JENNIFER WALKER ELROD, Circuit Judge:

The Patient Protection and Affordable Care Act (the Act or ACA) is a monumental piece of healthcare legislation that regulates a huge swath of the nation's economy and affects the healthcare decisions of millions of Americans. The law has been a focal point of our country's political debate since it was passed nearly a decade ago. Some say that the Act is a much-needed solution to the problem of increasing healthcare costs and lack of healthcare availability. Many of the amici in this case, for example, argue that the law has extensively benefitted everyone from children to senior citizens to local governments to small businesses. Others say that the Act is a costly exercise in burdensome governmental regulation that deprives people of economic liberty. Amici of this perspective argue, for example, that the Act "has deprived patients nationwide of a competitive market for affordable high-deductible health insurance," leaving "patients with no alternative to ... skyrocketing premiums." Association of American Physicians & Surgeons Amicus Br. at 15.

None of these policy issues are before the court. And for good reason—the courts are not institutionally equipped to address them. These issues are far better left to the other two branches of government. The questions before the court are far narrower: questions of law, not of policy. Those questions are: First, is there a live case or controversy before us even though the federal defendants have conceded many aspects of the dispute; and, relatedly, do the intervenor-defendant states and the U.S. House of Representatives have standing to appeal? Second, do the plaintiffs have standing? Third, if they do, is the individual mandate unconstitutional? Fourth, if it is, how much of the rest of the Act is inseverable from the individual mandate?

We answer those questions as follows: First, there is a live case or controversy because the intervenor-defendant states have standing to appeal and, even if they did not, there remains a live case or controversy between the plaintiffs and the federal defendants. Second, the plaintiffs have Article III standing to bring this challenge to the ACA; the individual mandate injures both the individual plaintiffs, by requiring them to buy insurance that they do not want, and the state plaintiffs, by increasing their costs of complying with the reporting requirements that accompany the individual mandate. Third, the individual mandate is unconstitutional because it can no longer be read as a tax, and there is no other constitutional provision that justifies this exercise of congressional power. Fourth, on the severability question, we remand to the district court to provide additional analysis of the provisions of the ACA as they currently exist.

I.

On March 23, 2010, President Barack Obama signed the ACA into law. See Patient Protection and Affordable Care Act, Pub. L. No. 111-148, 124 Stat. 119 (2010). The Act sought to "increase the number of Americans covered by health insurance and decrease the cost of health care" through several key reforms. See Nat'l Fed'n of Indep. Bus. v. Sebelius (NFIB), 567 U.S. 519, 538, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012).

Some of those reforms implemented new consumer protections, aiming primarily to protect people with preexisting conditions. For example, the law prohibits insurers from refusing to cover preexisting conditions. 42 U.S.C. § 300gg-3. The "guaranteed-issue requirement" forbids insurers from turning customers away because of their health. See 42 U.S.C. §§ 300gg, 300gg-1. The "community-rating requirement" keeps insurers from charging people more because of their preexisting health issues. 42 U.S.C. § 300gg-4.1 The law also requires insurers to provide coverage for certain types of care, including women's and children's preventative care. 42 U.S.C. § 300gg-13(a)(3)-(4).2

Other reforms sought to lower the cost of health insurance by using both policy "carrots" and "sticks."3 On the stick side, the individual mandate—which plaintiffs challenge in the instant case—requires individuals to "maintain [health insurance] coverage." 26 U.S.C. § 5000A(a). If individuals do not maintain this coverage, they must make a payment to the IRS called a "shared responsibility payment."4 Id.; see also King v. Burwell, ___ U.S. ___, 135 S.Ct. 2480, 2486, 192 L.Ed.2d 483 (2015).

The individual mandate was designed to lower insurance premiums by broadening the insurance pool. See 42 U.S.C. § 18091(2)(J) ("By significantly increasing... the size of purchasing pools, ... the [individual mandate] will significantly ... lower health insurance premiums."). When the young and healthy must buy insurance, the insurance pool faces less risk, which, at least in theory, leads to lower premiums for everyone. See 42 U.S.C. § 18091(2)(I) (positing that the individual mandate will "broaden the health insurance risk pool to include healthy individuals, which will lower health insurance premiums"). The individual mandate thus serves as a counterweight to the ACA's protections for preexisting conditions, which push riskier, costlier individuals into the insurance pool. Under the protections for consumers with preexisting conditions, if there were no individual mandate, there would arguably be an "adverse selection" problem: "many individuals would," in theory, "wait to purchase health insurance until they needed care." Id.5

The Act also sought to lower insurance costs for some consumers through policy "carrots," providing tax credits to offset the cost of insurance to those with incomes under 400 percent of the federal poverty line. See 26 U.S.C. § 36B; 42 U.S.C. §§ 18081, 18082. The Act also created government-run, taxpayer-funded health insurance marketplaces—known as "Exchanges"—which allow customers "to compare and purchase insurance plans." King, 135 S. Ct. at 2485; see also 42 U.S.C. § 18031. Opponents of the law argue that the law has led to unintended subsidies to keep plans afloat and insurance companies in the black. Texas points in its brief, for example, to a Congressional Budget Office study estimating that federal outlays for health insurance subsidies and related spending will rise by about 60 percent over the next ten years, from $58 billion in 2018 to $91 billion by 2028. CBO, The Budget and Economic Outlook: 2018 to 2028 at 51 (April 2018), available at https://tinyurl.com/CBOBudgetEconOutlook-2018-2028; State Plaintiffs' Br. at 13-14.

