct ground and theory of
standing, so we analyze them in turn. The standing of both the
Texas Public Policy Foundation and the Competitive
 134
Enterprise Institute falters on the first factor, injury in fact, so
we need not consider the remaining two factors.

The Texas Public Policy Foundation states that its mission
is to provide legal counseling and services on a broad swath of
matters, including promoting “a balanced approach to
environmental regulation” by providing “legal counseling,
referral, and advocacy services to individuals and businesses
injured by federal, state, or local government overreach[.]”
Decl. of Greg Sindelar ¶¶ 5, 7 (“[Its] mission is to promote,
defend, and ensure liberty, personal responsibility, property
rights, criminal justice reform, greater educational
opportunities for all, a balanced approach to environmental
regulation, free speech, state’s rights under the 10th
Amendment, energy sufficiency, and free enterprise[.]”). The
Foundation’s attorneys litigate cases on a wide range of issues
on behalf of clients and refer clients to private counsel when
necessary. Id. ¶ 8. The Foundation claims that the challenged
rule has “caused a drain on [its] resources because [it] has had
to divert significant time, effort, and resources from [its]
activities in the area of property rights and wetlands regulation,
for example,” in order to represent clients “who are forced to
deal with” the federal regulation of greenhouse gases. Id. ¶ 9.

It is well established that injury to an organization’s
advocacy activities does not establish standing. See, e.g.,
Center for Law & Educ. v. Department of Educ., 396 F.3d
1152, 1162 n.4 (D.C. Cir. 2005) (citing Sierra Club v. Morton,
405 U.S. 727, 739–740 (1972)). That is because “the
expenditure of resources on advocacy is not a cognizable
Article III injury.” Turlock Irrigation Dist. v. FERC, 786 F.3d
18, 24 (D.C. Cir. 2015). To hold otherwise “would eviscerate
standing doctrine’s actual injury requirement” by permitting an
interest group to generate its own standing merely by putting
an issue in its lawyers’ crosshairs. Id. (quoting Center for Law
 135
& Educ., 396 F.3d at 1162 n.4); see also National Taxpayers
Union, Inc. v. United States, 68 F.3d 1428, 1434 (D.C. Cir.
1995). The Texas Public Policy Foundation declares only that,
since the EPA issued the ACE Rule, it has increased its legal
counseling, referral, and advocacy on behalf of clients affected
by the regulation of greenhouse gases rather than other clients.
That is precisely the kind of injury to advocacy—and
expenditure of resources on such efforts—that we have held
does not amount to injury in fact.

The Foundation does not show the kind of perceptible
impairment to its mission that sufficed for standing in a case
like American Anti-Vivisection Society. There, we found injury
because the agency’s inaction—specifically, its failure to
promulgate standards regarding the humane treatment of
birds—deprived the organization of key information on which
its public educational activities depended. See 946 F.3d at 619.
That inaction compelled the organization to develop guidance
for the public that otherwise would have been provided by the
agency’s standards. Id. By contrast, the Foundation fails to
allege impairment of any similarly “discrete programmatic
concerns” aside from its non-cognizable advocacy activities.
National Taxpayers Union, 68 F.3d at 1433 (quoting American
Legal Found. v. FCC, 808 F.2d 84, 92 (D.C. Cir. 1987)).

The Foundation points to Abigail Alliance for Better
Access to Developmental Drugs v. Eschenbach, 469 F.3d 129
(D.C. Cir. 2006), in arguing that the cost associated with more
legal counseling, referral, and advocacy services is a source of
injury. But the “counseling, referral, advocacy, and
educational services” at issue in Abigail Alliance were medical
services, not legal services, and they directly furthered the
plaintiff’s mission of providing access to potentially life-saving
medical drugs and treatments. See id. at 132–133. The
Foundation’s transplantation of Abigail Alliance’s words into
 136
the context of legal representation and counseling cannot
change the outcome: the costs of litigation are not a cognizable
Article III organizational injury. See Turlock, 786 F.3d at 24.

The Competitive Enterprise Institute claims a different
injury, which also falls short: the risk that it will face higher
electricity bills. The Institute works to counter “economic
overregulation in areas ranging from technology and finance to
energy and the environment,” Decl. of Kent Lassman ¶ 3, and
avers that it relies on electricity to power its headquarters in
Washington, D.C., id. ¶¶ 2, 4. It says that the Regulatory
Impact Analysis for the ACE Rule shows that the Rule could
increase its electricity costs. That analysis estimated a 0.0% to
0.1% increase in average retail electricity prices nationwide
attributable to the Rule between 2025 and 2035. See S.A. 220
(projecting baseline prices, in cents per kilowatt-hour, of 10.49
and 10.71 in 2025 and 2030, respectively, as compared to 10.50
and 10.72 under the ACE Rule, and estimating no increase
attributable to the ACE Rule by 2035).

The Regulatory Impact Analysis that the Institute cites
modeled one “illustrative policy scenario on retail electricity
prices[,]” S.A. 220, and included the caveat that the estimates
were based on “inadequate and incomplete information[,]”
meaning that “costs could be lower[,]” S.A. 222. The analysis
acknowledged that “the EPA has not analyzed or modeled a
specific standard of performance,” and recognized that costs
could vary depending on “how states might apply the [best
system of emission reduction] taking account of source-
specific factors in setting standards of performance, and how
sources might comply with those standards.” S.A. 221–222. It
also identified “several key areas of uncertainty related to the
electric power sector[,]” including electricity demand, natural
gas supply and demand, and longer-term planning by utilities.
S.A. 222.
 137
Even a small injury may suffice to support standing, see,
e.g., Competitive Enter. Inst. v. FCC (CEI), 970 F.3d 372, 384
(D.C. Cir. 2020), but it must be “concrete and particularized
and actual or imminent, not conjectural or hypothetical,” id. at
381 (quoting Spokeo, Inc. v. Robins, 136 S. Ct. 1540, 1548
(2016)). “Were all purely speculative increased risks deemed
injurious, the entire requirement of actual or imminent injury
would be rendered moot, because all hypothesized,
nonimminent injuries could be dressed up as increased risk of
future injury.” Public Citizen, 489 F.3d at 1294 (quoting
Natural Res. Def. Council v. EPA, 464 F.3d 1, 6 (D.C. Cir.
2006)).

In recognition that standing must rest on a concrete injury
that is at least imminent, “we have repeatedly held that litigants
cannot establish an Article III injury based on the independent
actions of some third party not before this court.” Turlock, 786
F.3d at 25 (formatting modified) (quoting Florida Audubon
Soc’y v. Bentsen, 94 F.3d 658, 670 (D.C. Cir. 1996) (en banc)).
“This is because ‘predictions of future events (especially future
actions taken by third parties)’ are too speculative to support a
claim of standing.” Id. (quoting United Transp. Union v. ICC,
891 F.2d 908, 912 (D.C. Cir. 1989)).

The remoteness and contingency of the prospect that the
Competitive Enterprise Institute will in the future actually face
even the tiny rate increase tentatively projected as possibly
arising from the challenged ACE Rule renders its claimed
injury speculative and thus defeats its standing. In particular,
the effect the Institute anticipates on its future electricity rates
depends on how third parties—such as electricity generators,
electricity providers, public utility commissions, and state
pollution control agencies—might react to the ACE Rule. See
EPA Br. 192. It also turns on the nature of standards that States
decide to set, and on the compliance choices of regulated
 138
sources. Id. It remains entirely unclear what standards States
would develop in response to the “best system of emission
reduction,” how and whether those standards would have any
effect on the costs of generation and transmission of energy,
and whether rates will be affected by any offsetting savings
through state or federal support for different generation mixes.
A theory that “stacks speculation upon hypothetical upon
speculation * * * does not establish an ‘actual or imminent’
injury.” Turlock, 786 F.3d at 24 (quoting New York Reg’l
Interconnect, Inc. v. FERC, 634 F.3d 581, 587 (D.C. Cir.
2011)); see Arpaio v. Obama, 797 F.3d 11, 20–23 (D.C. Cir.
2015). In asking us to anticipate the future actions of various
third parties that are not before us, the Institute does just that.

At oral argument, the Competitive Enterprise Institute
identified as its strongest support our decision in Competitive
Enterprise Institute v. FCC. But the concrete and actual injury
claimed there was traceable through “a relatively simple causal
chain[,]” 970 F.3d at 383, unlike the harm asserted here, which
is based on “inadequate and incomplete information[,]”
S.A. 222, and dependent on third parties’ unpredictable
responses to the ACE Rule. Critically, the plaintiffs there
demonstrated that their internet prices in fact had increased
since the agency took its challenged action. CEI, 970 F.3d at
382–383. This record lacks any such evidence.

Because neither the Texas Public Policy Foundation nor
the Competitive Enterprise Institute shows injury in fact to
support the Robinson Petitioners’ standing, we cannot address
the merits of their NAAQS-related challenge to the ACE Rule.

IV. AMENDMENTS TO THE IMPLEMENTING REGULATIONS

When the EPA repealed the Clean Power Plan and
finalized the ACE Rule, it also changed the longstanding
implementing regulations generally applicable to emission
 139
guidelines promulgated under Section 7411(d) of the Clean Air
Act. See ACE Rule, 84 Fed. Reg. at 32,564–32,571. The
Public Health and Environmental Organization Petitioners (the
Public Health Petitioners) challenge the implementing
regulations insofar as they adopt new timing requirements that
substantially extend the preexisting schedules for state and
federal actions and sources’ compliance under Section 7411(d).
See 40 C.F.R. §§ 60.23a(a)(1), 60.27a(b), 60.27a(c), 60.24a(d);
see also ACE Rule, 84 Fed. Reg. at 32,567. Because the
challenged regulations lack reasoned support, they cannot
stand.

The new implementing regulations extend the time
allowed for States to submit their plans, for the EPA to review
those plans, for the Agency to promulgate federal plans where
state plans fall short, and for legally enforceable consequences
to attach to sources that are slow to comply. Those extended
timeframes apply unless the EPA otherwise specifies with
respect to particular emission guidelines. See ACE Rule, 84
Fed. Reg. at 32,568. The Public Health Petitioners argue that
the amendments are arbitrary and capricious because the
Agency altogether failed to address the urgency of controlling
harmful emissions—especially the greenhouse gas emissions
accelerating climate change.

