Commonwealth of Kentucky v. Federal Highway Administration

UNITED STATES DISTRICT COURT WESTERN DISTRICT OF KENTUCKY PADUCAH DIVISION COMMONWEALTH OF KENTUCKY, ET AL. v. NO. 5:23-cv-162-BJB FEDERAL HIGHWAY ADMINISTRATION, ET AL. * * * * * OPINION & ORDER The Federal Highway Administrator has issued a rule requiring each state to set declining targets for tailpipe CO2 emissions from vehicles on the National Highway System. The purported basis for this rule is a statute authorizing the Administrator to “establish … measures for the States to use to assess … the performance of the National Highway System.” 23 U.S.C. § 150 (c). Twenty-one states have sued to prevent enforcement of the rule. To them, this is a major question that Congress hasn’t or perhaps couldn’t delegate to the agency: requiring auto-emissions reductions would reach far beyond roadbuilding to dictate changes to state economies, narrow their transportation options, and displace their greenhouse-gas policy. Summary Judgment Motion (DN 72) at 10–13. A tailpipeemissions reduction forced on states under the guise of highway planning is just another example of a regulatory elephant squeezed into a statutory mousehole. See Whitman v. Am. Trucking, 531 U.S. 457, 468 (2001). During this litigation, however, the Administrator has characterized the rule as doing very little indeed. It doesn’t specify the amount or pace of CO2 decline and contains no enforcement mechanism for failing to achieve the reduction targets. SurReply (DN 84) at 5. In his view, the Final Rule is a modest, perhaps even hortatory, benchmarking exercise. The agency aims merely to use existing data to guide better state allocation of federal highway spending to eventually reduce theCO2 emissions those roads generate. Id. at 4. From this perspective, the case exemplifies only the “minor-questions doctrine”—a mere data-reporting obligation readily accepted by courts; an issue that doesn’t even implicate Chevron, much less the major-questions doctrine. Congress expressly gave the agency authority to define performance measures for the national highway system. See § 150(c). So why would states or courts care if the Administrator used these spending conditions to authorize states to fashion their own CO2 emission-reduction targets? The answer is found in the separate and carefully assigned roles given states and the federal executive under the federal highway-funding statutes—and indeed under our constitutional structure. Lawyers and judges, by now trained to consider multiple readings and inquire which is reasonable and which is major—not simply what is correct—are perhaps too quick to ascribe vastly different implications to the same agency action. But here, at least, Congress supplied a clear and sensible instruction: the Administrator may set standards and measures that states use to plan and assess the National Highway System. The Administrator, who retains significant delegated authority over the federal spending program, may review state planning reports for compliance and potentially even withhold conditional federal funding. But what the Administrator may not do is step into the shoes of sovereign states, which set their own targets for any standards and measures established by the agency. Understanding the scope of this reticulated statutory scheme, described below, allows courts approaching arguably “major” policy issues to first measure the size of any would-be mouseholes—and only thereafter determine whether what lies inside is an elephant. See Thomas B. Griffith, Haley N. Proctor, Deference, Delegation, and Divination: Justice Breyer and the Future of the Major Questions Doctrine, 132 Yale L.J. Forum 693, 699 (2022). If Congress did purport to give the Administrator authority to set state policy, that would raise a different and arguably bigger problem. Modern constitutional doctrine allows Congress to demand much from states, but it cannot commandeer or coerce the apparatus of state governments into mere administrative districts of the federal government. See, e.g., Printz v. United States, 521 U.S. 898, 935 (1997). If the Administrator were allowed to shove national greenhouse-gas policy into the mouths of uncooperative state Departments of Transportation, this would corrupt the separation of sovereigns central to our lasting and vibrant system of federalism. Neither the Constitution nor the Administrative Procedure Act authorizes administrative ventriloquism. Given this lack of a statutory basis to require declining CO2 targets, the Final Rule cannot be justified as a tool to influence state decisionmaking regarding emissions reductions—nor as a mere reporting requirement. The assumption that the Final Rule will influence state behavior is based on nothing more than the Administrator’s unfounded hope. And the state and federal governments apparently already have the data needed to calculate tailpipe CO2 emissions on the National Highway System. So the Rule’s purpose cannot be gathering previously unavailable information. Despite that existing data, moreover, the 21 State Plaintiffs insist they wouldn’t seek to achieve declining CO2 admissions from their national highway system roadways. Absent a lawful goal that this Rule reasonably advances, it is invalid: a statutorily unsupported and substantively capricious exercise of the Administrator’s rulemaking authority. But because the Final Rule operates on a state-by-state basis, with no one state’s compliance or coercion affecting that of any other state, the remedy is appropriately dispensed to the Plaintiff States only, not nationwide. Given potentially conflicting requirements from other courts and the prospect of a tailored injunction, however, the Court asks the parties to file supplemental briefs on the proper remedy within 21 days. I. Standing & Venue Are these Plaintiff States properly before this particular Court? The Defendants contend that the U.S. District Court for the Western District of Kentucky lacks subject-matter jurisdiction because Kentucky lacks standing. Cross-Motion for Summary Judgment (DN 81-3) at 9–12. More specifically, they argue Kentucky has not carried its burden of proof to establish standing at summary judgment because it hasn’t filed an affidavit describing any injury caused by the Final Rule. The Defendants don’t contest that other Plaintiff States have standing, however. Oral Arg. Tr. (DN 91) at 8:3–8; Sur-reply at 4. And, under current doctrine, one plaintiff’s standing suffices for all. Rumsfeld v. FAIR, Inc., 547 U.S. 47 , 52 n.2 (2006) (“[T]he presence of one party with standing is sufficient to satisfy Article III’s case-orcontroversy requirement.”). But this still might leave a venue problem: if “Kentucky is dismissed” for lack of standing, they contend, “neither 28 U.S.C. § 1391 (e) nor any other authority provides a basis for venue in this district.” Cross-Motion for Summary Judgment at 13. So the Defendants ask the Court to dismiss this case without prejudice or transfer it to either the Southern District of Ohio or the District of Columbia, where they admit this challenge could be heard. Id. at 14, n.7 ; DN 84 at 4. Constitutional standing—injury, causation, and redressability adding up to a “case or controversy” under Article III—is “not [a] mere pleading requiremen[t].” Lujan v. Defenders of Wildlife, 504 U.S. 555, 561 (1992). “[E]ach element must be supported in the same way as any other matter on which the plaintiff bears the burden of proof, i.e., with the manner and degree of evidence required at the successive stages of the litigation.” Id. At summary judgment, this requires “specific facts” that are “set forth” by “affidavit or other evidence.” Id. “[O]rdinarily,” however, “there is … little question” that a plaintiff has standing when it “is [it]self an object of the action (or forgone action) at issue.” Id. That is the precise situation here. The Final Rule operates on State Departments of Transportation directly; they are the “object” of the emissions target-and-reporting requirements. Final Rule (DN 1-1). So they have standing to complain that the Administrator overstepped his authority in imposing those requirements. This amply justifies opening the courthouse doors to the Commonwealth, regardless of any “special solicitude” the Constitution may afford states. See, e.g., Massachusetts v. EPA, 549 U.S. 497, 520 (2007) (“[T]he Commonwealth is entitled to special solicitude in our standing analysis.”); Arizona v. Biden, 40 F.4th 375 , 385–86 (6th Cir. 2022) (“States … may incur ‘quasi-sovereign’ injuries that private parties cannot.”).1 The 1 The Defendants don’t really dispute that Kentucky is in fact injured (in a legal sense) by the Final Rule. Oral Arg. Tr. at 8:4–8 (“Our argument is just that Kentucky hasn’t met its burden of establishing standing. We’re not contesting that either of these theories of standing, you know, could suffice…”). Rather they argue in their sur-reply that Kentucky’s injury is not “self-evident,” pointing to other states that have not challenged the Final Rule: “Some states may have concluded that the Final Rule is beneficial … [o]r states may already have similar programs … and thus concluded that any compliance costs associated with the Final Rule operates directly on an arm of sovereign state governments, so “there can be ‘little question’ that the Final Rule does injure the States, since they are ‘the object of’ its requirement….” West Virginia v. EPA, 597 U.S. 697 , 719 (2022) (quoting Lujan, 504 U.S. at 561–62). When pressed on this point, neither side could identify a single precedent holding that “states lack standing to challenge a rule that operates on States qua States.” Oral Arg. Tr. at 4:1–20. The Defendants pointed only to readily distinguishable precedents in which an agency action challenged by a non-state plaintiff directly regulated private actors, not the states themselves. See Tennessee Republican Party v. SEC, 863 F.3d 507 , 514–17 (6th Cir. 2017); Block Communications, Inc. v. FCC, 808 F. App’x 332 , 334–35 (6th Cir. 2020). No additional proof is necessary when a rule purports to impose legal obligations directly on a state plaintiff. Even if proof were required, Kentucky would still have standing. Despite the lack of a state-specific affidavit, the factual record proves the Final Rule injures Kentucky. “[C]ompliance costs are a recognized harm for purposes of Article III.” Kentucky v. Yellen, 54 F.4th 325 , 342 (6th Cir. 2022). The Defendants concede that other States asserting the same injury (the costs of setting emissions