The ACA also enlarged the class of people eligible for Medicaid to include childless adults with incomes up to 133 percent of the federal poverty line. 42 U.S.C. §§ 1396a(a)(10)(A)(i)(VII), 1396a(e)(14)(I)(i); NFIB, 567 U.S. at 541-42, 132 S.Ct. 2566. The ACA originally required each state to expand its Medicaid program or risk losing "all of its federal Medicaid funds." NFIB, 567 U.S. at 542, 132 S.Ct. 2566. In NFIB, however, the Supreme Court held that this exceeded Congress' powers under the Spending Clause. Id. at 585, 132 S.Ct. 2566 (plurality opinion). But the Court allowed those states that wanted to accept Medicaid expansion funds to do so. See id. at 585-86, 132 S.Ct. 2566 (plurality opinion); id. at 645-46, 132 S.Ct. 2566 (Ginsburg, J., concurring in part, concurring in the judgment in part, and dissenting in part). As a result, the states that have not participated in the expansion now subsidize, through their general tax dollars, the states that have participated in expansion.

Since the Act was passed, its opponents have attempted to attack it both through congressional amendment and through litigation. Between 2010 and 2016, Congress considered several bills to repeal, defund, delay, or amend the ACA. See Intervenor-Defendant States' Br. at 10. Except for a few modest changes, these efforts were closely fought but ultimately failed. Intervenor-Defendant States' Br. at 10-11. In 2017, the shift in presidential administrations reinvigorated opposition to the law, but many of these later legislative efforts failed as well. In March 2017, House leaders pulled a bill that would have repealed many of the ACA's essential provisions. In July 2017, the Senate voted on three separate bills that similarly would have repealed major provisions of the Act, but each vote failed.6 Finally, in September 2017, several Senators introduced another bill that would have repealed some of the ACA's most significant provisions, but Senate leaders ultimately chose not to bring it to the floor for a vote. Intervenor-Defendant States' Br. at 11.

The ACA's opponents also took their cause to the courts in a series of lawsuits, some of which reached the Supreme Court. Particularly relevant here, the Court, in NFIB, upheld the law's individual mandate. 567 U.S. at 574, 132 S.Ct. 2566. Through fractured voting and shifting majorities—explained in more detail in Part V of this opinion—the Court decided that the ACA's individual mandate could be read as a tax on an individual's decision not to purchase insurance, which was a constitutional exercise of Congress' taxing powers under Article I of the U.S. Constitution. Id.; U.S. Const. art. I, § 8, cl. 1. The Court favored this tax interpretation to save the provision from unconstitutionality. Reading the provision as a standalone command to purchase insurance would have rendered it unconstitutional. This reading could not have been justified under the Commerce Clause because it would have done more than "regulate commerce ... among the several states." U.S. Const. art. I, § 8, cl. 3. It would have compelled individuals to enter commerce in the first place.7 NFIB, 567 U.S. at 557-58, 132 S.Ct. 2566. The Court also held that the provision could not be justified under the Constitution's Necessary and Proper Clause. Id. at 561, 132 S.Ct. 2566 (Roberts, C.J.); id. at 654-55, 132 S.Ct. 2566 (Scalia, Kennedy, Thomas, and Alito, JJ., dissenting).

In December 2017, the ACA's opponents achieved some legislative success. As part of the Tax Cuts and Jobs Act, Congress set the "shared responsibility payment" amount—the amount a person must pay for failing to comply with the individual mandate—to the "lesser" of "zero percent" of an individual's household income or "$0," effective January 2019. Pub. L. No. 115-97, § 11081, 131 Stat. 2054, 2092 (2017); see also 26 U.S.C. § 5000A(c). The individual mandate is still "on the books" of the U.S. Code and still consists of the three fundamental components it always featured. Subsection (a) prescribes that certain individuals "shall ... ensure" that they and their dependents are "covered under minimum essential coverage." 26 U.S.C. § 5000A(a). Subsection (b) "impose[s]... a penalty" called a "[s]hared responsibility payment" on those who fail to ensure they have minimum essential coverage. 26 U.S.C. § 5000A(b). Subsection (c) sets the amount of that payment. All Congress did in 2017 was change the amount in subsection (c) to zero dollars. 26 U.S.C. § 5000A(c).

Two months after the shared responsibility payment was set at zero dollars, the plaintiffs here—two private citizens8 and eighteen states9—filed this lawsuit against several federal defendants: the United States of America, the Department of Health and Human Services and its Secretary, Alex Azar, as well as the Internal Revenue Service and its Acting Commissioner, David J. Kautter. The plaintiffs argued that the individual mandate was no longer constitutional because: (1) NFIB rested the individual mandate's constitutionality exclusively on reading the provision as a tax; and (2) the 2017 amendment undermined any ability to characterize the individual mandate as a tax because the provision no longer generates revenue, a requirement for a tax. The plaintiffs argued further that, because the individual mandate was essential to and inseverable from the rest of the ACA, the entire ACA must be enjoined. On this theory, the plaintiffs sought declaratory relief that the individual mandate is unconstitutional and the rest of the ACA is inseverable. The plaintiffs also sought an injunction prohibiting the federal defendants from enforcing any provision of the ACA or its regulations.

The federal defendants agreed with the plaintiffs that once the shared responsibility payment was reduced to zero dollars, the individual mandate was no longer constitutional. They also agreed that the individual mandate could not be severed from the ACA's guaranteed-issue and community-rating requirements. Unlike the plaintiffs, however, the federal defendants contended in the district court that those three provisions could be severed from the rest of the Act. Driven by the federal defendants' decision not to fully defend against the lawsuit, sixteen states10 and the District of Columbia intervened to defend the ACA.