At the threshold, the EPA asserts that the Public Health
Petitioners forfeited any challenge to the amended
implementing regulations, but we conclude the claim was
preserved. The EPA contends that Petitioners “barely
mention” this claim in their opening brief, EPA Br. 268–269
(citing CTS Corp. v. EPA, 759 F.3d 52, 60 (D.C. Cir. 2014)),
but it was adequately, if concisely, set forth, see Pub. Health &
Env’t Orgs. Br. 11–13. The issue is neither particularly
complex nor as momentous as others in the case; Petitioners
nonetheless clearly stated and supported the claim with
 140
citations to the record and sources of legal authority. Id. That
relatively abbreviated treatment suffices. See, e.g., Tribune Co.
v. FCC, 133 F.3d 61, 69 n.8 (D.C. Cir. 1998) (noting one
paragraph in a fifty-eight-page brief arguing that the agency’s
action was arbitrary and capricious sufficed to preserve the
claim).

Petitioners’ joint comment on this amendment as the EPA
proposed it in the rulemaking process, which Petitioners cite in
their brief, provides more detail. See Pub. Health & Env’t
Orgs. Br. 13 (citing Comments of Environmental and Public
Health Organizations on Proposed Revisions to Emission
Guideline Implementing Regulations 26–27, J.A. 973–974).
The EPA well understands the nature of the claim, see EPA Br.
268–269, and there is no indication the brevity of the
discussion in Petitioners’ opening brief prejudiced the Agency
at all. Cf. Avia Dynamics, Inc. v. Federal Aviation Admin., 641
F.3d 515, 521 (D.C. Cir. 2011) (forfeiture excused where
federal agency was placed on notice of arguments by extensive
substantive motion practice).

On the merits, the EPA failed to justify substantially
extending established compliance timeframes, including
deadlines that it has had in place since 1975. See State Plans
for the Control of Certain Pollutants from Existing Facilities,
40 Fed. Reg. 53,340, 53,345, 53,346–53,348 (Nov. 17, 1975).
Before we can sustain agency action as nonarbitrary under the
APA, “the agency must * * * articulate a satisfactory
explanation for its action including a ‘rational connection
between the facts found and the choice made.’” Motor Vehicle
Mfrs. Ass’n of U.S., Inc. v. State Farm Mut. Auto. Ins. Co., 463
U.S. 29, 43 (1983) (quoting Burlington Truck Lines v. United
States, 371 U.S. 156, 168 (1962)). Petitioners’ comments took
issue with the tepid justifications the Agency offered, but the
heart of their challenge is the EPA’s complete failure to say
 141
anything at all about the public health and welfare implications
of the extended timeframes.

The Agency principally relied on reviving an argument it
had considered and rejected when it first adopted the schedule
it now displaces: that timeframes for the regulation of existing
sources under Section 7411(d) should necessarily mimic or
exceed timeframes for adoption of National Ambient Air
Quality Standards (NAAQS) under Section 7410 of the Clean
Air Act. Compare ACE Rule, 84 Fed. Reg. at 32,568, with
State Plans for the Control of Certain Pollutants from Existing
Facilities, 40 Fed. Reg. at 53,345. Section 7411(d) calls for
regulations that “establish a procedure similar to that provided
by [S]ection 7410[,]” which, like Section 7411, requires States
to submit plans for the EPA’s approval and, if those plans are
either not submitted or fall short, requires the EPA to itself
prescribe a plan. See 42 U.S.C. § 7410(a)(1), (c)(1); id.
§ 7411(d)(1)–(2). The two sets of rules accordingly reflect
generally similar state-federal interactions.

But it is not evident that the statement that Section 7411(d)
would use “a procedure similar” to that employed under
Section 7410 even speaks to timing rules. As the Agency
recognized when it promulgated the 1975 rule, faster
compliance was appropriate under Section 7411(d) because
plans under this provision are far simpler. They apply only to
a single category of source, whereas state plans for NAAQS
under Section 7410 cover multiple types of sources. See 40
Fed. Reg. at 53,345 (commenting that “Section [7411](d) plans
will be much less complex than the [state implementation
plans]” required under Section 7410). The Public Health
Petitioners’ comment on the 2018 proposed amendments to the
implementing regulations explained that “a section [7410 state
implementation plan] must ensure that ambient air
concentrations of a given pollutant in the state will stay below
 142
the EPA-designated standard.” J.A. 971. That goal is “far
more complicated to both achieve and demonstrate” than
limiting source emissions under Section 7411(d), because
“meeting the ambient air quality standards involves air quality
monitoring, complex modeling procedures, close attention to
such factors as topography, wind patterns, cross-[border]
transport of air pollution, and many other considerations.”
J.A. 971. By the same token, Petitioners commented that the
EPA failed to justify giving itself as much time to review the
simpler Section 7411(d) plans as it has to review state plans
under Section 7410. J.A. 971–972. The EPA failed to engage
meaningfully with the different scale of the two types of plans,
dismissing Petitioners’ comment with the conclusory assertion
that Section 7411(d) plans “have their own complexities and
realities that take time to address.” ACE Rule, 84 Fed. Reg. at
32,568.

The EPA’s proposed rule also relied on more general
claims that the amended timelines are appropriate because of
the amount of work involved in States’ plan development and
in the EPA’s review of those plans. See Emission Guidelines
for Greenhouse Gas Emissions from Existing Electric Utility
Generating Units; Revisions to Emission Guideline
Implementing Regulations; Revisions to New Source Review
Program: Proposed Rule, 83 Fed. Reg. 44,746, 44,771 (Aug.
31, 2018). In response, Petitioners commented that the EPA
did not document any problems during the decades that the
existing timelines had been in place. J.A. 972 (“If the agency
is truly concerned that the timing provision[s] in the framework
regulations are unworkable, it must provide actual evidence of
this—which it has not done thus far—and must propose
amended provisions that correspond to the actual workload
involved in section [7411(d)] rulemakings[.]”). The Final Rule
failed to fill that gap. See, e.g., ACE Rule, 84 Fed. Reg. at
32,568. Indeed, the Agency at one point seemed to forget that
 143
it even had a burden of justification under the APA, going so
far as to suggest that the obligation was somehow on the
commenters to show that the various actors do not need any
additional time. Id.

It might be a close call whether, viewed in isolation, the
analogy to Section 7410 and the general claim of need for more
processing time could supply the “rational connection” the
APA requires. State Farm, 463 U.S. at 43. But we do not view
those reasons in isolation.

The EPA’s weak grounds for routinizing additional
compliance delays in the amended implementing regulations
are overwhelmed by its total disregard of the added
environmental and public health damage likely to result from
slowing down the entire Section 7411(d) regulatory process.
“Normally, an agency rule would be arbitrary and capricious if
the agency * * * entirely failed to consider an important aspect
of the problem[.]” State Farm, 463 U.S. at 43. The extensions
of implementation deadlines here give no consideration to the
need for speed. Control of emissions from existing sources
before they harm people and the environment is the central
purpose of Section 7411(d) of the Clean Air Act. Yet when it
deferred the compliance deadlines, the EPA did not even
mention the need for prompt reduction of those emissions or
the human and environmental costs of its substantial new delay.

In their comments, Petitioners emphasized the gravity and
urgency of impending harms from unlawfully uncontrolled
emissions as a reason the EPA must retain the tighter
timeframes in the existing rule, not promulgate a new rule to
build in additional years of delay. See Comments of
Environmental and Public Health Organizations on Proposed
Revisions to Emission Guideline Implementing Regulations
26–27, J.A. 973–974. They stressed in particular the broad and
 144
longstanding scientific consensus on the role of carbon dioxide
emissions in accelerating climate change, and insisted that
“deep emission reductions are needed immediately” in order to
avoid “the worst effects of climate change,” making time “of
the utmost essence.” Id. They explained how the timing
amendments stymie effective control of carbon dioxide
emissions:

[T]he amendments in question would permit up to 60
months to elapse between the time an EPA emission
guideline is finalized and the time that affected
sources must, at a minimum, begin reducing their
emissions through enforceable increments of
progress. Assuming EPA issues a final emission
guideline for power plant [carbon dioxide] emissions
in mid-2019, designated sources can be expected to
start reducing emissions in mid-2024. * * * [T]he
world has surpassed not only the 350 ppm threshold—
that atmospheric concentration of [carbon dioxide]
that is considered the maximum safe level—but the
400 ppm threshold as well. If we are to avoid the
worst effects of climate change, deep emission
reductions are needed immediately: time is simply of
the utmost essence. For EPA to inject even further
delay into the process * * * flouts the agency’s Clean
Air Act obligation to require emission reductions to
prevent this endangerment to public health and
welfare.

Comments of Environmental and Public Health Organizations
27, J.A. 974.

Not all source categories or types of emissions subject to
Section 7411(d) present problems of the magnitude and
urgency of those posed by unregulated carbon dioxide
 145
emissions from power plants. But the Public Health
Petitioners’ comments on the Agency’s proposed amendments
to the implementing regulations squarely called on the EPA to
explain how slowing the regulatory timeframe with respect to
any covered emissions or source category might be justified
and consistent with the Act’s objective. See Comments of
Environmental and Public Health Organizations 23, J.A. 970.

In response to Petitioners’ concrete objections, the final
rule neither changed nor better justified the timing provisions.
In fact, upon reading the rule’s explanation of the deadline
extensions, one would have no idea that the EPA actually
recognized that greenhouse gas pollution was causing a global
climate crisis requiring urgent remediation. In finalizing the
proposed extensions to key deadlines, the EPA tersely
reiterated its stated interest in giving itself, States, and
regulated parties more time to comply—despite no showing of
need—and, contrary to its explanation of the rule it displaced,
stated that it was important after all to align the timing of the
Section 7411(d) state-plan process with the compliance
schedule under Section 7410. See ACE Rule, 84 Fed. Reg. at
32,564, 32,568.

The EPA did not even hint at how or whether it determined
that prolonging public exposure to ongoing harms from
pollutants emitted by existing source categories could be
justified consistent with the core objectives of the Clean Air
Act. That failure is irrational, especially in the face of the
EPA’s continued adherence to its 2015 finding of an urgent
need to counteract the threats posed by unregulated carbon
dioxide emissions from coal-fired power plants. The EPA
made no mention whatsoever of the harms that Petitioners
warned would result if the Agency slackened the pace of state
and federal action to mitigate the harms Section 7411(d)
targets. In relation to the timing amendments, pollution
 146
control—whether in the context of carbon dioxide and the ACE
Rule or air pollution more generally—was simply not on the
EPA’s agenda. In short, Petitioners called the EPA’s attention
to an important aspect of the regulatory problem, and the EPA
looked away.