targets) have established standing. See DN 84 at 4 n.3 (accepting that venue would be proper in the Southern District of Ohio); DN 72-11 ¶ 8 (declaration that the Ohio Department of Transportation must “take on certain obligations” to comply with the Final Rule). The administrative record says much the same thing as such affidavits—and the record covers Ohio, Kentucky, and every other state. The Administrator’s own Regulatory Impact Analysis estimates that each state will incur $636,708 in compliance costs to establish the initial CO2 targets. See Initial Regulatory Impact Analysis at FHWA00099; Final Regulatory Impact Analysis at FHWA000196–97.2 Over a ten-year span, implementing the Final Rule will cost states between $10.8 million and $12.7 million. Final Regulatory Impact Analysis at FHWA000205. Kentucky obviously will incur a portion of these costs, which are traceable to the Final Rule: the very purpose of the Regulatory Impact Analysis is to estimate “the economic impact of the final rule in terms of costs that will be incurred by Federal, State, and local governments.” Final Regulatory Impact Analysis at FHWA000177. And this injury is redressable by a favorable decision. If the Court enjoins the Final Rule, Kentucky would not have to incur compliance costs because it would not have to Final Rule are either negligible or outweighed by those states’ ability to eliminate a redundant state program.” Sur-Reply at 3. Whatever other states have chosen to do, of course, is irrelevant to whether Kentucky may be compelled to do. A property owner, by analogy, is no less injured by a trespasser on his lawn because his neighbor would welcome the same person as an invitee. 2 Other courts have likewise relied upon regulatory impact analyses to establish standing. E.g., California v. Azar, 911 F.3d 558, 572 , (9th Cir. 2018); see also Sierra Club v. EPA, 292 F.3d 895 , 899–900 (D.C. Cir. 2002) (“In many if not most cases … no evidence outside the administrative record is necessary for the court to be sure of [plaintiff’s standing.]”). comply. See Kentucky v. Biden, 23 F.4th 585 , 601 (6th Cir. 2022) (enjoining federal mandate would redress state injuries). So the only question is whether some procedural or evidentiary rule limits the standing analysis to declarations and affidavits (which Kentucky lacks) or instead allows consideration of the administrative record (which covers Kentucky and every other state). Defense counsel has identified no such limit, which would stand at odds with basic administrative-law practice. Affidavits and declarations are not, in fact, the only ways to present proof establishing standing. “[O]n review of a final agency action, the petitioner must present specific facts supporting standing through citations to the administrative record or affidavits or other evidence … unless standing is self-evident.” Tennessee Republican Party, 863 F.3d at 517 (emphasis added) (cleaned up). By pointing to the Administrator’s own administrative record— specifically the Regulatory Impact Analysis—Kentucky has established a redressable injury in the form of compliance costs that flow from the Final Rule under review. So Kentucky has demonstrated its standing on summary judgment—both “through citations to the administrative record” and because “standing is selfevident” based on the Final Rule’s basic operation. And because Kentucky has standing, venue is proper in this district: “A civil action in which a defendant is an officer or employee of the United States or any agency thereof acting in his official capacity … may … be brought in any judicial district in which … the plaintiff resides if no real property is involved in the action.” 28 U.S.C. § 1391 (e). That means the Court can proceed to the merits. II. The Statutory Scheme Congress has provided highway funding to the States for more than 100 years. See FEDERAL HIGHWAY PROGRAMS: IN BRIEF, Congressional Research Service 1 (2022); South Dakota v. Dole, 483 U.S. 203 , 205–06 (1987) (considering whether conditional highway funding program violated the Spending Clause). In 2012, Congress enacted the Moving Ahead for Progress in the 21st Century Act, which (among many other things) created “performance management requirements” for the Federal-aid highway program. That grant-in-aid program funds—with conditions—what is largely state-directed construction and maintenance of the National Highway System. Cross-Motion for Summary Judgment at 4; see also 23 U.S.C. § 150 (a). This conditional-spending policy involves a form of “Cooperative Federalism” (or “marble-cake federalism”). See generally Ernest A. Young, A Research Agenda for Uncooperative Federalists, 48 TULSA L. REV. 427, 430 (2013) (citing Morton Grodzins, THE AMERICAN SYSTEM 8 (Daniel Elazar ed., 1966)). Congress called on its Spending Power to achieve national goals through the apparatus of state government. See id. at 440–41. Rather than “issu[ing] directives requiring the States to address particular problems,” Printz, 521 U.S. at 935 , conditional spending programs operate “much in the nature of a contract,” see Pennhurst State School & Hospital v. Halderman, 451 U.S. 1, 17 (1981). This performance-management scheme assigns different roles to Congress, state Departments of Transportation, and the Administrator of the Federal Highway Administration (exercising authority assigned to the Federal Secretary of Transportation but delegated to the Administrator, see 49 C.F.R. § 1.85 (a)). First, Congress set a number of “national goals” to “focus the Federal-aid highway program….” § 150(b). Second, the Administrator, exercising authority delegated by Congress to the Secretary, “shall promulgate a rulemaking that establishes performance measures and standards.” §§ 150(c)(1); see also § 101(a)(27); 49 C.F.R. § 1.85 (a). Third, States “set performance targets that reflect the measures” established by the Secretary. § 150(d). And, every two years, they submit a report to the Secretary describing (among other things) “the condition and performance of the National Highway System in the State” and the State’s “progress in achieving performance targets [it has] identified.” § 150(e). This trio together implements a system of performance benchmarks used across a number of different highway programs. One such program—and the subject of this lawsuit—is the “National Highway Performance Program.” § 119. Its purpose is “to provide support for the condition and performance of the National Highway System,” to support the “construction of new facilities on the National Highway System,” and “to ensure that investments of Federal-aid funds in highway construction are directed to support progress toward the achievement of performance targets established in an asset management plan of a State for the National Highway System.” § 119(b) (listing several other “purposes of the national highway performance program” as well). It requires state Departments of Transportation to “develop a risk-based asset management plan for the National Highway System” that will “improve or preserve the condition of the assets and the performance of the system” by setting forth strategies to make progress toward state targets for national goals. § 119(e). The specific standards and measures established by the Administrator under § 150(c)(3)(A) are used “for the purpose of carrying out section 119”—the National Highway Performance Program. If the substantive aims of these overlapping state-federal planning efforts isn’t entirely clear, that’s okay—at least for the purposes of this lawsuit. The main takeaway from this management-consultant-speak is structural: the statute defines distinct roles for state and federal officials. After states submit their plans, using the federal measures and standards promulgated under § 150(c), the Administrator reviews those plans and, in some cases, determines whether to certify them. See § 119(e)(6). If a plan doesn’t pass muster, the Administrator may specify corrective actions, deny certification, or provide at least a 90-day window in which the state may “cure the deficiencies of the plan, during which time period all penalties … of a denial of certification shall be stayed.” § 119(e)(5)–(6). If the process fails, however, § 119(e)(5)(A) authorizes the Administrator to withhold federal funds from states. See also 23 C.F.R. § 1.36 (“If the Administrator determines that a State has … failed to comply with the Federal laws or the regulations in this part with respect to a project, he may withhold payment to the State of Federal funds on account of such project, withhold approval of further projects in the State, and take such other action that he deems appropriate under the circumstances….”). III. The Administrator’s Greenhouse Gas Regulation Purporting to exercise his rulemaking authority under § 150(c)(3)(A)(ii), the Administrator in July 2022 notified the public of a forthcoming greenhouse-gas regulation. 87 Fed. Reg. 42401 , 47407. The Rulemaking Notice proposed a new measure,3 ostensibly issued under §§ 150(c)(3)(A)(ii)(IV) and (V), to “help” state Departments of Transportation assess the performance of the National Highway System.4 The Administrator proposed requiring state Departments of Transportation5 “to establish declining CO2 emissions targets to reduce CO2 emissions … that align with the Administration’s net-zero targets.” 87 Fed. Reg. at 42401. This “GHG Measure,” as the Administrator calls it, measures “the percent change in tailpipe CO2 emissions on the NHS compared to the reference year.” 23 C.F.R. § 490.507 (b). And state Departments of Transportation would be required to annually report their progress relative to that goal. The Proposed Rule aimed for a 50% reduction in greenhouse-gas emissions by 2030, relative to 2005 levels, and economy-wide net-zero emissions by 2050. Id. at 42402–03. Some of the nearly 40,000 comments submitted in the rulemaking docket lauded the proposed rule. 88 Fed. Reg. at 85371. These commenters asserted that “a GHG measure is important for understanding the long-term impact of transportation investments on GHG emissions and to better connect transportation decisions with climate goals.” Id. at 85377. Critics, however, questioned the statutory basis for the rule, the Administrator’s calculation of costs and benefits, and the practical ability of state Departments of Transportation to reduce tailpipe emissions on federal highways. Id. at 85375–79. Many took particular aim at the Proposed Rule’s declining-targets requirement: “A large number of commenters addressed the 3 The FHWA previously adopted a similar measure—sans declining-targets requirement—in 2017. See 82 Fed. Reg. 5970 , 5981 (“The rule includes a new GHG measure to assess system performance, specifically the percent change in CO2 emissions from 2017, generated by on-road mobile sources on the NHS. … All State DOTs … will be required to establish targets and report on progress.”). This rule was repealed in 2018 before it could take effect. DN 1-5. 