The district court agreed with the plaintiffs' arguments on the merits. Specifically, the court held that: (1) the individual plaintiffs had standing because the individual mandate compelled them to purchase insurance; (2) setting the shared responsibility payment to zero rendered the individual mandate unconstitutional; and (3) the unconstitutional provision could not be severed from any other part of the ACA. The district court granted the plaintiffs' claim for declaratory relief. Specifically, the district court's order "declares the Individual Mandate, 26 U.S.C. § 5000A(a), UNCONSTITUTIONAL," and the order further declares that "the remaining provisions of the ACA, Pub L. 111-148, are INSEVERABLE and therefore INVALID." The district court, however, denied the plaintiffs' application for a preliminary injunction. The district court entered partial final judgment11 as to the grant of summary judgment for declaratory relief, but stayed judgment pending appeal. This appeal followed.

On appeal, the U.S. House of Representatives intervened to join the intervenor-defendant states in defending the ACA.12 Also on appeal, the federal defendants changed their litigation position. After contending in the district court that only a few provisions of the ACA were inseverable from the individual mandate, the federal defendants contend in their opening brief for the first time that all of the ACA is inseverable. See Fed. Defendants' Br. at 43-49. Moreover, the federal defendants contend for the first time on appeal that— even though the entire ACA is inseverable—the court should not enjoin the enforcement of the entire ACA. The federal defendants now argue that the district court's judgment should be affirmed "except insofar as it purports to extend relief to ACA provisions that are unnecessary to remedy plaintiffs' injuries."13 Fed. Defendants' Br. at 49. They also now argue that the district court's judgment "cannot be understood as extending beyond the plaintiff states to invalidate the ACA in the intervenor states." Fed. Defendants' Supp. Br. at 10. Simply put, the federal defendants have shifted their position on appeal more than once.

II.

We review a district court's grant of summary judgment de novo. Time Warner Cable, Inc. v. Hudson, 667 F.3d 630, 638 (5th Cir. 2012). Summary judgment is appropriate when "the movant shows that there is no genuine dispute as to any material fact and the movant is entitled to judgment as a matter of law." Fed. R. Civ. P. 56(a); see also Dialysis Newco, Inc. v. Cmty. Health Sys. Grp. Health Plan, 938 F.3d 246, 250 (5th Cir. 2019). A dispute about a material fact is genuine if "the evidence is such that a reasonable jury could return a verdict for the non-moving party." Amerisure Ins. v. Navigators Ins., 611 F.3d 299, 304 (5th Cir. 2010) (quoting Gates v. Tex. Dep't of Protective & Regulatory Servs., 537 F.3d 404, 417 (5th Cir. 2008)). When ruling on a motion for summary judgment, the court views all inferences drawn from the factual record "in the light most favorable to the non-moving parties below." Trent v. Wade, 776 F.3d 368, 373 n.1 (5th Cir. 2015).

III.

We first must consider whether there is a live "[c]ase" or "[c]ontroversy" before us on appeal, as Article III of the U.S. Constitution requires. U.S. Const. art. III, § 1. A case or controversy does not exist unless the person asking the court for a decision—in this case, asking us to decide whether the district court's judgment was correct—has standing, which requires a showing of "injury, causation, and redressability." Sierra Club v. Babbitt, 995 F.2d 571, 574 (5th Cir. 1993). When "standing to appeal is at issue, appellants must demonstrate some injury from the judgment below." Id. at 575 (emphasis omitted).

We conclude, as all parties agree, that there is a case or controversy before us on appeal. Two groups of parties appealed from the district court's judgment: the federal defendants, and the intervenor-defendant states.14 There is a case or controversy before us because both of these groups have their own independent standing to appeal.15

The federal defendants have standing to appeal. The instant case is on all fours with the Supreme Court's decision in United States v. Windsor, 570 U.S. 744, 133 S.Ct. 2675, 186 L.Ed.2d 808 (2013). In that case, the executive branch of the federal government declined to defend a federal statute that did not allow the surviving spouse of a same-sex couple to receive a spousal tax deduction. Id. at 749-53, 133 S.Ct. 2675. The district court ruled that the statute was unconstitutional and ordered the executive branch to issue a tax refund to the surviving spouse. Id. at 754-55, 133 S.Ct. 2675. The executive branch agreed with the district court's legal conclusion, but it appealed the judgment and continued to enforce the statute by withholding the tax refund until a final judicial resolution. Id. at 757-58, 133 S.Ct. 2675.

The Supreme Court ruled that "the United States retain[ed] a stake sufficient to support Article III jurisdiction." Id. at 757, 133 S.Ct. 2675. That stake was the tax refund, which the federal government refused to pay. This threat of payment of money from the Treasury constituted "a real and immediate economic injury" to the federal government, which was sufficient for standing purposes. Id. at 757-58, 133 S.Ct. 2675 (quoting Hein v. Freedom From Religion Found., Inc., 551 U.S. 587, 599, 127 S.Ct. 2553, 168 L.Ed.2d 424 (2007) (plurality opinion)). As the Court explained, "the refusal of the Executive to provide the relief sought suffices to preserve a justiciable dispute as required by Article III." Windsor, 570 U.S. at 759, 133 S.Ct. 2675; see also Food Mktg. Inst. v. Argus Leader Media, ___ U.S. ___, 139 S.Ct. 2356, 2362, 204 L.Ed.2d 742 (2019) (concluding that there was a justiciable controversy because the government "represented unequivocally" that it would not voluntarily moot the controversy absent a final judicial order, and "[t]hat is enough to satisfy Article III"); INS v. Chadha, 462 U.S. 919, 939, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983) (holding that there was "adequate Art. III adverseness" because the executive branch determined that a federal statute was unconstitutional and refused to defend it but simultaneously continued to abide by it).