The EPA offered what is at best a radically incomplete
explanation for extending the compliance timeline. It offered
undeveloped reasons of administrative convenience and
regulatory symmetry, even as it ignored the environmental and
public health effects of the Rule’s compliance slowdown. The
EPA thus “failed to consider an important aspect of the
problem,” State Farm, 463 U.S. at 43—indeed, arguably the
most important aspect. We accordingly vacate the
implementing regulations’ extensions of the Section 7411(d)
compliance periods.

V. VACATUR AND REMAND

The ACE Rule expressly rests on the incorrect conclusion
that the plain statutory text clearly foreclosed the Clean Power
Plan, so that complete repeal was “the only permissible
interpretation of the scope of the EPA’s authority under
[Section 7411].” ACE Rule, 84 Fed. Reg. at 32,534; see also
id. at 32,532. “[T]hat error prevented it from a full
consideration of the statutory question here presented.”
Negusie v. Holder, 555 U.S. 511, 521 (2009). “Where a statute
grants an agency discretion but the agency erroneously believes
it is bound to a specific decision, we [cannot] uphold the result
as an exercise of the discretion that the agency disavows,”
United States v. Ross, 848 F.3d 1129, 1134 (D.C. Cir. 2017),
and the “regulation must be declared invalid, even though the
agency might be able to adopt the regulation in the exercise of
its discretion,” Prill v. NLRB, 755 F.2d 941, 948 (D.C. Cir.
1985) (quoting Planned Parenthood Federation of America,
 147
Inc. v. Heckler, 712 F.2d 650, 666 (D.C. Cir. 1983) (Bork, J.,
concurring in part and dissenting in part)); accord Arizona v.
Thompson, 281 F.3d 248, 259 (D.C. Cir. 2002) (quoting Prill,
755 F.2d at 948).

Because the ACE Rule rests squarely on the erroneous
legal premise that the statutory text expressly foreclosed
consideration of measures other than those that apply at and to
the individual source, we conclude that the EPA fundamentally
“has misconceived the law,” such that its conclusion “may not
stand.” SEC v. Chenery Corp., 318 U.S. 80, 94 (1943).
Accordingly, we hold that the ACE Rule must be vacated and
remanded to the EPA so that the Agency may “consider the
question afresh in light of the ambiguity we see.” Negusie, 555
U.S. at 523 (quoting Cajun Elec. Power Coop., Inc. v. FERC,
924 F.2d 1132, 1136 (D.C. Cir. 1991)); accord Peter Pan Bus
Lines, Inc. v. Federal Motor Carrier Safety Admin., 471 F.3d
1350, 1354 (D.C. Cir. 2006); Prill, 755 F.2d at 948.

VI. CONCLUSION

Because promulgation of the ACE Rule and its embedded
repeal of the Clean Power Plan rested critically on a mistaken
reading of the Clean Air Act, we vacate the ACE Rule and
remand to the Agency. We also vacate the amendments to the
implementing regulations that extend the compliance timeline.
Because the objections of the Coal Petitioners are without
merit, we deny their petitions. And because the Robinson
Petitioners lack standing, their petition is dismissed.

So Ordered.
WALKER, Circuit Judge, concurring in part, concurring in the
judgment in part, and dissenting in part: This case concerns two
rules related to climate change. The EPA promulgated both
rules under § 111 of the Clean Air Act.1
A major milestone in climate regulation, the first rule set
caps for carbon emissions. Those caps would have likely
forced shifts in power generation from higher-polluting energy
sources (such as coal-fired power plants) to lower-emitting

1
When this opinion refers to § 111, it is specifically referring to
§ 111(d). The codified version of § 111(d) is titled “Standards of
performance for existing sources; remaining useful life of source.”
42 U.S.C. § 7411(d). The first part reads:

(1) The Administrator shall prescribe regulations which
shall establish a procedure similar to that provided by
section 7410 of this title under which each State shall
submit to the Administrator a plan which (A) establishes
standards of performance for any existing source for any
air pollutant (i) for which air quality criteria have not been
issued or which is not included on a list published under
section 7408(a) of this title or emitted from a source
category which is regulated under section 7412 of this title
but (ii) to which a standard of performance under this
section would apply if such existing source were a new
source, and (B) provides for the implementation and
enforcement of such standards of performance.
Regulations of the Administrator under this paragraph
shall permit the State in applying a standard of
performance to any particular source under a plan
submitted under this paragraph to take into consideration,
among other factors, the remaining useful life of the
existing source to which such standard applies.

Id.
 2
sources (such as natural gas or renewable energy sources).2
That policy is called generation shifting.
Hardly any party in this case makes a serious and sustained
argument that § 111 includes a clear statement unambiguously
authorizing the EPA to consider off-site solutions like
generation shifting. And because the rule implicates “decisions
of vast economic and political significance,” Congress’s failure
to clearly authorize the rule means the EPA lacked the authority
to promulgate it.3
The second rule repealed the first and partially replaced it
with different regulations of coal-fired power plants. Dozens
of parties have challenged both the repeal and the provisions
replacing it.
In my view, the EPA was required to repeal the first rule
and wrong to replace it with provisions promulgated under
§ 111. That’s because coal-fired power plants are already
regulated under § 112, and § 111 excludes from its scope any
power plants regulated under § 112. Thus, the EPA has no
authority to regulate coal-fired power plants under § 111.
I.
When the Constitution’s ratifiers empowered Congress to
legislate on certain matters of national importance,4 they
understood that federal regulation came with risks. For
example, Congress might impose widely disbursed costs to

2
For ease of reading, this opinion refers to the technical term “coal-
fired electric utility generating units” by the slightly less precise but
lay-friendlier term “coal-fired power plants.”
3
Utility Air Regulatory Group v. EPA, 573 U.S. 302, 324 (2014)
(quoting FDA v. Brown & Williamson Tobacco Corp., 529 U.S. 120,
160 (2000)).
4
U.S. CONST. art. I, § 8.
 3
benefit insular groups in a nation of diverse economic and
political interests. The framers called those groups factions.5
To guard against factions, legislation requires something
approaching a national consensus. While a single state’s
simple majority can often subject that state to “novel social and
economic experiments,”6 federal legislation must survive
bicameralism and presentment.7 Only through that process can
ideologically aligned states use federal power to impose their
will on the unwilling.8 So too for ideologically aligned
environmentalists. Or polluters. Or big tech. Or big labor. Or
free traders. Or fair traders. Or farmers. Or fishers. Or
butchers. Or bakers.
In that process, each political institution probes legislative
proposals from the perspective of different constituencies.9
The House speaks for the people. The Senate, among other
roles, guards the interests of small states. The Electoral
College, with representation just short of proportional, strikes
a balance between the two. And by staggering elections over
two-, four-, and six-year cycles, we further impede fleeting
factions from ganging up on small states and unpopular
political minorities. The point is: It’s difficult to pass laws —
on purpose.

5
See The FEDERALIST No. 10, at 56-65 (J. Madison) (J. Cooke ed.,
1961).
6
New State Ice Co. v. Liebmann, 285 U.S. 262, 311 (1932) (Brandeis,
J., dissenting).
7
U.S. CONST. art. I, §§ 1, 7; see INS v. Chadha, 462 U.S. 919, 951
(1983).
8
Of course, even then, a legislative coalition cannot regulate outside
Congress’s enumerated powers. See U.S. CONST. art. I, § 8.
9
See Department of Transportation v. Association of American
Railroads, 575 U.S. 43, 61 (2015) (Alito, J., concurring) (“Our
Constitution, by careful design, prescribes a process for making law,
and within that process there are many accountability checkpoints.”).
 4
This legislative gauntlet sometimes produces unfortunate,
even tragic, consequences. Between the 1870s and 1960s, it
foreclosed desperately needed civil rights laws. For budget
hawks who predict a fiscal crisis, it has blocked entitlement
reform. And for those who fear a climate crisis, it has
prevented clear congressional guidance on how to cool the
planet and who will foot the bill.10
That, however, is the price we pay for bicameralism and
presentment. Major regulations and reforms either reflect a
broad political consensus, or they do not become law.
In its clearest provisions, the Clean Air Act evinces a
political consensus. For example, according to Massachusetts
v. EPA, carbon dioxide is clearly a pollutant, and the Act’s
§ 202 unambiguously directs the EPA to curb pollution from
new cars.11
But for every carbon question answered in that case, many
more were not even presented.12 For example, does the Clean
Air Act force the electric-power industry to shift from fossil
fuels to renewable resources? If so, by how much? And who
will pay for it? Even if Congress could delegate those
decisions, Massachusetts v. EPA does not say where in the
Clean Air Act Congress clearly did so.
In 2009, Congress tried to supply that clarity through new
legislation.