4 The Final Rule also requires State Departments of Transportation to measure the percentage change in tailpipe emissions on the Interstate System. 88 Fed. Reg. at 85392 (“There are three performance measures to assess the performance of the Interstate System … One measure is used to measure GHG emissions.”) (to be codified at 23 C.F.R. § 490.507 ). But because Congress has defined “the National Highway System [to] include[e] the Interstate System,” nothing in this litigation appears to turn on this distinction and this opinion refers only to the National Highway System. 23 U.S.C. § 103 (a). 5 The Final Rule also requires Metropolitan Planning Organizations (MPOs) to set declining targets and measure tailpipe emissions. 88 Fed. Reg. 85364 . But no MPOs are Plaintiffs, so this opinion refers only to States. requirement to establish declining targets …. Most of these commenters asserted that a declining target is inconsistent with 23 U.S.C. 150.” Id. at 85380 . The Final Rule removed the reference to the administration’s net-zero goals, acknowledging that “the reference to net-zero targets and national GHG goals in the NPRM may have caused confusion.” Id. But otherwise the Final Rule remained largely unchanged. State Departments of Transportation must “establish declining carbon dioxide (CO2) targets for the GHG Measure”—measured as “the percent change in on-road tailpipe CO2 emissions on the NHS relative to the reference year.” Id. at 85364 . To track progress towards their targets, states also have to calculate the GHG Measure and a related metric—“the calculation of tailpipe CO2 emissions on the NHS for a given year computed in million metric tons.” Id. at 85364, 85372 . And they must report all this information—the targets, measures, and metrics—to the FHWA. Id. at 85372 (codified as 23 C.F.R. §§ 490.105 (c)(5), (e)(10), 490.507(b)). The Administrator published the Rule on December 7, 2023, and set an effective date of January 8, 2024. Id. at 85364 . Upset with the Administrator’s decision, 21 states sued him, the FHWA, the U.S. Department of Transportation, Secretary Pete Buttigieg, and President Biden in this Court on December 21, 2023. Complaint (DN 1) ¶¶ 59–63.6 The Plaintiff States alleged that the Rule exceeds the Agency’s statutory authority, exercises that authority (to the extent it exists) in an arbitrary and capricious manner, and violates the major-questions doctrine and Spending Clause. ¶¶ 113–92. They asked the Court to vacate the Rule, issue preliminary and permanent injunctive relief to prevent the Administrator from enforcing the Rule, and issue a declaration that the Rule is unlawful. Several weeks later, these States requested a preliminary injunction. Preliminary Injunction Motion (DN 28). At a status conference the Court ordered the parties to confer and consider converting the motion for preliminary injunction to cross-motions7 for summary judgment. Status Order (DN 38). The parties agreed to 6 Texas, acting alone, sued these same defendants in the Northern District of Texas on December 19, 2023. The two suits largely proceeded in parallel until that court issued its ruling on March 27, 2024. See Texas v. U.S. Dep’t of Transp., No. 5:23-cv-304, 2024 WL 1337375 (N.D. Tex. Mar. 27, 2024). 7 Defendants insisted on filing a cross-motion for summary judgment to resolve this case, noting that “[n]ormally, APA cases are resolved on cross-motions for summary judgment.” DN 81 at 1–2, n. 1 (quoting Pinnacle Armor, Inc. v. United States, 923 F. Supp.2d 1226, 1245 (E.D. Cal. 2013)).” As it oftentimes does in administrative-law challenges, the parties’ desire for dueling motions created an unnecessary dispute over elementary briefing and scheduling questions: whether simultaneous briefs will talk past each other, who goes first, who gets more pages, who gets the last word, and so forth. See generally DNs 46, 58, 61. APA cases, like most civil litigation, can be resolved on a single motion for summary judgment if filed by the movant having the burden of proof and persuasion. Usually this is clear in the administrative-law context. Why practice has developed in a different and more verbose way is far less clear—and counsel for the States and Federal Defendants couldn’t offer a reason why administrative practice has developed in this way. do so, and the Administrator further agreed to postpone any enforcement of the Rule until March 29, 2024. Scheduling Order (DN 61). The Court held a hearing on the parties’ cross-motions for summary judgment on March 7, 2024. DN 86. IV. Statutory Authority When Congress granted the agency authority to establish performance measures, did its delegation reach as far as the “environmental performance measures” the Administrator created here? Framed this way, this case poses a classic Chevron question: do the statute’s text and context, viewed in the flattering light of the agency’s expertise, rule out the Administrator’s interpretation as unreasonable? See, e.g., Loper Bright Enterprises v. Raimondo, 45 F.4th 359 , 365 (D.C. Cir. 2022), cert. granted, 143 S.Ct. 2429 (2023) (No. 22-451) (“The court’s review … is limited to the familiar questions of whether Congress has spoken clearly, and if not, whether the implementing agency’s interpretation is reasonable.”) (citing Chevron U.S.A., Inc. v. NRDC, 467 U.S. 837 , 842–43 (1984)). Squinting at the statute’s various references to sustainability, the Administrator perceives that “environmental performance” finds a natural home within the various measures used to “carry out” the National Highway Performance Program. To implement the Program, § 150(c)(3)(A)(ii)(V) supplies an express rather than implicit delegation: he must “establish … measures for States to use to assess … the performance of the National Highway System.” § 150(c)(3)(A)(ii)(V). “Congress,” moreover, “did not define the term ‘performance’ as used in § 150(c)(3),” Cross-Motion for Summary Judgment at 23, so under Chevron the agency gets to fill any interpretive gaps. That term “is most logically defined,” according to the Administrator, “by reference to the seven ‘national goals’ set forth in the immediately preceding subsection,” one of which “is ‘environmental sustainability.’” Id. at 23–24 (citing § 150(b) (“enhance[ing] the performance of the transportation system while protecting and enhancing the natural environment”)). “Read together,” these provisions “make clear that Congress intended ‘performance’ to include environmental performance in furtherance of the goal of environmental sustainability.” Id. at 24. Surrounding provisions, according to the Administrator, corroborate the reasonableness of this view: Section 119(b) lists, among the purposes of the National Highway Performance Program, “increas[ing] the resiliency of the National Highway System to mitigate the cost of damages from sea level rise, extreme weather events, flooding, wildfires, and other natural disasters.” Section 101(b)(3)(G) likewise includes, among a dozen or so “policy declarations” regarding the “transportation needs of [the] 21st Century,” that “transportation should play a significant role in … improving the environment….” See also § 134(a)(1) (noting the “national interest” includes “minimizing transportation-related fuel consumption and air pollution through … statewide transportation planning processes….”). Under the strong-form version of Chevron deference that prevailed for decades, this might have been enough. See, e.g., Sierra Club v. EPA, 793 F.3d 656, 666 (6th Cir. 2015) (“statutory context alone is sufficiently ambiguous for EPA to clear the first step of Chevron.”); Solar Energy Industries Ass’n v. FERC, 59 F.4th 1287 , 1292 (“Because Congress has not spoken to the issue, we move to step two and must defer to any reasonable agency interpretation.”). But several aspects of this question indicate this is not a moment for abandoning independent judicial judgment in favor of administrative deference. For one thing, the Administrator doesn’t rely on Chevron. It appears nowhere in his papers. Asked whether the agency forfeited deference, defense counsel stated only that the Government “ha[s]n’t invoked it.” Oral. Arg. Tr., at 57:7–13. When pressed on the United States’ position regarding whether Chevron deference could be forfeited, he replied—like Bartleby—that he wouldn’t (or perhaps couldn’t) say. See Butte Cnty., Cal. v. Hogen, 613 F.3d 190, 195 (D.C. Cir. 2010) (quoting Herman Melville, BARTLEBY, THE SCRIVENER: A STORY OF WALL STREET 10 (Dover 1990) (1853)) (the Government’s response had “all the explanatory power of the reply of Bartleby the Scrivener to his employer: ‘I would prefer not to.’”); Oral Arg. Tr. 57:2022 (“Honestly, Your Honor, I’m not sure we have a formal position other than just we haven't invoked it in this case.”). Whether waivable or not,8 the Defendants’ failure to call on Chevron surely doesn’t help their case. Even if the Defendants had sought deference, this would make no difference. Chevron’s first step requires courts to “employ the traditional tools of statutory construction” to ascertain “whether Congress has spoken to the precise question at issue,” 467 U.S. at 842–43, n.5, or at least cabined the meaning of the disputed provision, Hardin, 65 F.4th at 901–02. And here those traditional tools snip most of the threads involving environmental considerations that the Administrator stitches together in support of his interpretation. Section 150(c)(2)(C) directs the FHWA to “limit performance measures only to those described in this subsection.” “Nothing in th[at] subsection,” the Plaintiff States note, “discusses a performance measure for on-road CO2 emissions.” Id. Rather, it discusses measures for “on-road mobile source emissions.” § 150(c)(5)(B). 