The instant case is similar. Though the plaintiffs and the federal defendants are in almost complete agreement on the merits of the case, the government continues to enforce the entire Act. The federal government has made no indication that it will begin dismantling any part of the ACA in the absence of a final court order. Just as in Windsor, then, effectuating the district court's order would require the federal government to take actions that it would not take "but for the court's order." Windsor, 570 U.S. at 758, 133 S.Ct. 2675. And just as in Windsor, the federal defendants stand to suffer financially if the district court's judgment is affirmed.16 As just one example, the district court's judgment declares the Act's Medicare reimbursement schedules unlawful, which, if given effect, would require Medicare to reimburse healthcare providers at higher rates. See, e.g., 42 U.S.C. § 1395ww(b)(3)(B)(xi)-(xii). Therefore, just as in Windsor, an appellate decision here will "have real meaning." 570 U.S. at 758, 133 S.Ct. 2675 (quoting Chadha, 462 U.S. at 939, 103 S.Ct. 2764).17

The intervenor-defendant states also have standing to appeal. While a party's mere "status as an intervenor below ... does not confer standing," Diamond v. Charles, 476 U.S. 54, 68, 106 S.Ct. 1697, 90 L.Ed.2d 48 (1986), intervenors may appeal if they can demonstrate injury from the district court's judgment. Sierra Club, 995 F.2d at 574; see also Va. House of Delegates v. Bethune-Hill, ___ U.S. ___, 139 S.Ct. 1945, 1951, 204 L.Ed.2d 305 (2019); Cooper v. Tex. Alcoholic Beverage Comm'n, 820 F.3d 730, 737 (5th Cir. 2016). The intervenor-defendant states have made this showing because the district court's judgment, if ultimately given effect, would: (1) strip these states of funding that they receive under the ACA; and (2) threaten to hamstring these states in possible future litigation because of the district court judgment's potentially preclusive effect.18

First, the intervenor-defendant states receive significant funding from the ACA, which would be discontinued if we affirmed the district court's judgment declaring the entire Act unconstitutional. "[F]inancial loss as a result of" a district court's judgment is an injury sufficient to support standing to appeal. United States v. Fletcher ex rel. Fletcher, 805 F.3d 596, 602 (5th Cir. 2015). In their supplemental briefing, the intervenor-defendant states identify a few examples of the funding sources they would lose under the district court's judgment. Evidence in the record shows that eliminating the Act's Medicaid expansion provisions alone would cost the original sixteen intervening state defendants and the District of Columbia a total of more than $418 billion in the next decade. See 42 U.S.C. §§ 1396a(a)(10)(A)(i)(VIII), (e)(14)(I)(i), 1396d(y)(1). Moreover, the Act's Community First Choice Option program gives states funding to care for the disabled and elderly at home or in their communities instead of in institutions. See 42 U.S.C. § 1396n(k). Record evidence shows that eliminating this program would cost California $400 million in 2020, and that Oregon and Connecticut have already received $432.1 million under this program. This evidence is more than enough to show that the intervenor-defendant states would suffer financially if the district court's judgment is given effect, an injury sufficient to confer standing to appeal. See Dep't of Commerce v. New York, ___ U.S. ___, 139 S.Ct. 2551, 2565, 204 L.Ed.2d 978 (2019).

The district court's judgment, if given effect, also threatens to injure the intervenor-defendant states with the judgment's potentially preclusive effect in future litigation. We have held that "[a] party may be aggrieved by a district court decision that adversely affects its legal rights or position vis-à-vis other parties in the case or other potential litigants." Leonard v. Nationwide Mut. Ins., 499 F.3d 419, 428 (5th Cir. 2007) (quoting Custer v. Sweeney, 89 F.3d 1156, 1164 (4th Cir. 1996)). If the federal defendants began unwinding the ACA, either in reliance on the district court's judgment or on their own, the district court's judgment would potentially estop the intervenor-defendant states from challenging that action in court. This case thus stands in contrast to the cases in which there was no chance whatsoever of a preclusive effect. See Klamath Strategic Inv. Fund ex rel. St. Croix Ventures v. United States, 568 F.3d 537, 546 (5th Cir. 2009) (holding that there was no threatened injury from potential estoppel from the appealed-from judgment because that judgment was interlocutory, not final, and therefore could not estop the appealing party).

Finally, we examine the standing of the U.S. House of Representatives, which intervened after the case had been appealed. The Supreme Court's recent decision in Virginia House of Delegates v. Bethune-Hill calls the House's standing to intervene into doubt. 139 S. Ct. at 1953 ("This Court has never held that a judicial decision invalidating a state law as unconstitutional inflicts a discrete, cognizable injury on each organ of government that participated in the law's passage."). However, we need not resolve the question of the House's standing. "Article III does not require intervenors to independently possess standing" when a party already in the lawsuit has standing and seeks the same "ultimate relief" as the intervenor. Ruiz v. Estelle, 161 F.3d 814, 830 (5th Cir. 1998). That is the case here: the intervenor-defendant states have standing to appeal, and the House seeks the same relief as those states. We accordingly pretermit the issue of whether the House has standing to intervene.