10
Cf. Bowsher v. Synar, 478 U.S. 714, 722 (1986) (“That this system
of division and separation of powers produces conflicts, confusion,
and discordance at times is inherent, but it was deliberately so
structured to assure full, vigorous, and open debate on the great
issues affecting the people and to provide avenues for the operation
of checks on the exercise of governmental power.”).
11
549 U.S. 497, 532-35 (2007).
12
In this opinion, “carbon” is used as shorthand for carbon dioxide
and other greenhouse gases.
 5
The House succeeded.13
The President supported it.14
But that effort stalled in the Senate.15
Since climate change is real, man-made, and important,
Congress’s failure to act was, to many, a disappointment. But
the process worked as it was designed.16 In general, Senators
from small states blocked legislation they viewed as adverse to
their voters.17 And because small states have outsized
influence in the Senate,18 no bill arrived on the President’s

13
American Clean Energy and Security Act, H.R. 2454, 111th Cong.
(2009).
14
See Interview with President Obama on Climate Bill, N.Y. TIMES
(June 28, 2009),
https://www.nytimes.com/2009/06/29/us/politics/29climate-
text.html.
15
See Richard Cowan & Thomas Ferraro, Senator Graham Calls
Cap-and-Trade Plan Dead, REUTERS (Mar. 2, 2010, 2:26 PM),
https://www.reuters.com/article/us-climate-usa-congress/senator-
graham-calls-cap-and-trade-plan-dead-idUKTRE62142T20100302.
16
Cf. Association of American Railroads, 575 U.S. at 61 (Alito, J.,
concurring) (“Bicameralism and presentment make lawmaking
difficult by design[.]”) (cleaned up).
17
Due to opposition to the 2009 climate bill, it never received a
Senate vote. The closest analogue is the 2008 climate bill, which
received a cloture vote. And of the states with no Senator voting for
the 2008 bill, most of those states have populations smaller than 1/50
of the nation. Roll Call Vote 110th Congress – 2nd Session, U.S.
SENATE,
https://www.senate.gov/legislative/LIS/roll_call_lists/roll_call_vote
_cfm.cfm?congress=110&session=2&vote=00145#state (all internet
materials last visited Jan. 10, 2021).
18
In 2008, see supra, for twenty-four state delegations, there was no
Senate opposition to the climate bill. That’s short of a majority of
state delegations, and well short of the 3/5 necessary to break a
 6
desk. Nor have dozens of other climate-related bills introduced
since then.19
So President Obama ordered the EPA to do what Congress
wouldn’t.20 In 2015, after “years of unprecedented outreach

filibuster. But those twenty-four states equal 60% of the
population. So the Senate’s equal-state representation was
critical. If representation were proportional to population, the
climate bill would have been more likely to pass. Roll Call Vote
110th Congress – 2nd Session, U.S. SENATE,
https://www.senate.gov/legislative/LIS/roll_call_lists/roll_call_vote
_cfm.cfm?congress=110&session=2&vote=00145#state.
19
See, e.g., American Clean Energy and Security Act, H.R. 2454,
111th Cong. (2009); Integrated Energy Systems Act, S. 2702, 116th
Cong. (2019); Clean Industrial Technology Act, S. 2300, 116th
Cong. (2019); Advancing Grid Storage Act, H.R. 7313, 115th Cong.
(2018); Climate Risk Disclosure Act, S. 3481, 115th Cong. (2018);
American Energy and Conservation Act, S. 3110, 114th Cong.
(2016); Climate Solutions Commission Act, H.R. 6240, 114th Cong.
(2016); Super Pollutants Act, S. 2911, 113th Cong. (2014);
American Renewable Energy and Efficiency Act, H.R. 5301, 113th
Cong. (2014); End Polluter Welfare Act, S. 3080, 112th Cong.
(2012); Save Our Climate Act, H.R. 3242, 112th Cong. (2011);
Carbon Dioxide Capture Technology Prize Act, S. 757, 112th Cong.
(2011); Clean Energy Standard Act, S. 20, 111th Cong. (2010).
20
Evan Lehmann & Nathanael Massey, Obama Warns Congress to
Act on Climate Change, or He Will, SCIENTIFIC AMERICAN (Feb. 13,
2013), https://www.scientificamerican.com/article/obama-warns-
congress-to-act-on-climate-change-or-he-will/ (“‘But if Congress
won’t act soon to protect future generations, I will,’ Obama said. ‘I
will direct my Cabinet to come up with executive actions we can
take, now and in the future, to reduce pollution, prepare our
communities for the consequences of climate change, and speed the
transition to more sustainable sources of energy.’”).
 7
and public engagement”21 — including 4.3 million public
comments22 (about 4.25 million more than in Massachusetts v.
EPA)23 — the EPA promulgated a rule aimed at “leading global
efforts to address climate change.”24
Entitled the Clean Power Plan, the EPA’s rule used the
Clean Air Act’s § 111 to set limits for carbon emissions that
would likely be impossible to achieve at individual coal-fired
power plants because of costs, unavailable technologies, or a
need to severely reduce usage.25 In that sense, the limits
required generation shifting: shifting production from coal-
fired power plants to facilities that use natural gas or renewable
resources.
To be clear, the 2015 Rule did not expressly say, “Power
plants must adopt off-site solutions.” But it did set strict
emission limits in part by considering off-site solutions. And
those emission limits would likely have been unachievable or
too costly to meet if off-site solutions were off the table.
A political faction opposed generation shifting. It
challenged the 2015 Rule in this Court, arguing that § 111 does
not allow the EPA to consider off-site solutions when
determining the best system of emission reduction. The faction
included about twenty-four states, represented by many

21
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html.
22
Id.
23
549 U.S. at 511.
24
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html.
25
Respondents’ Br. at 32-37. For the codified text of § 111(d), see
the first footnote of this opinion.
 8
Senators who opposed the 2009 legislation.26 Conversely, a
political faction of about eighteen states defended the rule.
Many of their Senators had supported the stymied legislation.27
At that litigation’s outset, our Court refused to stay the
rule’s implementation.28 But in an unprecedented intervention,
the Supreme Court did what this Court would not.29 And
through its stay, the Supreme Court implied that the challengers
would likely succeed on the case’s merits.30
Taking the Supreme Court’s not-so-subtle hint, in 2019
President Trump’s EPA repealed the 2015 Rule and issued the
Affordable Clean Energy Rule. Like the rule it replaced, the
2019 Rule relies on the Clean Air Act’s § 111 to reduce carbon
emissions. But unlike its predecessor, the 2019 Rule did not
include generation shifting in its final determination of the best
system of emission reduction.
A new faction then challenged the 2019 Rule. It looked a
lot like the faction that had defended the 2015 Rule. Arrayed
against that faction were many states and groups that had
opposed the old rule. And so once again, politically diverse

26
See Legislative Hearing on S. 1733, Clean Energy Jobs and
American Power Act Before the Committee on Environment and
Public Works, 111th Cong. (2009) (For example, Senators from
Oklahoma, Ohio, Wyoming, and Louisiana expressed opposition or
concern about the legislation.).
27
See id. (For example, Senators from California, Massachusetts,
Minnesota, New Mexico, Oregon, Rhode Island, Vermont, and
Maryland expressed support for the legislation.).
28
West Virginia v. EPA, No. 15-1363 (D.C. Cir. Jan. 21, 2016) (per
curiam) (order).
29
West Virginia v. EPA, 136 S. Ct. 1000 (2016) (mem.).
30
See Winter v. Natural Resources Defense Council, 555 U.S. 7, 20
(2008).
 9
states and politically adverse special interest groups brought
their political brawl into a judiciary designed to be apolitical.
In this latest round, the briefing’s word count exceeded a
quarter of a million words. The oral argument lasted roughly
nine hours. The case’s caption alone runs beyond a dozen
pages. And yet, in all that analysis, hardly any of the dozens
of petitioners or intervenors defending the 2015 Rule make a
serious and sustained argument that § 111 includes a clear
statement unambiguously authorizing the EPA to consider a
system of emission reduction that includes off-site solutions or
that § 111 otherwise satisfies the major-rules doctrine’s clear-
statement requirement. Neither does the EPA.
In light of that,31 I doubt § 111 authorizes the 2015 Rule
— arguably one of the most consequential rules ever proposed
by an administrative agency:
• It required a “more aggressive transformation in the
domestic energy industry,” marking for President
Obama a “major milestone for his presidency.”32
• It aspired to reduce that industry’s carbon emissions
by 32 percent — “equal to the annual emissions
from more than 166 million cars.”33

31
Cf. ARTHUR CONAN DOYLE, Silver Blaze, in THE COMPLETE
SHERLOCK HOLMES 312, 325 (2009) (“Before deciding that question
I had grasped the significance of the silence of the dog, for one true
inference invariably suggests others.”).
32
J.A. 2076 (White House Fact Sheet).
33
FACT SHEET: Clean Power Plan By The Numbers, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-clean-power-
plan-numbers.html; What Is the Clean Power Plan?, NATIONAL
RESOURCES DEFENSE COUNCIL, INC. (Sept. 29, 2017),
https://www.nrdc.org/stories/how-clean-power-plan-works-and-
why-it-
 10
• Leaders of the environmental movement considered
the rule “groundbreaking,”34 called its
announcement “historic,” 35
and labeled it a
“critically important catalyst.”36
The potential costs and benefits of the 2015 Rule are
almost unfathomable. Industry analysts expected wholesale
electricity’s cost to rise by $214 billion.37 The cost to replace
shuttered capacity? Another $64 billion.38 (“A billion here, a
billion there, and pretty soon you’re talking real money.”39)

matters#:~:text=According%20to%20EPA%20projections%2C%20
by,nationally%2C%20relative%20to%202005%20levels.&text=Th
e%20shift%20to%20energy%20efficiency,its%20electricity%20bill
s%20in%202030 (“According to EPA projections, by 2030, the
Clean Power Plan would cut the electric sector’s carbon pollution by
32 percent nationally, relative to 2005 levels.”).
34
Save the Clean Power Plan, NATIONAL RESOURCES DEFENSE
COUNCIL, INC., https://www.nrdc.org/save-clean-power-plan.
35
The Clean Power Plan, ENVIRONMENTAL DEFENSE FUND,
https://www.edf.org/clean-power-plan-resources.
36
Press Release, Michael Brune, Sierra Club Executive Director,
Repealing the Clean Power Plan Will Threaten Thousands of Lives
(Oct. 9, 2017), https://www.sierraclub.org/press-
releases/2017/10/repealing-clean-power-plan-will-threaten-
thousands-lives.
37
EPA’s Clean Power Plan An Economic Impact Analysis, NMA, 2,
http://nma.org/attachments/article/2368/11.13.15%20NMA_EPAs
%20Clean%20Power%20Plan%20%20An%20Economic%20Impac
t%20Analysis.pdf.
38
Id.
39
Senator Everett McKinley Dirksen Dies, U.S. SENATE (Sept. 7,
1969),
https://www.senate.gov/artandhistory/history/minute/Senator_Evere
tt_Mckinley_Dirksen_Dies.htm; cf. id. (“Researchers have been
unable to track down the quotation most commonly associated with
 11
True, you can dismiss that research as industry-funded.
But the EPA itself predicted its rule would cost billions of
dollars and eliminate thousands of jobs.40
On the benefits side of the ledger, the White House labeled
the 2015 Rule a “Landmark,”41 and the President called it “the
single most important step America has ever taken in the fight
against global climate change.”42 With that in mind,
calculating the rule’s benefits requires a sober appraisal of that
fight’s high stakes. According to the rule’s advocates, victory
over climate change will lower ocean levels; preserve glaciers;
reduce asthma; make hearts healthier; slow tropical diseases;
abate hurricanes; temper wildfires; reduce droughts; stop many
floods; rescue whole ecosystems; and save from extinction up
to “half the species on earth.”43