8 Courts and scholars debate whether parties can waive or forfeit Chevron deference. See, e.g., Amaya v. Rosen, 986 F.3d 424 , 430 n.4 (4th Cir. 2021) (highlighting the debate); James Durling & E. Garrett West, May Chevron Be Waived?, 71 STAN. L. REV. ONLINE 183, 184 (2019) (“Chevron waiver conflicts with the well-settled principle that litigants may not waive legal propositions … [and] violates basic administrative law doctrines.”); Cargill v. Garland, 57 F.4th 447 , 465 (5th Cir. 2023) (en banc) (“[T]he Chevron argument has been waived … [because] the Government’s expressed interpretation is just a legal argument, and a party waives a legal argument if it fails to raise the argument when presented with the opportunity.”). The Sixth Circuit has at times suggested it hasn’t yet answered this question, and at other times acted as if waiver were possible. Compare Hardin v. ATF, 65 F.4th 895 , 899 (6th Cir. 2023) (“We need not resolve the question of whether the government can waive the application of Chevron deference …”), with CFTC. v. Erskine, 512 F.3d 309, 314 (6th Cir. 2008) (agency “waived any reliance on Chevron deference by failing to raise it to the district court”). And that measure is to carry out a separate program—the Congestion Mitigation and Air Quality Program—which reaches only ozone, carbon monoxide, and particulate matter, not CO2. § 149(b). Taken together, the explicit reference on-road mobilesource emissions tracking of these specific substances, but not CO2, implies that Congress didn’t give the Administrator parallel authority with respect to CO2. See Antonin Scalia & Bryan A. Garner, READING LAW 107 (2012) (“The expression of one thing implies the exclusion of others.”). That negative inference is reinforced by the express directive “limit[ing] performance measures only to those described in this section.” § 150(c)(2)(C). Read as a whole, this suggests that measures of “on-road mobile source emissions” for substances not covered by § 149–such as CO2—are impermissible. And while § 119 does authorize states to spend highway funds on certain “[e]nvironmental mitigation efforts,” § 119(d)(2)(O), the Plaintiff States emphasize that most discussions of environmental mitigation in the statute “relate to natural habitats and wetlands–not CO2.” Summary Judgment Motion at 20. So the Administrator is left with only his core argument, delivered with the style if not the precedential force of Chevron, that “performance” in § 150(c) means “environmental performance.” The Plaintiff States, for their part, urge the Court not to countenance the Secretary’s interpretation, under Chevron or otherwise, based on the so-called major-questions doctrine. Under it, an agency “must point to clear congressional authorization for the power it claims” in “extraordinary cases” addressing issues of vast “economic and political significance.” West Virginia v. EPA, 597 U.S. 697 , 721–23 (2022) (quotation marks omitted). Whether climate change is a “major question” in the abstract, however, is not the right place to start. Courts must examine the specific statutory provision invoked and the regulatory authority it purportedly authorizes. In West Virginia, for example, the question the Court considered “major” was not simply “climate change.” Rather, it was “whether restricting the Nation’s overall mix of electricity generation, to transition from 38% coal to 27% coal by 2030, can be the ‘best system of emission reduction’ within the meaning of Section 111.” 597 U.S. at 720. So too in Biden v. Nebraska: the question was not “loan forgiveness” in general but whether the Secretary of Education’s specific statutory authority to “waive or modify any … provision applicable to the student financial assistance programs” gave the agency “the authority to cancel $430 billion of student loan principal.” 143 S. Ct. 2355 , 2363, 2368 (2023). Confronted as questions of policy outside the judicial ken, these might’ve been difficult assessments indeed. But as questions of statutory interpretation in which judges are expert, these challenges were readily resolvable. See id. at 2383 (Barrett, J., concurring) (“[B]y my lights, the Court arrived at the most plausible reading of the statute in these [major-question] cases.”). By beginning with two competing textual interpretations, rather than competing views of governmental policy, courts approach the “words of a statute … in their context and with a view to their place in the overall statutory scheme.” Davis v. Mich. Department of Treasury, 489 U.S. 803, 809 (1989). That assessment here makes the answer clear. The Administrator reads “performance” as “environmental performance.” But that ignores the surrounding statutory scheme. The more fundamental problem is that “each State”—not the Administrator—“shall set performance targets that reflect the measures identified in [the National Highway Performance Program].” § 150(d)(1). Structurally, even more than linguistically, therefore, the Administrator’s reading of the statute falls short of the Plaintiff States’ interpretation; it doesn’t matter whether § 150(c)(3) authorizes the Administrator’s reading of “performance … measure” if the statute doesn’t let the Administrator set “performance targets” that are reserved for state decisionmaking. Nothing in § 150(c) or (d) contemplates the Administrator appropriating that targetsetting authority by defining performance measure in a way that dictates the choice of targets given states by Congress. So the Administrator may not create a performance measure that forces states to set declining targets in CO2 tailpipe emissions on the National Highway System. Section § 150(c), which authorizes federal performance measures, says nothing about targets. The statute clearly differentiates between the two.9 The very next section, § 150(d), explains that after the FHWA has established a measure, “each State shall set performance targets ….” This language places no limits on states’ discretion in setting targets. Certainly the Administrator cannot justify this targetsetting as establishing a measure by another name; his role is to set the “standards” and “measures” and then wait to assess the states’ own use of those benchmarks in their planning. Compare §§ 119(e)(6) (federal certification of State assetmanagement plans), 150(c) (federal standard setting), with §§ 119(e)(5) (state planning), 135 (same). The concepts are not identical10—and as discussed above, see § II, the statutory scheme is built around their assignment to the separate governmental actors engaged in the planning process. Yet the Administrator does not discuss § 150(d) at all in the Final Rule. He simply asserts that state Departments of Transportation “have flexibility to set targets that are appropriate for their communities and that work for their respective climate change and other policy priorities, as long as the targets aim to reduce emissions over time.” 88 Fed. Reg. at 85364. The assertions in its briefing are similar. There, the Defendants note that “FHWA has authority to establish measures and that states must set targets that reflect those measures.” Cross-Motion for Summary Judgment at 22. But these statements say nothing about what if any statutory provision authorizes the FHWA to require states to set a specific type of target. Cf. See New York Stock Exchange LLC v. SEC, 962 F.3d 541 , 554 (D.C. Cir. 9 The agency itself recognizes this difference; it has issued a regulation defining “measures” and “targets” differently. 23 C.F.R § 490.101. 10 Indeed, an economics adage illustrates the point: “Goodhart’s Law” teaches that “[w]hen a measure becomes a target, it ceases to be a good measure.” The former concerns counting, while the latter implies control. . 2020) (“an agency cannot purport to act with the force of law without delegated authority from Congress”). It strains the English language to contend that a declining-targets requirement still affords states “flexibility to set targets … so long as the targets are declining.” 88 Fed. Reg. at 85382; see also Sur-Reply at 12 (“Declining targets are also entirely consistent with States’ authority to set their own targets [because] States retain the ability to set their own targets.”). For one thing, that statement is false. State Departments of Transportation do not have flexibility under the Final Rule. They cannot set static or increasing targets, even if “their policies and priorities” dictate such targets. Id. Instead, they must set declining targets.11 And contrary to the Defendants’ assertion that “FHWA does not mandate any specific degree of decline,” Sur-Reply at 16, the FHWA’s own reporting form requires the target “represent an anticipated decline of -0.1% or more.” Initial GHG Reporting Form at FHWA000161.12 As Plaintiffs’ counsel put it, the Final Rule evinces all the discretion of “Henry Ford’s Rule for Model T’s”— any color so long as it’s black. Reply (DN 83) at 9. For another thing, “a declining target” is not “just part of the FHWA’s decision regarding how to measure performance and set targets.” Sur-Reply at 11. Again, that begs the question. If the FHWA seeks to play a role in setting targets, it must identify a statutory provision authorizing it to do so. After all, “an agency literally has no power to act … unless and until Congress confers power upon it.” La. Public Service Commission, v. F.C.C., 476 U.S. 355, 374 (1986). The Administrator may not displace state authority and contradict the statutory division of labor by 11 The Administrator’s question-begging argument that when Congress “provided for safety, sustainability, and performance goals … they certainly did not mean worsening safety, sustainability, or performance” is beside the point. Sur-Reply at 16. Congress appears to have allowed states to hold sustainability constant while improving safety, or vice versa. Nothing in the statute says otherwise. Yet by mandating declining targets, the Administrator denies the states this option. 