IV.

We now turn to the issue of whether any of the plaintiffs had Article III standing to bring this case at the time they brought the lawsuit. To be a case or controversy under Article III, the plaintiffs must satisfy the same three requirements listed above. First, a plaintiff must have suffered an "injury in fact"—a violation of a legally protected interest that is "concrete and particularized," as well as "actual or imminent, not `conjectural' or `hypothetical.'" Lujan v. Defs. of Wildlife, 504 U.S. 555, 560, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992) (quoting Whitmore v. Arkansas, 495 U.S. 149, 155, 110 S.Ct. 1717, 109 L.Ed.2d 135 (1990)). Second, that injury must be "fairly ... trace[able] to the challenged action of the defendant, and not ... th[e] result [of] the independent action of some third party not before the court." Id. (alterations in original) (quoting Simon v. E. Ky. Welfare Rights Org., 426 U.S. 26, 41-42, 96 S.Ct. 1917, 48 L.Ed.2d 450 (1976)). Third, it must be "likely"—not merely "speculative"—that the injury will be "redressed by a favorable decision." Id. at 561, 112 S.Ct. 2130 (quoting Simon, 426 U.S. at 38, 43, 96 S.Ct. 1917).

The instant case has two groups of plaintiffs: the individual plaintiffs and the state plaintiffs. Only one plaintiff need succeed because "one party with standing is sufficient to satisfy Article III's case-or-controversy requirement."19 Texas v. United States (DAPA), 809 F.3d 134, 151 (5th Cir. 2015) (quoting Rumsfeld v. Forum for Acad. & Institutional Rights, Inc., 547 U.S. 47, 52 n.2, 126 S.Ct. 1297, 164 L.Ed.2d 156 (2006)).20 The individual plaintiffs and the state plaintiffs allege different injuries. We evaluate each in turn and conclude that both the individual plaintiffs and the state plaintiffs have standing.

A.

The standing issues presented by the individual plaintiffs are not novel. The Supreme Court faced a similar situation when it decided NFIB in 2012. At oral argument in that case, Justice Kagan asked Gregory Katsas, representing NFIB, whether he thought "a person who is subject to the [individual] mandate but not subject to the [shared responsibility payment] would have standing." Transcript of Oral Argument at 68, Dep't of Health and Human Servs. v. Florida, 567 U.S. 519, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012). Mr. Katsas replied, "Yes, I think that person would, because that person is injured by compliance with the mandate." Id. Mr. Katsas explained, "the injury— when that person is subject to the mandate, that person is required to purchase health insurance. That's a forced acquisition of an unwanted good. It's a classic pocketbook injury." Id. at 68-69.

In 2012, this questioning made sense because neither the individual mandate nor the shared responsibility payment would be assessed for another two years. Patient Protection and Affordable Care Act, Pub. L. No. 111-148, § 1501, 124 Stat. 119, 244 (2012) (requiring insurance coverage "for each month beginning after 2013" and applying the shared responsibility payment for any failure to purchase insurance "during any calendar year beginning after 2013"). It was thus certainly imminent that the private plaintiffs would be subject to the individual mandate, which applies to everyone, but not certain that they would be subject to the shared responsibility payment, which exempts certain people. 26 U.S.C. § 5000A(e) (prescribing that "[n]o penalty shall be imposed" on certain groups of people).21 The distinction was important because a plaintiff "must demonstrate standing for each claim he seeks to press." Davis v. Fed. Election Comm'n, 554 U.S. 724, 734, 128 S.Ct. 2759, 171 L.Ed.2d 737 (2008) (quoting Daimler-Chrysler Corp. v. Cuno, 547 U.S. 332, 352, 126 S.Ct. 1854, 164 L.Ed.2d 589 (2006)). To bring a claim against the individual mandate, therefore, the plaintiffs needed to show injury from the individual mandate—not from the shared responsibility payment.

Accordingly, the district court in NFIB ruled that the private plaintiffs were injured by the ACA "because of the financial expense [they would] definitively incur under the Act in 2014," and the private plaintiffs' need "to take investigatory steps and make financial arrangements now to ensure compliance then." Florida ex rel. Bondi v. U.S. Dep't of Health & Human Servs., 780 F.Supp.2d 1256, 1271 (N.D. Fla. 2011), aff'd in part and rev'd in part, 648 F.3d 1235 (11th Cir. 2011), aff'd in part and rev'd in part, 567 U.S. 519, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012). The record evidence in that case supported this conclusion. Mary Brown, one of the private plaintiffs in that case, for example, had declared that "to comply with the individual insurance mandate, and well in advance of 2014, I must now investigate whether and how to rearrange my personal finance affairs." Appendix of Exhibits in Support of Plaintiffs' Motion for Summary Judgment, Florida v. U.S. Dep't of Health & Human Servs., No. 3:10-cv-91-RV/EMT (N.D. Fla. Nov. 10, 2010), ECF No. 80-6. At the Eleventh Circuit, all parties agreed that Mary Brown had standing. Florida ex rel. Atty. Gen. v. U.S. Dep't of Health & Human Servs., 648 F.3d 1235, 1243 (11th Cir. 2011), aff'd in part and rev'd in part, 567 U.S. 519, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012) ("Defendants do not dispute that plaintiff Brown's challenge to the minimum coverage provision is justiciable."). Congress could have reasonably contemplated people like Mary Brown. As Mr. Katsas explained at oral argument in the Supreme Court, "Congress reasonably could think that at least some people will follow the law precisely because it is the law." Transcript of Oral Argument at 67, Dep't of Health & Human Servs. v. Florida, 567 U.S. 519, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012).