Dirksen. Perhaps he never said it, but the comment would have been
entirely in character.”).
40
J.A. 336; see, e.g., Regulatory Impact Analysis for the Clean
Power Plan Final Rule, EPA, 6-25 (Oct. 23, 2015),
https://19january2017snapshot.epa.gov/sites/production/files/2015-
08/documents/cpp-final-rule-ria.pdf.
41
Fact Sheet: President Obama to Announce Historic Carbon
Pollution Standards for Power Plants, THE WHITE HOUSE (Aug. 3,
2015), https://obamawhitehouse.archives.gov/the-press-
office/2015/08/03/fact-sheet-president-obama-announce-historic-
carbon-pollution-standards.
42
Andrew Rafferty, Obama Unveils Ambitious Plan to Combat
Climate Change, NBC NEWS (Aug. 3, 2015, 3:05 PM),
https://www.nbcnews.com/politics/barack-obama/obama-unveils-
ambitious-plan-combat-climate-change-n403296.
43
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html; Al Gore, Al Gore: The Climate Crisis Is
the Battle of Our Time, and We Can Win, N.Y. TIMES (Sept. 20,
2019), https://www.nytimes.com/2019/09/20/opinion/al-gore-
 12
These are, to put it mildly, serious issues. Lives are at
stake. And even though it’s hard to put a dollar figure on the
net value on what many understandably consider invaluable,
the EPA tried: $36 billion, it said, give or take about a $10-
billion margin of error.44
So say what you will about the cost-benefit analysis behind
generation shifting, it’s hardly a minor question. Minor
questions do not forestall consequences comparable to “the
extinction event that wiped out the dinosaurs 65 million years
ago.”45 Minor questions are not analogous to “Thermopylae,
Agincourt, Trafalgar, Lexington and Concord, Dunkirk, Pearl
Harbor, the Battle of the Bulge, Midway and Sept. 11.”46
Minor rules do not inspire “years of unprecedented outreach
and public engagement.”47 Minor rules are not “the single most

climate-change.html; Effects of Climate Change, WORLD WILDLIFE
FUND, https://www.worldwildlife.org/threats/effects-of-climate-
change.
44
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html.
45
AN INCONVENIENT TRUTH (Lawrence Bender Productions, 2006)
(“Global warming, along with the cutting and burning of forests and
other critical habitats, is causing the loss of living species at a level
comparable to the extinction event that wiped out the dinosaurs 65
million years ago. That event was believed to have been caused by
a giant asteroid. This time it is not an asteroid colliding with the
Earth and wreaking havoc: it is us.”).
46
Al Gore, Al Gore: The Climate Crisis Is the Battle of Our Time,
and We Can Win, N.Y. TIMES (Sept. 20, 2019),
https://www.nytimes.com/2019/09/20/opinion/al-gore-climate-
change.html; see id. (“This is our generation’s life-or-death
challenge.”).
47
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html.
 13
important step America has ever taken in the fight against
global climate change.”48 Minor rules do not put thousands of
men and women out of work.49 And minor rules do not
calculate $10 billion in net benefits as their margin of error.50
Rather, the question of how to make this “the moment
when the rise of the oceans began to slow and our planet began
to heal”51 — and who should pay for it — requires a “decision[]
of vast economic and political significance.”52 That standard
is not mine. It is the Supreme Court’s. And no cocktail of
factors informing the major-rules doctrine can obscure its
ultimate inquiry: Does the rule implicate a “decision[] of vast
economic and political significance”?

48
Andrew Rafferty, Obama Unveils Ambitious Plan to Combat
Climate Change, NBC NEWS (Aug. 3, 2015, 3:05 PM),
https://www.nbcnews.com/politics/barack-obama/obama-unveils-
ambitious-plan-combat-climate-change-n403296.
49
See, e.g., Regulatory Impact Analysis for the Clean Power Plan
Final Rule, EPA, 6-25 (Oct. 23, 2015),
https://19january2017snapshot.epa.gov/sites/production/files/2015-
08/documents/cpp-final-rule-ria.pdf.
50
FACT SHEET: Overview of the Clean Power Plan, EPA,
https://archive.epa.gov/epa/cleanpowerplan/fact-sheet-overview-
clean-power-plan.html.
51
Barack Obama, Barack Obama’s Remarks in St. Paul, N.Y. TIMES
(June 3, 2008),
https://www.nytimes.com/2008/06/03/us/politics/03text-
obama.html.
52
Utility Air Regulatory Group, 573 U.S. at 324 (quoting Brown &
Williamson Tobacco, 529 U.S. at 160) (cleaned up); see Gonzales v.
Oregon, 546 U.S. 243, 267 (2006) (quoting Brown & Williamson
Tobacco, 529 U.S. at 160) (cleaned up).
 14
Proponents of the 2015 Rule say it doesn’t.53 They have
to. If it did, it’s invalid — because a clear statement is
missing.54 And according to the Supreme Court, that is exactly
what a major rule requires.
To be sure, if we frame a question broadly enough,
Congress will have always answered it. Does the Clean Air
Act direct the EPA to make our air cleaner? Clearly yes. Does
it require at least some carbon reduction? According to
Massachusetts v. EPA, again yes.
But how should the EPA reduce carbon emissions from
power plants? And who should pay for it? To those major
questions, the Clean Air Act’s answers are far from clear.
I admit the Supreme Court has proceeded with baby steps
toward a standard for its major-rules doctrine. But “big things
have small beginnings.”55 And even though its guidance has
been neither sweeping nor precise, the Supreme Court has at
least drawn this line in the sand: Either a statute clearly
endorses a major rule, or there can be no major rule.56
Moreover, if Congress merely allowed generation shifting
(it didn’t), but did not clearly require it, I doubt doing so was
constitutional. For example, imagine a Congress that says,
“The EPA may choose to consider off-site solutions for its best
system of emission reduction, but the EPA may choose not to
consider off-site solutions.” In that instance, Congress has

53
See Oral Arg. Tr. at 23 (Counsel for State and Municipal
Petitioners on the 2015 Rule: “We do not think it implicates the
Major Questions Doctrine here for a couple of reasons.”).
54
See supra p. 9.
55
LAWRENCE OF ARABIA (Columbia Pictures, 1962).
56
MCI Telecommunications Corp. v. American Telephone &
Telegraph Co., 512 U.S. 218, 230-31 (1994); Brown & Williamson
Tobacco, 529 U.S. at 126-27, 133; Gonzales, 546 U.S. at 267; Utility
Air Regulatory Group, 573 U.S. at 322-25.
 15
clearly delegated to the EPA its legislative power to determine
whether generation shifting should be part of the best system
of emission reduction — a “decision[] of vast economic and
political significance.”57
Such delegation might pass muster under a constitution
amended by “moments” rather than the “reflection and choice”
prescribed by Article V.58 But if ever there was an era when
an agency’s good sense was alone enough to make its rules
good law, that era is over.59
Congress decides what major rules make good sense. The
Constitution’s First Article begins, “All legislative Powers
herein granted shall be vested in a Congress of the United
States, which shall consist of a Senate and House of
Representatives.”60 And every “law” must “pass[] the House
of Representatives and the Senate” and “be presented to the
President.”61 Thus, whatever multi-billion-dollar regulatory

57
Utility Air Regulatory Group, 573 U.S. at 324 (2014) (quoting
Brown & Williamson Tobacco, 529 U.S. at 160) (cleaned up); see
also Gonzales, 546 U.S. at 267 (quoting Brown & Williamson
Tobacco, 529 U.S. at 160).
58
See U.S. CONST. art. V; compare BRUCE ACKERMAN, We the
People: Foundations 22 (1991) (“moments”) with MICHAEL S.
GREVE, The Upside-Down Constitution 13 (2012) (“reflection and
choice”) (quoting The FEDERALIST No. 1, at 3-7 (A. Hamilton) (J.
Cooke ed., 1961)).
59
See, e.g., SAS Institute, Inc. v. Iancu, 138 S. Ct. 1348, 1358-59
(2018) (“The Director may (today) think his approach makes for
better policy, but policy considerations cannot create an ambiguity
when the words on the page are clear. Neither may we defer to an
agency official’s preferences because we imagine some hypothetical
reasonable legislator would have favored that approach. Our duty is
to give effect to the text that actual legislators (plus one President)
enacted into law.”) (cleaned up).
60
U.S. CONST. art. I, § 1.
61
Id. § 7.
 16
power the federal government might enjoy, it’s found on the
open floor of an accountable Congress, not in the impenetrable
halls of an administrative agency — even if that agency is an
overflowing font of good sense.62
Over time, the Supreme Court will further illuminate the
nature of major questions and the limits of delegation. And
under that caselaw, federal regulation will undoubtedly endure.
So will federal regulators. Administrative agencies are
constitutional, and they’re here to stay.63
Beyond that, I leave it for others to predict what the
Supreme Court’s emerging jurisprudence may imply for those
agencies’ profiles. Here, regardless of deference and
delegation doctrines, the regulation of coal-fired power plants
under § 111 is invalid for a more mundane reason: A 1990
amendment to the Clean Air Act forbids it.