12 The parties and the Court litigated this case on the assumption that the Administrator’s initial GHG reporting form required state Department of Transportation to select a target that declines by at least -0.1%. On March 20, however, the Plaintiff States made a supplemental filing explaining that on January 17—while this litigation was underway—the Administrator rescinded this form an issued a new version that omitted the -0.1% requirement. Declaration of Victor Maddox (DN 93-1) at ¶6. The new reporting form states that “the target represents the anticipated decrease in on-road tailpipe CO2 emissions on NHS facilities….” Updated Initial GHG Reporting Form (DN 93-3) at 1. But, as the Defendants note, the new reporting form is not part of the administrative record. So the Court does not consider the new form. See Little Traverse Lake Property Owners Association v. National Park Service, 883 F.3d 644, 657 (6th Cir. 2018) (“The task of the reviewing court is to apply the appropriate APA standard of review, 5 U.S.C. § 706 , to the agency decision based on the record the agency presents to the reviewing court.” (quotation marks and emphasis omitted)). In any event, the new form would not have changed this Court’s decision. appropriating for himself the ability to dictate the targets states may set. “Merely because an agency has rulemaking power does not mean that it has delegated authority to adopt a particular regulation.” New York Stock Exchange LLC v. SEC, 962 F.3d 541 , 554 (D.C. Cir. 2020). Although the statute assigns the federal agency ample authority in designing these benchmarks, reviewing state planning, seeking “corrective action,” and potentially withholding federal funds, that authority doesn’t extend to displacing the state role in setting their own targets based on their own sovereign priorities. By eliding this crucial distinction, the Administrator would shift to state Departments of Transportation the responsibility—and presumably the political accountability— for choosing what level of CO2 tailpipe emissions is optimal. See Printz, 521 U.S. at 929–30. This brand of federalism could collapse the marble cake. It becomes difficult to tell where national authority ends and state “discretion” begins. “At best, the [Administrator’s interpretation] is counterintuitive;” at worst, it “comes close to violating the plain language of the statute.” Goldstein v. SEC, 451 F.3d 873, 881 (D.C. Cir. 2006). Either way, the Final Rule requiring states to set declining targets exceeds the agency’s statutory authority. Id. Under any tenable reading of the statutory text, therefore, the specific measure before the Court—declining targets for the percent change in tailpipe CO2 emissions on the NHS relative to 2022—exceeds the agency’s statutory authority. Cf. Sur-Reply at 11 (“[A] declining target is just part of FHWA’s decision regarding how to measure performance and set targets.”). Because the Administrator has identified no source of authority that authorizes him to require states to set declining targets, the Final Rule exceeds his statutory authority. V. Arbitrary & Capricious Even assuming Congress gave the Administrator authority to set environmental performance standards that embrace CO2, the Administrator exercised that authority in an arbitrary and capricious manner. “[A]gency action” must “be reasonable and reasonably explained.” FCC v. Prometheus Radio Project, 592 U.S. 414 , 423 (2021). It must seek a lawful end in a reasonably logical manner—based on “relevant data,” “a satisfactory explanation,” and “a rational connection between the facts found and the choice made.” Motor Vehicle Manufacturer’s Association of U.S., Inc. v. State Farm Mutual Automobile Insurance Co., 463 U.S. 29, 43 (1983). Here, however, the Administrator “offered an explanation for [his] decision that runs counter to the evidence before the agency.” Id. He justified the rule on two grounds: it would reduce CO2 emissions and serve as an informational tool for state policymakers. See 88 Fed. Reg. 85364 , 85371 (Dec. 7, 2023) (Final Rule); 87 Fed. Reg. 42401–02, 42407, 42410 (July 15, 2022) (Rulemaking Notice). Neither adds up. A. Emissions Reductions. In both the Rulemaking Notice and Final Rule the Administrator explained that declining state-level tailpipe emissions targets would “address the largest source of U.S. CO2 emissions” by “aid[ing] State DOTs … in planning GHG emissions reductions and evaluating progress toward national, State and local GHG goals.” 87 Fed. Reg. at 42406, 42410; see also 88 Fed. Reg. at 85369 (“[E]stablishing declining targets will help State DOTs … plan towards reductions in GHG emissions and make Federal Infrastructure investment decisions that reduce climate pollution.”). The Rulemaking Notice specifically required States to “establish declining targets that reduce CO2 emissions” to “align with the Administration’s target… of net-zero emissions, economy-wide, by 2050.” 87 Fed. Reg. at 42402. Doing so, the Administrator asserted, would “improve transportation sector GHG performance and work toward achieving net-zero emissions economy-wide by 2050.” Id. at 42403. The Final Rule echoed this rationale, but excised any mention of the net-zero goal. One benefit of the rule, however, remained “potentially significant reductions in GHG emissions….” 88 Fed. Reg. at 85389. “[E]stablishing the GHG emissions performance measure,” the Administrator insisted, would “improve the transportation sector’s GHG performance, which has lagged behind other major U.S. sectors.” Id. at 85379. But not directly. The Final Rule does “not mandat[e] any [emissions] reductions” of a particular level. Id. at 85374. Nor does it even require “that emissions actually decrease.” Id. at 85379. And the Administrator has disclaimed any enforcement or other consequences for States that fail to meet their emissions targets. Id. at 85378. So the most direct causal relationship between the Rule’s end and means (assuming emissions reduction is a statutorily authorized end) is off the table. Instead, the Administrator advocates an indirect effect on CO2 emissions: the Final Rule’s influence on state Department of Transportation decisionmaking. Requiring states to set unenforceable annual emissions targets, the Administrator maintains, would help State Departments of Transportation “plan toward reductions in GHG emissions and Federal infrastructure investment decisions that reduce climate pollution.” Id. at 85369. This is an admittedly contingent assumption: “The benefits of the final rule,” he explained, “would depend on whether and how State DOTs … shift their investment decisions as a result of the requirements of the rule.” Regulatory Impact Analysis, FHWA000207; see also FHWA000087 (same). Again, however, this is up to the states, because the Final Rule doesn’t insist on particular projects or types of projects; “FHWA reject[ed] the characterization that State[s] … are being pressured … to select any specific project based on this measure.” 88 Fed. Reg. at 85380; see also id. at 85375 (similar). “The FHWA expects—but does not require—that this measure will help State DOTs and MPOs select projects that will reduce GHG emissions.” Id. at 85375. The assumption is a head-scratcher. Why would making states calculate an unenforceable GHG Measure cause them to choose highway projects that reduce tailpipe emissions? The rulemaking and litigation record is replete with indications that states (at least these 21) don’t want to or plan to set and pursue declining CO2 targets for their federally funded highways. In response, the Administrator offers neither carrot nor stick—only his own hope. But that is not a strategy—at least not a rational one that could survive arbitrary-and-capricious review. And nothing in the record explains why step 1 (collecting state emissions targets) should induce step 2 (states reducing emissions). Rather than citing evidence, logic, or expertise, the Regulatory Impact Analysis released alongside the Rulemaking Notice instead misquoted a business aphorism: “The reasoning behind” this Rule, the agency maintained, “is that what gets measured gets improved.” Initial Regulatory Impact Analysis at FHWA00086. Often attributed (or misattributed) to Peter Drucker, the actual aphorism asserts that “what gets measured gets managed.” See Measurement Myopia, Drucker Institute, (Jul. 7, 2013) https://drucker.institute/thedx/measurement-myopia/.13 Managed by whom? In this case, by the States—not by FHWA. Some states presumably already want to reduce emissions, in which case the Final Rule is unnecessary. See 87 Fed. Reg. at 42411 (“[T]he 2018 repeal final rule identified that several States … were already tracking CO2 emissions….”) And other states (those in this lawsuit) have already rejected the notion that their policy preferences would change if only they engaged in annual benchmarking. Nothing in the record or in basic reasoning supports this informational-inducement rationale. As incentives go, an extra homework assignment is a pretty paltry one. The Administrator’s calculation of the Final Rule’s benefits—that is, its ability to reduce emissions—may have revealed his lack of conviction in this rationale. He conceded in the Rulemaking Notice and Final Rule that “FHWA is unable to quantify the benefits” of the Final Rule. 87 Fed. Reg. at 42418; 88 Fed. Reg. at 85389. At first, the Initial Regulatory Impact Analysis stated that the Rule “will lead to an enhanced performance of the NHS,” FHWA000086, while later the Final Regulatory Impact Analysis stated only that the Final Rule “could lead to an enhanced performance of the NHS because of the established GHG measure,” FHWA000207. Ultimately, however, the Administrator conceded that “[t]he potential benefits [of the Final Rule] are difficult to forecast, quantify, and monetize.” Final Regulatory Impact analysis at FHWA000207; see also 88 Fed. Reg. at 85365 (“[T]he anticipated quantitative benefits are difficult to forecast and monetize.”) Yet the Administrator still asserted that that “the benefits of this rulemaking … are substantial and justify finalizing this action.” 13 Perhaps that is why the Administrator removed this statement from the Final Regulatory Impact Analysis. Compare FHWA000086 with FHWA000207. Ironically, its initial inclusion seemingly ignores one of the primary lessons of this saying: managers can choose to measure the wrong thing, and measurement does not necessarily imply improvement. 88 Fed. Reg. at 85369.14 Without any apparent mechanism by which this Rule would achieve those benefits, this is pure ipse dixit. Although the Plaintiff States don’t challenge, and this Court doesn’t scrutinize, the Administrator’s weighing of identified costs and benefit, these evolving assessments surely support the view that the links between policy and outcomes are not the tightest. See Louisiana v. U.S. Department of Energy, 90 F.4th 461 , 473 (5th Cir. 2024) (“[T]he States argue, and DOE does not appear to dispute, that one important aspect of that problem is whether appliance regulations actually reduce energy and water consumption. Yet the administrative record contains ample evidence that DOE's efficiency standards likely do the opposite….”). The Administrator likewise gave only cursory consideration to concerns that State Departments of Transportation have limited ability to select transportation projects that will actually reduce emissions in a cost-effective way. True, the Rule identified some options, see 88 Fed. Reg. at 85377–78, but commenters asserted that these didn’t account for rural settings, low population density, and the disproportionate use of “heavy duty vehicles required for business or agriculture.”). See, e.g., id. at 85377 ; Idaho Comment at FHWA4200–01. The Administrator rejected these comments out of hand, claiming they were “predicated on the misconception that FHWA is requiring any specific behavior by State[s].” 