The district court in the instant case followed a similar approach with regard to the individual plaintiffs' standing.22 It concluded that because the individual plaintiffs are the object of the individual mandate, which requires them to purchase health insurance that they do not want, those plaintiffs have demonstrated two types of "injury in fact": (1) the financial injury of buying that insurance; and (2) the "increased regulatory burden" that the individual mandate imposes. In concluding that these injuries were caused by the individual mandate, the court made specific fact findings that both Nantz and Hurley purchased insurance solely because they are "obligated to comply with the ... individual mandate." The district court made these findings based on Nantz's and Hurley's declarations, which the intervenor-defendant states never challenged. Because the undisputed evidence showed that the individual mandate caused these injuries, the district court reasoned that a favorable judgment would redress both injuries, allowing the individual plaintiffs to forgo purchasing health insurance and freeing them "from what they essentially allege to be arbitrary governance."

We agree with the district court. The Supreme Court has held that when a lawsuit challenges "the legality of government action or inaction, the nature and extent of facts that must be averred (at the summary judgment stage) or proved (at the trial stage) in order to establish standing depends considerably upon whether" the plaintiffs are themselves the "object[s] of the action (or forgone action) at issue." Lujan, 504 U.S. at 561, 112 S.Ct. 2130; see also Texas v. EEOC, 933 F.3d 433, 446 (5th Cir. 2019). "Whether someone is in fact an object of a regulation is a flexible inquiry rooted in common sense." EEOC, 933 F.3d at 446 (quoting Contender Farms, L.L.P. v. U.S. Dep't of Agric., 779 F.3d 258, 265 (5th Cir. 2015)). If a plaintiff is indeed the object of a regulation, "there is ordinarily little question that the action or inaction has caused [the plaintiff] injury, and that a judgment preventing or requiring the action will redress it." Lujan, 504 U.S. at 561-62, 112 S.Ct. 2130.

It is undisputed that Hurley and Nantz are the objects of the individual mandate and that they have purchased insurance in order to comply with that mandate. Record evidence supports these conclusions. In his declaration in the district court, Nantz stated, "I continue to maintain minimum essential health coverage because I am obligated." Similarly, Hurley averred in his declaration that he is "obligated to comply with the ACA's individual mandate." They both explain in their declarations that they "value compliance with [their] legal obligations" and bought insurance because they "believe that following the law is the right thing to do." Accordingly, the district court expressly found that Hurley and Nantz bought health insurance because they are obligated to, and we must defer to that factual finding. The evidentiary basis for this injury is even stronger than it was in NFIB. In the instant case, the individual mandate has already gone into effect, compelling Nantz and Hurley to purchase insurance now as opposed to two years in the future.

The intervenor-defendant states fail to point to any evidence contradicting these declarations, and they did not challenge this evidence in the district court. In fact, some of the evidence these parties rely on actually supports the conclusion that Nantz and Hurley purchased insurance to comply with the individual mandate. The intervenor-defendant states acknowledge a 2017 report from the Congressional Budget Office indicating that "a small number of people" would continue to buy insurance without a penalty "solely because" of a desire to comply with the law. Cong. Budget Office, Repealing the Individual Health Insurance Mandate: An Updated Estimate 1 (Nov. 2017). This report is at least somewhat consistent with a 2008 Congressional Budget Office report, relied on by the state plaintiffs, that "[m]any individuals" subject to the mandate, but not the shared responsibility payment, will obtain coverage to comply with the mandate "because they believe in abiding by the nation's laws." Cong. Budget Office, Key Issues in Analyzing Major Health Insurance Proposals 53 (Dec. 2008). Whether this group of law-abiding citizens includes "many individuals" or "a small number of people," Nantz and Hurley have undisputed evidence showing that they are a part of this group.

In this context, being required to buy something that you otherwise would not want is clearly within the scope of what counts as a "legally cognizable injury." "Economic injury" of this sort is "a quintessential injury upon which to base standing." Tex. Democratic Party v. Benkiser, 459 F.3d 582, 586 (5th Cir. 2006); see also Vt. Agency of Nat. Res. v. United States, 529 U.S. 765, 772-77, 120 S.Ct. 1858, 146 L.Ed.2d 836 (2000) (finding Article III injury from financial harm); Clinton v. New York, 524 U.S. 417, 432, 118 S.Ct. 2091, 141 L.Ed.2d 393 (1998) (same); Sierra Club v. Morton, 405 U.S. 727, 733-34, 92 S.Ct. 1361, 31 L.Ed.2d 636 (1972) (same); DAPA, 809 F.3d at 155 (same). In Benkiser, for example, we held that a political party would suffer an injury in fact because it would need to "expend additional funds" in order to comply with the challenged regulation. 459 F.3d at 586. In the instant case, the undisputed record evidence shows that the individual plaintiffs have spent "additional funds" to comply with the statutory provision that they challenge on constitutional grounds.

This injury, moreover, is "actual," not merely a speculative fear about future harm that may or may not happen. Lujan, 504 U.S. at 560, 112 S.Ct. 2130. The record shows that, at the time of the complaint, Hurley and Nantz held health insurance, spending money every month that they did not want to spend. Nantz reports that his monthly premium is $266.56, and Hurley says his is $1,081.70. The injury is also "concrete" because it involves the real expenditure of those funds. See Barlow v. Collins, 397 U.S. 159, 162-63, 164, 90 S.Ct. 832, 25 L.Ed.2d 192 (1970) (finding a concrete injury when a regulation caused economic harm from lost profit).