62
See id.; id. § 1; A.L.A. Schechter Poultry Corp. v. United States,
295 U.S. 495 (1935); Marshall Field & Co. v. Clark, 143 U.S. 649,
692 (1892); Gundy v. United States, 139 S. Ct. 2116, 2130-31 (2019)
(Alito, J., concurring in the judgment); see generally MIKE LEE, Our
Lost Constitution (2015); PHILIP HAMBURGER, Is Administrative
Law Unlawful? (2014); Cody Ray Milner, Comment, Into the
Multiverse: Replacing the Intelligible Principle Standard With a
Modern Multi-Theory of Nondelegation, 28 GEO. MASON L. REV.
395 (2020); cf. Talk America, Inc. v. Michigan Bell Telephone Co.,
564 U.S. 50, 68 (2011) (Scalia, J., concurring) (“When the legislative
and executive powers are united in the same person, or in the same
body of magistrates, there can be no liberty . . . .”) (quoting
MONTESQUIEU, Spirit of the Laws bk. XI, ch. 6, pp. 151-52 (O. Piest
ed., T. Nugent transl. 1949)); In re Aiken County, 725 F.3d 255, 264
(D.C. Cir. 2013) (same).
63
Gundy, 139 S. Ct. at 2145 (Gorsuch, J., dissenting) (“Nor would
enforcing the Constitution’s demands spell doom for what some call
the administrative state.”) (cleaned up).
 17
II.
The Clean Air Act Amendments of 1990 prohibit the EPA
from subjecting power plants to regulation under § 111 if they
are already regulated under § 112. The 2015 Rule and the 2019
Rule rely on § 111 for the authority to regulate coal-fired power
plants. Because the EPA already regulates those coal-fired
power plants under § 112, the rules are invalid.
A.
Before 1990, the Clean Air Act’s § 112 told the EPA to
create a list of hazardous air pollutants. Section 112 directed
the EPA to regulate the pollutants on that list. And § 111
provided authorization to regulate pollutants not on that list.
Carbon is not on the § 112 list. So, under the pre-1990
scheme, the EPA could regulate carbon under § 111.
But Congress amended § 112 in 1990. Rather than just
telling the EPA to make a § 112 list of pollutants, Congress
created its own § 112 list.
That same year, Congress also amended § 111. As a result,
the codified version of § 111 prohibits the regulation of
pollutants “emitted from a source category which is regulated
under [§ 112].”64
Coal-fired power plants are a source regulated under
§ 112.65 Therefore, under the codified version of the Clean Air
Act, coal plants cannot be regulated under § 111. And since
the 2015 Rule and the 2019 Rule use § 111 to regulate carbon
emitted from coal plants, those rules purport to do what the
codified version of § 111 says the EPA cannot.
But that is not the whole story. Congress’s Office of the
Law Revision Counsel codifies statutes. And when it

64
42 U.S.C. § 7411(d) (emphasis added).
65
Their mercury emissions are regulated under § 112.
 18
mistakenly codifies text different from the Statutes at Large,
the Statutes at Large controls.66 And the Statutes at Large
differs from the codified text here.
The question concerns two amendments, one from each
house of Congress, which both ended up in the final bill.67
Under the House Amendment:
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant (i) for
which air quality criteria have not been issued
or which is not included on a list published
under section 7408(a) of this title or emitted
from a source category which is regulated
under section 112 [of the Clean Air Act.]68

66
Cheney Railroad Co. v. Railroad Retirement Board, 50 F.3d 1071,
1076 (D.C. Cir. 1995); see also United States National Bank of
Oregon v. Independent Insurance Agents of America, Inc., 508 U.S.
439, 448 & n.3 (1993).
67
The section, before the 1990 Amendments, read:

The Administrator shall prescribe regulations which
shall establish a procedure . . . under which each
State shall submit to the Administrator a plan which
(A) establishes standards of performance for any
existing source for any air pollutant (i) for which air
quality criteria have not been issued or which is not
included on a list published under section 7408(a) or
7412(b)(1)(A) of this title . . . .

42 U.S.C. § 7411(d)(1) (1988) (emphasis added).
68
Pub. L. No. 101-549, § 108(g), 104 Stat. 2399, 2467 (1990)
(emphasis added); 42 U.S.C. § 7411(d)(1).
 19
Under the Senate Amendment:
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant (i) for
which air quality criteria have not been issued
or which is not included on a list published
under section 7408(a) of this title or 112(b) [of
the Clean Air Act.]69
Let’s compare those two versions with the most relevant
text bolded, the divergent text underlined, and the other text
struck through.
House Version:
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant (i) for
which air quality criteria have not been issued
or which is not included on a list published
under section 7408(a) of this title or emitted
from a source category which is regulated
under section 112 . . . .
Senate Version:
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant (i) for

69
Pub. L. No. 101-549, § 302(a), 104 Stat. 2399, 2574 (1990)
(emphasis added).
 20
which air quality criteria have not been issued
or which is not included on a list published
under section 7408(a) of this title or
112(b) . . . .
Finally, let’s look at only the most relevant text.
House:
The Administrator shall prescribe regulations
for any air pollutant which is not emitted from
a source category which is regulated under
section 112.
Senate:
The Administrator shall prescribe regulations
for any air pollutant which is not included on
a list published under 112(b).
To sum up so far, in my view:
• The House said the EPA can’t use § 111 to regulate
pollutants emitted from a source category regulated
under § 112.70
o Coal-fired power plants are a source
category regulated under § 112.
• The Senate said the EPA can’t use § 111 to regulate
pollutants published under § 112.
o Carbon is not a pollutant published under
§ 112.
Some parties argue the House and Senate Amendments
conflict with each other or otherwise produce an absurd result.
Others say they don’t. In my view, it doesn’t matter. If there’s

70
The EPA adopts a different interpretation of the House
Amendment. That interpretation is addressed below in Part II.C.
 21
a conflict, the House Amendment controls. And if there’s no
conflict, the Senate Amendment takes nothing away from the
House Amendment. In either scenario — conflict or no
conflict — regulation of coal-fired power plants under § 111 is
invalid.
B.
Let’s start with the first scenario: Assume the two
amendments conflict.71 If that creates an absurd result, “a
mistake of expression (rather than of legislative wisdom) [may
have] been made.”72 Such a mistake of expression — a
“scrivener’s error” — is typically viewed as a typo.73 Where
the reading “makes entire sense grammatically but produces a
disposition that makes no substantive sense,” a “drafter’s error”
may exist.74 That said, the distinction between a scrivener’s
error and a drafter’s error “is generally not a principled one.”75
Here, the Senate and House Amendments do not have obvious
typos or mistakes, but some may think that including both in
the statute “makes no substantive sense” — in the same way

71
Cf. 70 Fed. Reg. 15,994, 16,030-32 (Mar. 29, 2005) (“EPA is
therefore confronted with the highly unusual situation of an enacted
bill signed by the President that contains two different and
inconsistent amendments to the same statutory provision.”).
72
Antonin Scalia, Common-Law Courts in a Civil-Law System: The
Role of United States Federal Courts in Interpreting the Constitution
and Laws, in A MATTER OF INTERPRETATION: FEDERAL COURTS
AND THE LAW 3, 20 (Amy Gutmann ed., 1997); see also West
Virginia v. EPA, No. 15-1363, Oral Arg. Tr. at 111 (Kavanaugh, J.)
(“When [a conflict] happens[,] you [may] have a scrivener’s error.”).
73
See ANTONIN SCALIA & BRYAN A. GARNER, Reading Law: The
Interpretation of Legal Texts 234 (2012) (quoting Daniel A. Farber,
Statutory Interpretation and Legislative Supremacy, 78 Geo. L.J.
281, 289 (1989)).
74
Id. at 235.
75
Id.
 22
that a single order to “always drive fast” and “never drive fast”
makes no substantive sense.
In these rare circumstances, judges may read the text in a
way that accounts for these errors. In doing so, “we are not
revising the apparent meaning of the text.”76 Instead, we give
the text “the meaning that it would convey to a reasonable
person, who would understand that misprints had occurred.”77
But the “meaning genuinely intended but inadequately
expressed must be absolutely clear; otherwise we might be
rewriting the statute rather than correcting a technical
mistake.”78
How then to discover the “meaning genuinely intended”?
Some might say “defer to the EPA” because of the text’s
ambiguity. But unintentional ambiguity from a drafter’s error
is nothing like the intentional ambiguity that typically receives
Chevron deference. Chevron applies to deliberate gaps for an

76
Id.; see also id. at 234 (quoting Grey v. Pearson, [1857] 6 H.L.
Cas. 61, 106 (per Lord Wensleydale)) (cleaned up).
77
Id. at 235.
78
United States v. X-Citement Video, Inc., 513 U.S. 64, 82 (1994)
(Scalia, J., dissenting).
 23
agency to fill.79 So deference is arguably faithful to a statute’s
meaning — at least in theory.80
In contrast, drafter’s errors are accidents. So there’s no
reason to believe deference was “genuinely intended.” And to
the extent an office or agency with expertise is entitled to
deference here — none is81 — Congress’s Office of the Law
Revision Counsel is the leading candidate. Its whole job is to
produce the United States Code, and it dismissed the Senate
Amendment as a drafter’s error.
Others might say the default should be freedom from
regulation when a drafter’s error creates ambiguity over an
agency’s authority to promulgate a major rule. After all, if
Congress doesn’t clearly endorse a major regulation, there can
be no major regulation.82

79
Chevron, U.S.A., Inc. v. Natural Resources Defense Council, Inc.,
467 U.S. 837, 862 (1984); cf. Pereira v. Sessions, 138 S. Ct. 2105,
2121 (2018) (Kennedy, J., concurring) (“Given the concerns raised
by some Members of this Court it seems necessary and appropriate
to reconsider, in an appropriate case, the premises that
underlie Chevron and how courts have implemented that
decision. The proper rules for interpreting statutes and determining
agency jurisdiction and substantive agency powers should accord
with constitutional separation-of-powers principles and the function
and province of the Judiciary.”) (cleaned up).
80
But see Gutierrez-Brizuela v. Lynch, 834 F.3d 1142, 1153 (10th
Cir. 2016) (Gorsuch, J., concurring) (“The fact is, Chevron’s claim
about legislative intentions is no more than a fiction — and one that
requires a pretty hefty suspension of disbelief at that.”).
81
Ganem v. Heckler, 746 F.2d 844, 851 (D.C. Cir. 1984) (“[T]he
changes made by the codifiers, whose choice, made without the
approval of Congress[,] should be given no weight, are of no
substantive moment.”) (cleaned up).
82
See MCI Telecommunications, 512 U.S. at 230-31; Brown &
Williamson Tobacco, 529 U.S. at 126-27, 133; Gonzales, 546 U.S. at
267; Utility Air Regulatory Group, 573 U.S. at 322-25.
 24
But as with Chevron, the major-rules doctrine draws
meaning from ambiguity: Because Congress does not hide
elephants in mouseholes, we presume the absence of clarity
means Congress intentionally chose not to endorse a major
regulation. So as with Chevron’s premise, the premise of the
major-rules doctrine is inapplicable to a drafter’s error. Here,
to the extent an elephant’s in a mousehole, we don’t know
whether the misprint is the mousehole or the elephant.
That leaves us with a third option: inquiring into legislative
history. True, as a general matter, courts should reject any
significant reliance on legislative history. Hamilton did.83 So
did Marshall.84 And Madison.85 And Story.86 “From the
beginnings of the republic, American law followed what is
known as the ‘no-recourse doctrine’ — that in the
interpretation of a text, no recourse may be had to legislative
history.”87 And although many judges abandoned the no-
recourse doctrine by the second half of the twentieth century,88
leading textualists like Justice Scalia have made important
progress in reviving it.
But “[w]hen you have a scrivener’s error[,] everyone,
including Justice Scalia, would look at the legislative
history.”89 Indeed, he “believed that the only time it was