88 Fed. Reg. at 85375. This is a complete non sequitur, however. The key question at this stage is whether the Administrator has offered a plausible reason to believe the Final Rule will cause states—such as the rural ones discussed in these comments—to take steps to reduce emissions. That the Administrator will not require emissions reductions through specific projects says nothing about whether state Departments of Transportation will be able to achieve emissions reductions through their independently chosen projects. Such “conclusory statements … do not constitute adequate agency consideration of an important aspect of the problem.” (quotation marks omitted)). Louisiana, 90 F.4th at 473. Lacking any rational connection between the Administrator’s non-enforcement, state unwillingness to voluntarily reduce emissions, and the Final Rule’s benchmarking process, the Administrator has failed to establish “a rational connection between the facts found and the choice made.” State Farm, 463 U.S. at 43 ; see also Louisiana, 90 F.4th at 475. 14 The break-even analysis does not help the Administrator. That analysis “calculates … the level that benefits must achieve in order to equal costs while holding every other variable in the analysis constant.” Final Regulatory Impact Analysis at FHFWA000208 (cleaned up). The Administrator concluded from that analysis that only small reductions in CO2 emissions would be necessary for the benefits of the Rule to exceed the costs (of compliance). Id. at FHWA000211–12. But that “does not quantify the direct benefits achieved by this final rule,” id. at FHWA000208, benefits that (as discussed above) are far from certain to materialize. That only minor reductions in CO2 might, in the Administrator’s eyes, cause the benefits of the Rule to equal the costs is beside the point, as there is no reason to believe that those reductions will occur at all. B. Informational Tool. The Administrator’s second rationale—that the Final Rule is an informational tool, fares no better. “Measuring and reporting complete, consistent, and timely information for onroad mobile source emissions,” he explained, “is necessary so that all levels of government and the public can monitor changes on GHG emissions over time and make more informed decisions about the role of transportation investments and other strategies in achieving GHG reductions.” 88 Fed. Reg. at 85371; 87 Fed. Reg. at 42403. As the Defendants put it in their briefing, the Final Rule is nothing more “than a data-collection and reporting rule.” Sur-Reply at 4.15 But the Final Rule is not a data-collection tool. First of all, data collection has nothing to do with the Rule’s core requirement: declining emissions targets. That reflects an aspirational future planning process, not a descriptive past reporting process. When the Defendants’ briefs discuss emissions reporting, they do so in furtherance of the declining-target requirement: states must report their “progress” toward the emissions-reduction goal. Cross-Motion for Summary Judgment at 6 (“Under the rule ... states would be required to establish … emissions reduction targets … [and] report on [their] progress meeting the targets.”); Sur-Reply at 7 (The Final Rule “merely requires states to set targets and submit reports.”). Second, this purported goal is entirely illusory. The FHWA already has the factual information needed to calculate the percentage change in annual emissions attributable to the National Highway System (that is, the “GHG Measure”). Indeed, the efficiency of using existing numbers was originally described as a justification for this measure: “The FHWA has retained the GHG performance measure proposed in the NPRM … because of its … reliance on data States already report to FHWA.” 88 Fed. Reg. at 85377. The GHG Measure is calculated using three types of data: fuel volume, emissions factors, and vehicle miles traveled (VMT). Final Regulatory Impact Analysis at FHWA000180. The calculation entails “multiplying motor fuel sales volumes … by emissions factors for CO2 per gallon of fuel, and the percentage of VMT on the NHS.” 87 Fed. Reg. at 42416. First, the fuel volume data is “already reported by State[s] … to FHWA” through the Fuels and Financial Analysis System (Fuels/FASH). Id; see also 88 Fed. Reg. at 85386 (“The FHWA’s Fuels & FASH database will serve as the source of fuel use data since it is a national, established, and validated source of fuel use information as reported by the States.”). Second, the agency supplies the emissions factors itself, so it doesn’t need to collect anything that 15 The Final Rule advanced both the emissions-reduction rationale and the informational rationale. See 88 Fed. Reg. at 85389 (“[B]enefits [of the Rule] include potentially significant reductions in GHG emissions … and better information on the impact of transportation decisions on GHG emissions.”). In this litigation, however, the Defendants have relied primarily on the informational rationale. See, e.g., Sur-Reply at 5 (“In other words, the Final Rule is a data-collecting and reporting device designed to improve transportation decisionmaking in a transparent manner.”). the states might report. FHWA000180; see also 88 Fed. Reg. at 85385 (“FHWA will publish uniform CO2 emissions factors for each fuel type to be used by all states in calculated the … metric for the GHG measure.”) And third, the VMT Data comes “from the FHWA’s Highway Performance Monitoring System (HPMS).” FHWA000180; see also FHWA000186 (similar); 88 Fed. Reg. at 85372 (“The VMT data used by State[s] … will represent the prior calendar year and should be consistent with the final VMT data submitted by the State[s] … to HPMS.”). So the agency is not “collecting” anything from the states. It already has the information necessary to calculate the GHG measure. What, then, is the point of foisting onto the states the requirement to calculate and report the GHG measure? The Administrator never says. If the Administrator, rather than recalcitrant states, seeks to disseminate information about each state’s on-road tailpipe CO2 emissions, why wouldn’t he simply perform the calculations and publish the results. That wouldn’t require a new rulemaking and state-level enforcement. Yet it would “demonstrate Federal leadership in the assessment and disclosure of climate pollution from the transportation sector,” 88 Fed. Reg. at 85371, an aim the Administrator seeks to pursue.16 More to the point, it would avoid the political-accountability concerns that sometimes plague cooperative-federalism programs. The Supreme Court has warned that “where the Federal Government directs the states to regulate, it may be state officials who will bear the brunt of public disapproval, while the federal officials who devised the regulatory program may remain insulated from the electoral ramifications of their decision.” New York v. United States, 505 U.S. 144, 169 (1992). This is true even where the regulatory program in question forces states officers to perform only “ministerial tasks.” Printz, 521 U.S. at 929 . When the federal government forces state governments to 16 The Defendants did not argue that the GHG Measure is justifiable separate and apart from the declining-targets requirement. The two are intertwined: states would calculate and report the GHG measure to signify to the agency their progress toward the annual decliningemissions target. Cross-Motion for Summary Judgment at 6 (“Under the rule ... states would be required to establish … emissions reduction targets … [and] report on [their] progress meeting the targets.”). Indeed, the Defendants explained that “a declining target is … the very purpose of performance management.” Sur-Reply at 11–12. True, the Final Rule included a severability clause. See 88 Fed. Reg. at 85373 (to be codified at 23 C.F.R. § 490.515 ). But it makes no difference here. Even if the Court considered the declining-targets requirement and the GHG Measure separately, both are invalid. As discussed above, the declining-targets requirement exceeds the agency’s statutory authority. And whether considered in tandem or isolation, the GHG Measure is arbitrary and capricious because the Administrator’s two justifications are unsupported by a rational explanation and run counter to the evidence before the agency. See, e.g., North Carolina v. EPA, 531 F.3d 896 , 929–30 (D.C. Cir. 2008) (vacating entire rule when the agency’s analysis was “fundamentally flawed” and “no amount of … revising of the explanations” would result in an “acceptable rule”). implement a federal program, it foists the blame for the program’s “burdensomeness” and “defects” onto the States. Id. at 930. The Final Rule seems to neglect these risks—if it’s not affirmatively embracing the shifting of accountability for climate-change policy. Why require state Departments of Transportation to calculate (without penalty) a measure they have no interest in, using data the Administrator already has, in the faint hope that they will take action that they disclaim? True, Printz left open the possibility that federal programs “which require only the provision of information to the Federal Government” might be permissible. Id. at 918. Certainly nothing in the Court’s caselaw suggests “data monitoring rules,” Cross-Motion for Summary Judgment at 27, are somehow insulated from arbitrary-and-capriciousness review. And in any event, as described above, it’s hard to describe the Final Rule as a “mere reporting requiremen[t].” But it’s relatively easy to perceive the accountability concerns the Court has found problematic. Though the Court need not resolve this case on Spending Clause grounds to reject the Final Rule’s lawfulness, the legitimate concerns raised by the Plaintiff States make clear why the Administrator cannot rely on an innocuous “data reporting” rationale to save an otherwise arbitrary and capricious rule. VI. Remedy In their complaint the Plaintiffs asked the Court to both “[v]acate and set aside the Final Rule in its entirety” and “[i]ssue permanent injunctive relief prohibiting the Agencies from implementing, applying, enforcing, or otherwise proceeding on the basis