Causation and redressability "flow naturally" from this concrete, particularized injury. Contender Farms, 779 F.3d at 266. The evidence in the record from Hurley's and Nantz's declarations show that they would not have purchased health insurance but for the individual mandate, and the intervenor-defendant states have no evidence to the contrary. A judgment declaring that the individual mandate exceeds Congress' powers under the Constitution would allow Hurley and Nantz to forgo the purchase of health insurance that they do not want or need. They could purchase health insurance below the "minimum essential coverage" threshold, or even decide not to purchase any health insurance at all.

The intervenor-defendant states make several arguments against this straightforward injury, and all of them come up short. They first argue that there is no legally cognizable injury because there is no longer any penalty for failing to comply. In one sense, this argument misses the point. The threat of a penalty that Hurley and Nantz would face under the pre-2017 version of the statute is one potential form of injury, but it is far from the only one. We have held that the costs of compliance can constitute an injury just as much as the injuries from failing to comply. See, e.g., Benkiser, 459 F.3d at 586. Thus, in this instance, it is this injury—the time and money spent complying with the statute, not the penalty for failing to do so—that constitutes the plaintiffs' injury.

But the intervenor-defendant states also argue that even the costs of compliance cannot count as an injury in fact if there is no consequence for failing to comply. The individual mandate's compulsion cannot inflict a cognizable injury, they say, because it is not a compulsion at all. Because the enforcement mechanism has been removed, the U.S. House contends, it is now merely a suggestion, at most. We recently rejected this argument in Texas v. EEOC, when the Equal Employment Opportunity Commission tried to argue that Texas could not challenge its allegedly non-final administrative guidance because "the Guidance does not compel Texas to do anything." 933 F.3d at 448. We concluded that it would "strain credulity to find that an agency action targeting current `unlawful' discrimination among state employers—and declaring presumptively unlawful the very hiring practices employed by state agencies—does not require action immediately enough to constitute an injury-in-fact."23 Id. The individual mandate is no different. Just like the agency guidance, the individual mandate targets as "unlawful" the decision to go without health insurance.

The dissenting opinion grounds its discussion of the issue in the Supreme Court's decision in Poe v. Ullman, 367 U.S. 497, 81 S.Ct. 1752, 6 L.Ed.2d 989 (1961). There, the Supreme Court rejected a challenge to Connecticut's criminal prohibition on contraception. The dissenting opinion states that if there was no standing in Ullman, then there cannot be standing here. The dissenting opinion seems to treat Ullman as part of the "pre-enforcement challenge" line of cases in which the Supreme Court analyzed claims of injury based on future enforcement to determine whether the future enforcement was sufficiently imminent. Ullman, however, is not cited in the seminal Supreme Court cases of that line. See, e.g., Susan B. Anthony List v. Driehaus, 573 U.S. 149, 158-61, 134 S.Ct. 2334, 189 L.Ed.2d 246 (2014); Holder v. Humanitarian Law Project, 561 U.S. 1, 15, 130 S.Ct. 2705, 177 L.Ed.2d 355 (2010); Virginia v. Am. Booksellers Ass'n, Inc., 484 U.S. 383, 392-93, 108 S.Ct. 636, 98 L.Ed.2d 782 (1988); Babbitt v. United Farm Workers Nat'l Union, 442 U.S. 289, 298, 99 S.Ct. 2301, 60 L.Ed.2d 895 (1979); see also Abbott Labs. v. Gardner, 387 U.S. 136, 154, 87 S.Ct. 1507, 18 L.Ed.2d 681 (1967). More importantly, as we have explained, this case is not a pre-enforcement challenge because the plaintiffs have already incurred a financial injury.24

The plurality opinion in Ullman said there was insufficient adversity between the parties because there was overwhelming evidence—eighty years' worth of no enforcement of the statute—of "tacit agreement" between prosecutors and the public not to enforce the anti-contraceptive laws that the plaintiffs challenged. 367 U.S. at 507-08, 81 S.Ct. 1752. As a result, the Court held that the lawsuit before it was "not such an adversary case as will be reviewed here." Id. The fifth, controlling vote in that case—Justice Brennan, who concurred in the judgment—emphasized that this adverseness was lacking because of the case's "skimpy record," devoid of evidence that the "individuals [were] truly caught in an inescapable dilemma." Id. at 509, 81 S.Ct. 1752 (Brennan, J., concurring).

By contrast, as documented above, the record in the instant case contains undisputed evidence that Nantz and Hurley feel compelled by the individual mandate to buy insurance and that they bought insurance solely for that reason. Especially in light of the fact that the individual mandate lacks a similar eighty-year history of nonenforcement, Nantz and Hurley have gone much further in demonstrating that they are caught in the "inescapable dilemma" that the Ullman plaintiffs were not.

The intervenor-defendant states also argue that there is no causation between the individual mandate and Hurley and Nantz's purchase of insurance because Hurley and Nantz exercised a voluntary "choice" to purchase insurance. Because Nantz and Hurley would face no consequence if they went without insurance, the intervenor-defendant states argue that their purchase of insurance is not fairly traceable to the federal defendants. Instead, they claim that Nantz and Hurley impermissibly attempt to "manufacture standing merely by inflicting harm on themselves." Glass v. Paxton, 900 F.3d 233, 239 (5th Cir. 2018) (quoting Clapper v. Amnesty Int'l USA, 568 U.S. 398, 416, 133 S.Ct. 1138, 185 L.Ed.2d 264 (2013)).