83
ANTONIN SCALIA & BRYAN A. GARNER, Reading Law: The
Interpretation of Legal Texts 370 (2012).
84
Id. at 370-71.
85
Id. at 371.
86
Id. at 371-72.
87
Id. at 369.
88
Id. at 388.
89
West Virginia v. EPA, No. 15-1363, Oral Arg. Tr. at 111
(Kavanaugh, J.); see also John Copeland Nagle, CERCLA’s
Mistakes, 38 WM. & MARY L. REV. 1405, 1414 (1997) (“[E]ven
textualists like Justice Scalia acknowledge that the courts can remedy
a ‘scrivener’s error’ notwithstanding plain statutory language.”).
 25
appropriate for a court to use legislative history was when there
was a credible claim of scrivener’s error.”90 For example,
concurring in the judgment in Green v. Bock Laundry Machine
Co., Justice Scalia considered “it entirely appropriate to
consult . . . legislative history . . . to verify that what seems . . .
an unthinkable disposition . . . was indeed unthought of, and
thus to justify a departure from the ordinary meaning of the
word” at issue.91
So, to recap: (1) The House and Senate Amendments may
conflict; (2) if they do, there may have been a drafter’s error;
and (3) legislative history can illuminate a drafter’s error.
What then, if anything, does the legislative history tell us?
(Buckle up.)
In 1990, the House passed a bill with many amendments
to the Clean Air Act. The Senate passed a different bill. A
Conference Committee reconciled them. But it made (at least)
two drafter’s errors — assuming again our two amendments
conflict.
First, the Conference Committee put both the House and
Senate Amendments in the Conference Report, which became
the final bill.92

90
Megan McDermott, Justice Scalia’s Bankruptcy Jurisprudence:
The Right Judicial Philosophy for the Modern Bankruptcy Code?,
2017 UTAH L. REV. 939, 974 (2017) (emphasis added).
91
490 U.S. 504, 527 (1989) (Scalia, J., concurring in the judgment).
92
The Conference Report says “[t]hat the Senate recede[s] from its
disagreement to the amendment of the House to the text of the bill
and agree[s] to the same with an amendment as follows: In lieu of
the matter proposed to be inserted by the House amendment insert
the following: . . . Sec. 108. Miscellaneous provisions.” H.R. Rep.
No. 101-952, 101st Cong., at 1 (1990) (cleaned up). Section 108(g)
under “Miscellaneous provisions” was the House Amendment that
 26
Second, the Conference Committee botched the “Joint
Explanatory Statement of the Committee of Conference.”93
The Joint Statement said, “The House amendment to the
text of the bill struck out all of the Senate bill after the enacting
clause and inserted a substitute text.”94 That “amendment”
refers to the House’s entire set of amendments to the Clean Air
Act. Clear enough so far.
The Joint Statement then said, “The Senate recedes from
its disagreement to the amendment of the House . . . .”95
Again, that seems straightforward.
But the Joint Statement didn’t stop there. The full sentence
excerpted just above says:
The Senate recedes from its disagreement to the
amendment of the House with an amendment
which is a substitute for the Senate bill and the
House amendment.96
That is drivel. The Senate recedes with an amendment?
What amendment? And how is that receding? And did the
House recede to the Senate’s amendment to the House’s
amendment that the Senate receded to?
The next day, the bill’s Senate Managers issued a
statement attempting to clarify the previous day’s materials.

struck “or 112(b)(1)(A)” and inserted “or emitted from a source
category which is regulated under section 112.” Id. at 73. But later
in the report we find the Senate’s original proposed amendment —
replacing “112(b)(1)(A)” with “112(b).” Id. at 183. So the Senate
says it receded to the House, and yet we still see the Senate’s original
language in the document.
93
Id. at 335-55.
94
Id. at 335.
95
Id.
96
Id.
 27
The statement notes that for two unrelated portions of the § 111
amendments, the House receded to the Senate.97 But it said the
Senate receded to the House regarding all other § 111 changes,
including the change at issue in this case.98
To the extent a statement by Senate Managers can ever
clear up a question of statutory meaning — count me
skeptical99 — theirs did.
Here’s where that leaves me. I’m frankly not convinced
the House and Senate Amendments are the product of a
drafter’s error. But if they are, the most lucid piece of
legislative history says the Senate intended to recede to the
House.
That would leave the House Amendment as the last man
standing. And under the House Amendment, the EPA can’t
regulate air pollutants from coal-fired power plants under § 111
when the plants are already regulated under § 112. Therefore,
if the House and Senate Amendments conflict, the 2015 Rule
and the 2019 Rule are invalid.100

97
Chafee-Baucus Statement of Senate Managers, S. 1630, the Clean
Air Act Amendments of 1990, 136 Cong. Rec. 36007, 36067 (Oct. 27,
1990).
98
Id. The full sentence about the two amendments at issue here
reads: “Conference agreement. The Senate recedes to the House
except that with respect to the requirement regarding judicial review
of reports, the House recedes to the Senate and with respect to
transportation planning, the House recedes to the Senate with certain
modifications.” In other words, except for judicial review of reports
(immaterial here) and transportation planning (immaterial here), the
Senate receded to the House.
99
Environmental Defense Fund, Inc. v. EPA, 82 F.3d 451, 460 n.11
(D.C. Cir. 1996).
100
The EPA doesn’t like that result. For thirty years it has either
ignored or misconstrued the House Amendment. But the EPA’s
 28
C.
As for the second (and more likely) of the two scenarios:
Assume the House and Senate Amendments do not conflict. In
that case, we don’t strike the Senate Amendment as a drafter’s
error.101 But even then, the House Amendment retains its full
effect.
Recall that each amendment does two things. First,
it creates a category of air pollutants. And second, it excludes
that category from regulations authorized under § 111.
For the House Amendment, that category covers any
pollutant “emitted from a source category which is regulated
under section 112.” And for the Senate Amendment, that
category covers any pollutant “published under
section . . . 112(b).”
So to see what’s in the House Amendment’s category,
you’d start by making a list of every source regulated under
§ 112. As far as § 111 regulation goes, any air pollutants from
those sources — including coal-fired power plants — are
forbidden fruit under the House Amendment.
To create the Senate Amendment’s list, you’d simply pull
the 180 or so pollutants from § 112(b), as modified by the EPA
since 1990. As far as § 111 regulation goes, those pollutants

long-running error is no reason to ignore plain text. To the extent I
glean anything from the EPA’s thirty-year mistake, it’s that the EPA
might be entitled to less deference than it thinks it deserves.
101
For the reader’s convenience, here again is the codified version of
§ 111(d): “The Administrator shall prescribe regulations which shall
establish a procedure . . . under which each State shall submit to the
Administrator a plan which (A) establishes standards of performance
for any existing source for any air pollutant (i) for which air
quality criteria have not been issued or which is not included on
a list published under [§ 108(a)] or emitted from a source
category which is regulated under [§ 112] . . . .” (emphasis added).
 29
— mercury compounds, asbestos, and more than 180 others —
are forbidden fruit under the Senate Amendment.102
In general, the House Amendment sweeps more broadly
than the Senate Amendment. For example, the House
Amendment’s list includes pollution from coal-fired power
plants, since they are regulated for mercury. So under the
House Amendment, § 111 cannot be used to regulate coal-fired
power plants at all.
In contrast, the Senate Amendment’s list includes
mercury, but it does not include all other pollution from
sources that emit mercury. So under the Senate Amendment,
§ 111 cannot be used to regulate coal-fired power plants’
emissions of mercury. But the Senate Amendment does not by
itself stop the EPA from using § 111 to regulate coal-fired
power plants’ emissions of pollutants like carbon, since carbon
isn’t on the Senate Amendment’s list.
That the House Amendment generally sweeps more
broadly than the Senate Amendment, however, does not mean
that fidelity to the House Amendment fails to give full effect to
the Senate Amendment. For example, imagine two parents
choosing a name for their child. The father says, “There’s no
way we’re naming our baby after a president from Virginia.”
And the mother says, “There’s no way we’re naming our baby
after any president.”
Just like the House and the Senate each took certain
regulations off § 111’s table, the mother and father have each
taken certain names off the table. And just as the House
Amendment excludes from § 111 every regulation excluded by

102
Initial List of Hazardous Air Pollutants with Modifications, EPA,
https://www.epa.gov/haps/initial-list-hazardous-air-pollutants-
modifications.
 30
the Senate Amendment (and then some), the mother has said
no way to every name excluded by the father (and then some).
When you give full effect to the mother’s no-way list, you
are not ignoring the father’s no-way list — because the father’s
list only excludes names and thus does not require the inclusion
of any names. And for the same reason, when you give full
effect to the father’s list, you are not ignoring the mother’s —
because the mother’s no-way list does not require the inclusion
of names excluded by the father.
Like the father’s list, the Senate Amendment has a lot to
say about what’s excluded from § 111. But like the father’s
list, the Senate Amendment says nothing about what’s
included. So when the House Amendment excludes coal-fired
power plants from § 111’s scope, it doesn’t ignore the Senate
Amendment. It supplements it — by excluding from § 111’s
scope a category of regulations not already excluded by the
Senate Amendment.
That’s the situation that will occur most often — air
pollutants excluded from § 111 regulation because they’re on
the Senate Amendment’s list will also be excluded from § 111
regulation because they’re on the House Amendment’s list.
But there may exist situations, at least in theory, when only
the Senate Amendment does any work.
For example, consider a hazardous air pollutant listed
under § 112 but “emitted by sources that Section [112] does not
reach.”103 That pollutant is barred from § 111 regulation by the
Senate Amendment (because it’s a pollutant listed under
§ 112), but it is arguably not barred by the House Amendment
(because it’s emitted from a source not regulated under § 112).