of the Final Rule.” Complaint at 62. They also sought declaratory relief. Id. During the summary-judgment hearing, however, the Plaintiffs sought relief limited to the challenger states—not vacatur of the Rule in its entirety. “[N]ationwide relief is not something we’re asking for …. The states that have asked for the relief [would] be the appropriate extent of an injunction or an order invalidating the rule.” Oral Arg. Tr. at 62:4–5. Defendants, meanwhile, asked the Court to “follow Justice Gorsuch’s Texas concurrence and decline to enter a nationwide order of vacatur.” Cross-Motion for Summary Judgment at 33 (citing United States v. Texas, 599 U.S. 670, 695 (Gorsuch, J., concurring in the judgment)). Must a court, anytime it rules an agency rule unlawful, vacate the entire rule? Even if the challengers seek narrower relief? That would seem counterintuitive—at a minimum—and likely at odds with traditional notions of remedial discretion and minimalism: typically, “relief should be no greater than necessary to protect the rights of the prevailing litigants.” Doe v. Rokita, 54 F.4th 518 , 519 (7th Cir. 2022); see also United States v. Miami University, 294 F.3d 797, 816 (6th Cir. 2002) (“If injunctive relief is proper, it should be no broader than necessary to remedy the harm at issue.”). Vacatur, after all, operates on the rule itself—“‘strik[ing] down’ an agency’s work” so “the disapproved agency action is treated as though it had never happened.” Jonathan F. Mitchell, The Writ-of-Erasure Fallacy, 104 VA. L. REV. 933, 1013 (2018). It “prevents [the rule’s] application to all those who would otherwise be subject to its operation.” East Bay Sanctuary Covenant v. Garland, 994 F.3d 962 , 987 (9th Cir. 2020) (emphasis added). Many courts nevertheless consider it “the normal remedy under the APA.” Long Island Power Association v. FERC, 27 F.4th 705 , 717 (D.C. Cir. 2022). Normal, but not automatic. “Although vacatur is the presumptive remedy for a violation of the Administrative Procedure Act, courts have discretion to craft other remedies.” Johnson v. U.S. Office of Personnel Management, 783 F.3d 655, 663 (7th Cir. 2015). The Sixth Circuit recently acknowledged that “[r]eviewing courts certainly have the power to vacate an agency action they find unlawful.” Sierra Club v. EPA, 60 F.4th 1008 , 1021 (6th Cir. 2023). And in several opinions it has stated, seemingly in passing and without confronting the vacatur-versus-injunction question, that courts “shall” or “must” set aside agency action that is unlawful.17 But these statements cannot literally mean that a court must vacate every agency action that is arbitrary and capricious or unlawful, for the Sixth Circuit also recognizes that remand without vacatur may be an appropriate remedy in some cases. See Sierra Club, 60 F.4th at 1022–23. Even the judges of the D.C. Circuit, hardly a bashful lot when it comes to the exercise of administrative-law remedies, have acknowledged that “vacatur need not be the remedy for an invalidly adopted rule.” Heartland Regional Medical Center v. Sebelius, 566 F.3d 193, 199 (D.C. Cir. 2009). And perceptive judges have explained that remedial discretion may retain an important role even if vacatur is compelled by text and precedent. See, e.g., Honeywell Int’l Inc. v. EPA, 374 F.3d 1363 , 1375 (D.C. Cir. 2005) (Randolph, J., concurring) (stay motion offers a “safety valve” if unlawful rule must be vacated). Why wouldn’t an injunction represent a more tailored and appropriate remedy? Judges may enjoin federal officers from executing unlawful regulations. See, e.g., Franklin v. Massachusetts, 505 U.S. 788, 802 (1992) (“[I]injunctive relief against executive officials like the Secretary of Commerce is within the courts’ power….”). And in so doing they may exercise their remedial discretion to fashion equitable remedies to particular parties and applications. “[C]laims for equitable relief … appeal to the ‘remedial discretion’ of the courts.” Winzler v. Toyota Motor Sales U.S.A., 681 F.3d 1208, 1210 (10th Cir. 2012) (Gorsuch, J.). And “[t]he 17 See, e.g., Simms v. National Highway Traffic Safety Admin., 45 F.3d 999 , 1003 (6th Cir. 1995) (“We review the administrative record, applying the standards set forth in section 706 of the APA … and must set aside agency action that is ‘arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law.”); Kentucky Waterways Alliance v. Johnson, 540 F.3d 466, 473 (6th Cir. 2008) (“The APA directs that when reviewing the decision of an administrative agency, a court shall ‘hold unlawful and set aside the agency action’ if the action is ‘arbitrary, capricious, an abuse of discretion or otherwise not in accordance with law.’”); Mann Construction, Inc. v. United States, 27 F.4th 1138 , 1148 (6th Cir. 2022) (“Because the IRS's process for issuing Notice 2007-83 did not satisfy the noticeand-comment procedures for promulgating legislative rules under the APA, we must set it aside.”). chancellor’s decision is based on all the circumstances; his discretion is necessarily broad and a strong showing of abuse must be made to reverse it.” United States v. W. T. Grant Co., 345 U.S. 629, 633 (1953). Despite this greater flexibility, injunctive relief is often treated as the more extreme remedy, even in the APA context. The Supreme Court has characterized an injunction as “a drastic and extraordinary remedy” as compared to the “less drastic” remedy of vacatur. Monsanto v. Geertson Seed Farms, 561 U.S. 139 , 165–66 (2010). In what way? Vacatur is susceptible to many of the by-now-familiar critiques of nationwide injunctions: binding nonparties, preventing “percolation” of difficult legal questions, and triggering sprints to the courthouse (sometimes a particular courthouse). Is “[t]his helter-skelter behavior” really “what our legal system was designed to encourage”? See CASA de Md., Inc. v. Trump, 971 F.3d 220 , 260 (4th Cir. 2020). A party-specific injunction seems tailor-made to fit relief—and the judicial role—to their proper size.18 This apparent mismatch flows from the notion that injunctive relief is somehow more extreme, or “drastic,” because it may bind a particular official or freeze an administrative process—and does so backed (at least theoretically) by a court’s contempt power. A court cannot award injunctive relief, even in an APA case, unless the prevailing party establishes that the traditional four-factor test—irreparable harm, inadequate remedies at law, balance of hardships, and public interest—is satisfied. See Monsanto, 561 U.S. at 157 (2010) (“An injunction should issue only if the traditional four-factor test is satisfied.”); eBay Inc. v. MercExchange, L.L.C., 547 U.S. 388 , 391 (2006). And the D.C. Circuit, at least, has held that “[s]uccess on an APA claim does not automatically entitle the prevailing party to a permanent injunction.” In re Fed. Bureau of Prisons' Execution Protocol Cases, 980 F.3d 123 , 137 (D.C. Cir. 2020). This makes sense in cases like Monsanto, in which an injunction went far beyond merely setting aside a rule. The district court did not merely enjoin the effectiveness of (i.e., vacate) a rule while leaving officials to respond as they saw fit, but also ordered an agency “not to act on [a petition] in whole or in part” until it completed a court-ordered environmental assessment, “enjoined almost all future” use of the product at issue, carved out an exception to its own prohibition for some producers, and literally “replaced [the agency’s] decision with the terms of the [court’s] judgment itself.” Monsanto, 561 U.S. at 144, 147 . Compared to this remedial bonanza, “partial or complete vacatur” was indeed “a less drastic remedy” and “sufficient to redress respondents’ injury.” Id. at 165–66. 18 That seems particularly true here given how the Final Rule functions. None of the Plaintiff States are affected in any way if other States comply with the Final Rule. Uniformly binding fifty-one separate sovereigns—many of which are non-parties—in the context of a statute and regulation that both speak in terms of autonomous state decisionmaking threatens not only the Case-or-Controversy requirement, but also basic principles of federalism. But an injunction is not always “drastic and extraordinary”; sometimes it will reflect the same judicial restraint the Supreme Court found lacking in Monsanto. The Court’s opinion in that case, however, seems to have been taken rather literally and mechanistically by some lower courts. See, e.g., Alliance for Hippocratic Medicine v. FDA, 78 F.4th 210 , 254 (5th Cir. 2023); League of Wilderness Defenders/Blue Mountains Biodiversity Project v. Pena, 3:12-cv-2271, 2015 WL 1567444 , at *2 (D. Or. Apr. 6, 2015). Assuming that vacatur is always less intrusive runs counter to practically all received remedial wisdom; the very precedent Justice Alito cited for the “extraordinary” nature of injunctions described, in the immediately preceding sentence, that “[f]lexibility rather than rigidity has distinguished” equity jurisdiction. Weinberger v. Romero-Barcelo, 456 U.S. 305, 312 (1982) (“The essence of equity jurisdiction has been the power of the Chancellor to do equity and to mould each decree to the necessities of the particular case.”). This precedent shouldn’t be overread to deny that vacatur is in practice often the more drastic remedy: “Unlike an injunction, which merely blocks enforcement, vacatur unwinds the challenged agency action.” Driftless Area Land Conservancy v. Valcq, 16 F.4th 508 , 522 (7th Cir. 2021). This places district courts in a bind when the parties agree that “universal” relief is unnecessary yet do not address the traditional four-factor test for injunctive relief. That is the situation here. The Plaintiff States’ motion for summary judgment neglects to address the test for injunctive relief at all. While their motion for a preliminary injunction does address those related factors, see DN 28-1, it primarily focuses on the likelihood of success on the merits, giving only cursory treatment to the remaining considerations—which are far more important when seeking a permanent rather than preliminary injunction. Success on the merits is no longer part of the test for permanent injunctive relief; it is a prerequisite. See Oglala Sioux Tribe v. C&W Enterprises, Inc., 542 F.3d 224, 229 (8th Cir. 2008) (“If a court finds actual success on the merits, it then considers the following factors in deciding whether to grant a permanent injunction ….”) (emphasis added). And a court must make the “specific findings concerning each of the four factors” required