103
Majority Op. at 119-20 n.19.
 31
In that scenario, it’s possible only the Senate Amendment
would bar § 111 regulation.104
In other words, these § 111 exclusions might form a Venn
diagram: Some air pollutants are excluded from § 111
regulation only because of the House Amendment (like carbon
from coal-fired power plants), some pollutants are only
excluded because of the Senate Amendment (as in the
hypothetical I just described), and some pollutants are excluded
because of both amendments (like mercury from coal-fired
power plants). Recognizing both amendments as operative
gives “maximum possible effect” to each.105
The EPA says Chevron applies to this question. Even so,
the outcome is the same. At Chevron step one, the plain text
of the Senate Amendment takes nothing away from the plain
text of the House Amendment and vice versa. And because the
House Amendment expressly precludes the regulation of coal-
fired power plants under § 111, the plain text precludes the
2015 Rule and the 2019 Rule — both of which depended on
§ 111 to regulate coal-fired power plants.
In American Electric Power Co. v. Connecticut, the
Supreme Court agreed with this reading. It said the “EPA may
not employ § [111(d)] if existing stationary sources of the
pollutant in question are regulated under the national ambient

104
As another theoretical example, consider a source that emits a
pollutant on § 112’s list and assume the EPA is required to regulate
that source based on § 112’s parameters. But now imagine that,
notwithstanding that requirement, the EPA has not yet regulated the
source. After all, sometimes these things take time. In that situation
too, the Senate Amendment might exclude from § 111 regulation
pollutants that the House Amendment might not (yet).
105
Citizens to Save Spencer County v. EPA, 600 F.2d 844, 870 (D.C.
Cir. 1979).
 32
air quality standard program . . . or the ‘hazardous air
pollutants’ program, § [112].”106
The EPA adopts a different approach to the House
Amendment. In “any air pollutant . . . emitted from a source
category which is regulated under section 112,” the EPA reads
the phrase “which is regulated under section 112” to modify
“air pollutant,” rather than “source category.” So it would
exclude from § 111’s scope only an “air pollutant . . . which is
regulated under § 112”:

106
American Electric Power Co. v. Connecticut, 564 U.S. 410, 424
n.7 (2011) (citing § 7411(d)(1)). The EPA notes that this footnote
was dicta and that it conflicted with national ambient air quality
standard regulations at the time. But the EPA can’t have it both
ways: It can’t dismiss an inconvenient part of American Electric
Power that is directly on point and then rely on other parts of that
case where the precise meaning and contours of § 111(d) were not at
issue.

As for American Electric Power’s holding, it depended on the
Supreme Court’s understanding that § 111(d) “speaks directly” to
carbon emissions from fossil-fuel plants. Id. at 424. I agree that
§ 111(d) “speaks directly” to whether the EPA can or cannot regulate
carbon from coal-fired power plants: The provision directly says that
the EPA can regulate pollutants from existing sources unless the EPA
already regulates those sources under § 112. Compare id. with id. at
424 n.7.
 33
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant (i) for
which air quality criteria have not been issued
or which is not included on a list published
under section 108(a) or emitted from a source
category which is regulated under
section 112 . . . .
To get to the EPA’s preferred reading — to make “which
is regulated by section 112” modify “air pollutant” — the EPA
needs to read into § 111(d)(1)(A)(i) a triplet of three
whiches:107
The Administrator shall prescribe
regulations . . . under which each State shall
submit to the Administrator a plan which (A)
establishes standards of performance for any
existing source for any air pollutant [1] for
which air quality criteria have not been issued
or [2] which is not included on a list published
under § 108(a) or emitted from a source
category [3] which is [not] regulated under
§ 112 . . . .
My alterations — including [1], [2], [3], and [not] —
reflect the tripartite division implied by the EPA. But of course
the alterations were not in the original. If they were, the EPA’s
grammatically unconventional reading might work. They’re
not, so it doesn’t.
For four reasons, the EPA’s approach is not persuasive.

107
Cf. WILLIAM SHAKESPEARE, MACBETH act 1, sc. 1.
 34
First, “ordinarily, and within reason, modifiers and
qualifying phrases attach to the terms that are nearest.”108
Under that canon, a modifying phrase, such as “which is
regulated under section 112,” should apply to the closest noun
possible — “source category,” not “air pollutant.”
Second, the EPA all but reads out of § 111 the following
words: “emitted from a source category.” To be sure, Congress
will sometimes “include words that add nothing of substance,”
so the canon against surplusage has limits.109 That’s why “a
court may well prefer ordinary meaning to an unusual meaning
that will avoid surplusage.”110 But amputating the words
“emitted from a source category” does not clarify § 111’s
“ordinary meaning.” Instead, doing so transforms that
meaning.
Third, and most importantly, Congress put a conjunction
(“or”) between parts one and two of the imagined triplet, but
not between parts two and three. If the EPA’s triplet exists,
Congress’s approach to English was, to put it kindly, novel.
In formal English, you usually separate a triplet with a
conjunction between the second and third parts. (Life, liberty,
or property.) Informal English sometimes puts a conjunction
between the first and second, and between the second and third.
(Life or liberty or property.) Sometimes you see a triplet with
no conjunction. (Life, liberty, property.) But you rarely if ever
see a triplet’s conjunction separate the first and second parts

108
Grecian Magnesite Mining, Industrial & Shipping Co., SA v.
Commissioner, 926 F.3d 819, 824 (D.C. Cir. 2019); see also
Lockhart v. United States, 136 S. Ct. 958, 962 (2016); ANTONIN
SCALIA & BRYAN A. GARNER, Reading Law: The Interpretation of
Legal Texts 144-46 (2012).
109
ANTONIN SCALIA & BRYAN A. GARNER, Reading Law: The
Interpretation of Legal Texts 176 (2012).
110
Id.
 35
without also separating the second and third parts. (Life or
liberty property). That’s why it’s not:
• Stop and drop roll; or
• Red and white blue; or
• Reduce and reuse recycle; or
• Blood and sweat tears; or
• Huey and Dewey Louie.
Thus, the EPA would require us to read into § 111 a triplet
written in a way no one writes.111
Fourth and finally, the EPA says a plain-text reading of the
House Amendment would leave § 111 almost no work to do.
111
Whatever else the savings clause in § 112(d)(7) might save, it
can’t save that. Cf. 42 U.S.C. § 7412(d)(7) (“No emission standard
or other requirement promulgated under this section shall be
interpreted, construed or applied to diminish or replace the
requirements of a more stringent emission limitation or other
applicable requirement established pursuant to section 7411 of this
title, part C or D, or other authority of this chapter or a standard
issued under State authority.”).

Note that § 112(d)(7) applies only to requirements “established
pursuant to” § 111. And even the EPA says regulations cannot be
established pursuant to § 111 if they target pollutants already
regulated under § 112. See also American Electric Power, 564 U.S.
at 424 n.7. So everyone agrees the § 111 amendments exclude
something from § 111 based on § 112. And § 112(d)(7) does not
cover whatever is excluded.

What’s more, § 111(d)’s exclusion is more specific than
§ 112(d)(7)’s generalities, and the specific usually controls the
general. See ANTONIN SCALIA & BRYAN A. GARNER, Reading Law:
The Interpretation of Legal Texts 183 (2012).
 36
But if so, that was a choice for Congress. After all, the 1990
Clean Air Act Amendments added more than one hundred
pollutants to § 112’s scope, with a mechanism for the EPA to
add even more later.112 Maybe Congress thought § 111(d)
shouldn’t be much more than a rarely used gap-filler in light of
a beefed up § 112 — at least until Congress passed another law
saying otherwise.
Of course, in the end, it doesn’t matter what Congress was
thinking.113 “It is the law that governs, not the intent of the
lawgiver.”114 That’s because, among other reasons, “it is
simply incompatible with democratic government, or indeed,
even with fair government, to have the meaning of a law
determined by what the lawgiver meant, rather than by what
the lawgiver promulgated.”115
Thus, an oddity of timing doesn’t trigger Chevron
deference.116 Nor does ambiguity arise every time an agency
wishes a statutory provision did more work than it does. When
statutory text informed by structure and context is clear, “that
is the end of the matter.”117

112
Pub. L. No. 101-549, § 301, 104 Stat. 2399, 2532-37 (1990).
113
Cf. Gutierrez-Brizuela, 834 F.3d at 1153 (Gorsuch, J.,
concurring) (“Trying to infer the intentions of an institution
composed of 535 members is a notoriously doubtful business under
the best of circumstances.”).
114
Antonin Scalia, Common-Law Courts in a Civil-Law System: The
Role of United States Federal Courts in Interpreting the Constitution
and Laws, in A MATTER OF INTERPRETATION: FEDERAL COURTS
AND THE LAW 3, 17 (Amy Gutmann ed., 1997).
115
Id.
116
Cf. Public Health & Environmental Respondent-Intervenors’ Br.
at 10-11.
117
Chevron, 467 U.S. at 842.
 37
* * *
This case touches on some of administrative law’s most
consequential, unresolved issues. What is the reach of
Massachusetts v. EPA? What is the meaning of a major
question? What are the limits of congressional delegation?
Each of those issues — and a dozen or two more — might
have mattered if the EPA had relied on a section of the Clean
Air Act other than § 111 to promulgate both rules at issue in
this case. But a 1990 amendment to § 111 excluded a category
of regulations from § 111’s scope. And because that category
covers the regulations challenged today, those other legal
questions are academic.
Both houses of Congress voted that amendment — the
House Amendment — into law. And as explained above, if it
conflicts with the Senate-proposed amendment to § 111, the
Senate Amendment was a drafter’s error.
On the other hand, if the House and Senate Amendments
can coexist, the House Amendment simply excludes from
§ 111’s scope a category of regulations in addition to the
regulations excluded by the Senate Amendment.
Either way, the law precludes what the House Amendment
precludes. And the House Amendment precludes § 111
regulations of coal-fired power plants already covered by
§ 112.
Therefore, the EPA correctly repealed the 2015 Rule, but
its replacement rule improperly applied § 111 to coal-fired
power plants already regulated under § 112.
 38
Those conclusions lead to this respectful concurrence in
part, concurrence in the judgment in part, and dissent in part.118

118
The majority’s thoughtful opinion (I) describes this case’s
regulatory and procedural history; (II) vacates the 2019 Rule; (III.A)
rejects most of the Coal Petitioners’ arguments, including their
contention that the EPA cannot use § 111 to regulate carbon
emissions from power plants already regulated under § 112; (III.B)
dismisses the Robinson Petitioners’ challenge for lack of standing;
(IV) vacates the EPA’s implementing regulations for emission
guidelines promulgated under § 111(d); (V) describes the remedy;
and (VI) concludes. I concur in part of the judgment with respect to
Part II, concur with respect to Part III.B, and concur in the judgment
with respect to Part IV.