to grant injunctive relief. Six Clinics Holding Corp. II v. Cafcomp Systems, Inc., 119 F.3d 393, 399 (6th Cir. 1997); see also FED. R. CIV. P. 65(d); Winter v. NRDC, 555 U.S. 7, 24 (2008) (“A[n] … injunction is an extraordinary remedy never awarded as of right.”). Issuing an injunction despite the Plaintiffs States’ failure to address these factors in their motion for summary judgment would be inconsistent with a limited judicial role and shift responsibility for establishing injunctive relief from the plaintiff to the judge. The Defendants’ briefing is likewise of little help. It does not address the fourfactor test for injunctive relief or chart a path toward limited injunctive relief short of that showing. Instead, it simply states that the Court should adopt Justice Gorsuch’s approach in Texas and decline to vacate the rule nationwide—never grappling with the definitional truism that vacatur is necessarily nationwide. And, unlike Justice Gorsuch, this Court is bound by vertical stare decisis. See Texas, 599 U.S. at 701 (“At the end of the day, the States fall back on … lower court decisions…. But … they do not bind us.”). Precedent casts doubt on the Court’s ability to issue an injunction absent any discussion by the parties about whether it is appropriate here. The Supreme Court has stated clearly (though not in a case presenting an injunction as less drastic than vacatur) that courts may not assume that an injunction is “generally the appropriate” remedy for unlawful agency action. Monsanto, 561 U.S. at 157 . “It is not enough for a court considering a request for injunctive relief to ask whether there is a good reason why an injunction should not issue; rather, a court must determine that an injunction should issue under the traditional four-factor test set out above.” Id. at 158 (emphasis in original). Vacating the rule, on the other hand, would exceed the parties’ requested relief, transgress the “bedrock practice” of courts to issue only “case-by-case judgments with respect to the parties” before them, and “sweep up nonparties who may not wish to receive the benefit of the court’s decision.” Arizona v. Biden, 40 F.4th 375 , 396 (6th Cir. 2022) (Sutton, C.J., concurring); Texas, 599 U.S. at 703 (Gorsuch, J., concurring).19 So the Court will instead issue declaratory relief—which the Plaintiff States sought in their complaint—and invite additional briefing on the propriety of injunctive relief. The Declaratory Judgment Act allows the Court to “declare the rights and other legal relations of any interested party seeking such declaration[.]” 28 U.S.C. § 2201 (a). This provides “an alternative remedy” to an injunction. Safety Specialty Insurance. Co. v. Genesee County Board of Commissioners, 53 F.4th 1014 , 1021 (6th Cir. 2022) (cleaned up). Importantly, “a declaratory judgment binds the parties, but only the parties[.]” Skyworks, Ltd. v. CDC 542 F. Supp. 3d 719 , 728 (N.D. Ohio 2021). The Act authorizes the Court to “declare the rights … of any interested party.” § 2201(a) (emphasis added). On its face, this statutory language binds the parties alone—unlike vacatur or a nationwide injunction. A declaratory judgment, similar to an injunction, requires some measure of judicial discretion (which the Defendants don’t resist here). See Grand Trunk W. R.R. Co. v. Consol. Rail Corp., 746 F.2d 323 , 19 Indeed, for these reasons (and others) some judges and scholars have begun to question whether vacatur is in fact authorized under the APA. See Texas, 599 U.S. at 695–99 (Gorsuch, J., concurring). John Harrison, for example, has argued that as a matter of historical practice, “vacatu[r] was unknown when the APA was adopted and for at least two decades afterwards.” John Harrison, Vacatur of Rules under the Administrative Procedure Act, 40 YALE J. ON REG. BULL. 119, 120 (2023). On this view, the statutory provision authorizing vacatur—5 U.S.C. §706(2)—does not speak in terms of remedies at all. Its directive to “set aside” agency action instead instructs courts “not to decide in accordance with the agency action.” John Harrison, Section 706 of the Administrative Procedure Act Does Not Call for Universal Injunctions or Other Universal Remedies, 37 YALE J. ON REG. BULL. 37, 46 (2020). Viewed this way, “§ 706(2) uses ‘set aside’ to mean ‘disregard’ rather than ‘vacate.’” Texas, 599 U.S. at 696 (Gorsuch, J., concurring). 326 (6th Cir. 1984) (listing considerations relevant to discretionary declaratory remedy). But unlike injunctive relief, a declaratory judgment is not an “extraordinary remedy”; it may be granted “even though [a court] chooses not to issue an injunction or mandamus.” Powell v. McCormack, 395 U.S. 486, 499 (1969). True, “[a] declaratory judgment is simply a statement of rights, not a binding order supplemented by continuing sanctions.” Steffel v. Thompson, 415 U.S. 452, 482 (1974) (Rehnquist, J., concurring). But it still “has the force and effect of a final judgment.” § 2201(a). And, as the Supreme Court has noted, “declaratory relief alone has virtually the same practical impact as a formal injunction would.” Samuels v. Mackell, 401 U.S. 66, 72 (1971). Just without “the strong medicine of the injunction.” Winter, 555 U.S. at 33 (quoting Steffel, 415 U.S. at 466 ). This remedy can secure the rights of the challenger without purporting to freeze and dictate all manner of actions by the agency, as the district court did in Monsanto. An agency could revisit the rulemaking, engage in emergency or temporary actions, avoid nonmutual estoppel, and consider intercircuit nonacquiescence—all historically familiar steps halted by a nationwide injunction or its cousin, vacatur. See CASA de Maryland, 971 F.3d at 260–61. Correspondingly, a declaratory judgment entitles the Plaintiff States to additional relief, if they so choose. “Further necessary or proper relief based on a declaratory judgment or decree may be granted, after reasonable notice and hearing, against any adverse party whose rights have been determined by such judgment.” § 2202; see also Powell, 395 U.S. at 499 (“A declaratory judgment can … be used as a predicate to further relief, including an injunction.”). Should the Administrator (or any of the other Defendants) “thum[b] his nose at the declaration” the Plaintiff States can file a motion for further relief under § 2202. Bench Billboard Co. v. City of Covington, No. 6-cv-75, 2012 WL 2919158 , at *4 (E.D. Ky. July 17, 2012). Such a motion is not constrained by the limits of Rule 59(e). Continental Casualty Co. v. Indian Head Industries, 941 F.3d 828 , 833 (6th Cir. 2019). In these circumstances, declaratory relief is the remedy most consistent with rule-of-law norms of notice, the right to be heard, and reasoned decisionmaking based on the record before the Court—participatory aspects of our judicial system that render the “power to decide” more constrained and therefore more credible. Whatever its label, “universal relief, whether by way of injunction or vacatur, strains our separation of powers.” Texas, 599 U.S. at 703 . * The litigation over this Rule encapsulates the problem with universal remedies. Shortly before this opinion issued, another district court adjudicating a challenge to the Rule brought by Texas alone issued its opinion. Texas v. DOT, 5:23cv-304, 2024 WL 1337375 (N.D. Tex. Mar. 27, 2024). Despite adjudicating a suit brought by a single state, and despite candidly recognizing the infirmities of universal relief, the judge quite reasonably thought himself constrained by Fifth Circuit precedent to vacate the Rule nationwide. Id., slip op. at 44–45. Thanks to the vacatur remedy, this functioned as a literal race to the courthouse—or perhaps out of it. Amidst concerns about the pace and percolation of judicial decisionmaking on the socalled emergency docket, when real-world events often accelerate adjudication, judges should think twice about a system that spurs additional time pressure of our own making. And depending on how one conceives of the reach of a single district judge’s order, the vacatur remedy swept in 49 non-party sovereign states—28 of which have not challenged the Rule and many of which may already be complying (voluntarily or not) with its requirements. The consequences could have been worse, still. If the two opinions disagreed, could one have required States to comply with a vacated Rule? Cf. Ohio v. EPA, 969 F.3d 306 , 308-09 (6th Cir. 2020) (state plaintiff seeking preliminary injunction of repealed rule to avoid “reasonable possibility” that rule would take effect again if another court enjoined the repeal nationwide). Judges across the country have catalogued the problems with nationwide injunctions. To date, few have recognized that vacatur suffers from the same problems. See Texas, 599 U.S. at 693-703 (Gorsuch, J., concurring). Fewer still have given attention to district courts’ ability to avoid these snares while still providing meaningful relief to the parties before them. One way or another, this unstable situation may soon add to the list of remedial challenges in American public law. ORDER The Court grants the Plaintiff States’ motion to withdraw their preliminary injunction (DN 69) and grants their motion for summary judgment (DN 72). The Court denies, without prejudice, the Plaintiff States’ requests to vacate the Final Rule or permanently enjoin its enforcement (DN 1) and instead grants the Plaintiffs’ request for declaratory relief. Given this holding, the Court denies the Defendants’ cross-motion for summary judgment and partial motion to dismiss for lack of jurisdiction and improper venue (DN 81). The Court also grants the motion for leave to file an amicus curiae brief by the American Road and Transportation Builders Association and Associated General Contractors of America, Inc. (DN 78). Last, the Court grants the Plaintiff States’ motion for leave to supplement the record (DN 93). The Court declares that the Final Rule exceeds the Federal Highway Administration’s statutory authority and is arbitrary and capricious. What additional relief, if any, might be necessary remains unclear— particularly given the intervening event of the Northern District of Texas’s vacatur order. See Texas v. DOT, 5:23-cv-304, 2024 WL 1337375 (N.D. Tex. March 27, 2024); 28 U.S.C. § 2202 . The Court invites the parties to confer and file, within 21 days, supplemental briefs on the need for additional remedial action by this Court. Those briefs may address how the traditional four-factor test for injunctive relief applies in these circumstances and on this record. Ve April 1, 2024