This argument fails, however, because it conflates the merits of the case with the threshold inquiry of standing. The argument assumes that 26 U.S.C. § 5000A presents not a legal command to purchase insurance, but an option between purchasing insurance and doing nothing. Because this option exists, the argument goes, any injury arising from Hurley's and Nantz's decisions to buy insurance instead of doing nothing (the other putative option) is entirely self-inflicted. This, however, is a merits question that can be reached only after determining the threshold issue of whether plaintiffs have standing.

Texas v. EEOC makes clear that courts cannot fuse the standing inquiry into the merits in this way. There, in addition to the injury described above from the Guidance's rebuke of Texas's employment practices as "unlawful," Texas claimed it was injured by the EEOC's curtailing of Texas's procedural right to notice and comment before being subject to a regulation. EEOC, 933 F.3d at 447. In rejecting the suggestion that Texas was not truly injured because the EEOC had not in fact violated the Administrative Procedure Act's notice-and-comment rules, we held that "[w]e assume, for purposes of the standing analysis, that Texas is correct on the merits of its claim that the Guidance was promulgated in violation of the APA." Id. (citing Sierra Club v. EPA, 699 F.3d 530, 533 (D.C. Cir. 2012)); see also Bennett v. Spear, 520 U.S. 154, 177-78, 117 S.Ct. 1154, 137 L.Ed.2d 281 (1997) (treating constitutional standing and finality as distinct inquiries).

Indeed, allowing a consideration of the merits as part of a jurisdictional inquiry would conflict with the Supreme Court's express decision in Steel Co v. Citizens for a Better Environment to not abandon "two centuries of jurisprudence affirming the necessity of determining jurisdiction before proceeding to the merits." 523 U.S. 83, 98, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998). That case presented both the question of Article III standing and the merits question of whether the relevant statute authorized lawsuits for purely past violations. Id. at 86, 118 S.Ct. 1003. The Court rejected any "attempt to convert the merits issue ... into a jurisdictional one." Id. at 93, 118 S.Ct. 1003. The Court further rejected the "doctrine of hypothetical jurisdiction," under which certain courts of appeals had "proceed[ed] immediately to the merits question, despite jurisdictional objections" in certain circumstances. Id. at 93-94, 118 S.Ct. 1003. As the district court correctly noted, that is exactly what the appellants ask this court to do. They urge us to "skip ahead to the merits to determine § 5000A(a) is non-binding and therefore constitutional and then revert to the standing analysis to use its merits determination to conclude there was no standing to reach the merits in the first place."

Moreover, even if we were to consider the merits as part of our jurisdictional inquiry, it would not make a difference in this case. Because we conclude in Part IV of this opinion that the individual mandate is best read as a command to purchase insurance (and an unconstitutional one at that), rather than as an option between buying insurance or doing nothing, the individual plaintiffs would have standing even if we considered the merits.25

B.

We next consider whether the eighteen state plaintiffs have standing, and we conclude that they do.26 The state plaintiffs allege that the ACA causes them both a fiscal injury as employers and a sovereign injury "because it prevents them from applying their own laws and policies governing their own healthcare markets." State Plaintiffs' Br. at 25. In DAPA, we determined that the state of Texas was entitled to special solicitude because it was "exercising a procedural right created by Congress and protecting a `quasi-sovereign' interest." DAPA, 809 F.3d at 162 (quoting Massachusetts v. EPA, 549 U.S. 497, 520, 127 S.Ct. 1438, 167 L.Ed.2d 248 (2007)); see also id. at 154-55. Because the state plaintiffs in this case have suffered fiscal injuries as employers, we need not address special solicitude or the alleged sovereign injuries.

Employers, including the state plaintiffs, are required by the ACA to issue forms verifying which employees are covered by minimum essential coverage and therefore do not need to pay the shared responsibility payment. See 26 U.S.C. § 6055(a) ("Every person who provides minimum essential coverage to an individual during a calendar year shall, at such time as the Secretary may prescribe, make a return described in subsection (b)."); 26 U.S.C. § 6056(a) ("Every applicable large employer [that meets certain statutory requirements] shall ... make a return described in subsection (b)."). These provisions have led to Form 1095-B and 1095-C statements that employees receive from their employers around tax time, which include a series of check boxes indicating the months that employees had health coverage that complies with the ACA. State Plaintiffs' Br. at 23. These legally required reporting practices exist on top of state employers' own in-house administrative systems for managing and tracking their employees' health insurance coverage.

The record is replete with evidence that the individual mandate itself has increased the cost of printing and processing these forms and of updating the state employers' in-house management systems. For example, Thomas Steckel, the director of the Division of Employee Benefits within the South Dakota Bureau of Human Resources, submitted a declaration documenting the administrative costs that the individual mandate has imposed by way of these reporting requirements. He said, "[t]he individual mandate caused significant administrative burdens and expenses to program our IT system to track and report ACA eligible employees and complete mandatory IRS Form 1095 annual reports." Steckel noted specifically that "the individual mandate caused ... $100,000.00 [in] ongoing costs" for Form 1095-C administration alone. The dissenting opinion discards this evidence as conclusory. But as even counsel for the intervenor-defendant states admitted at oral argument, nobody challenged this evidence as conclusory in the district court or in the appellate court.27 Oral Argument at 5:12.

South Dakota is far from the only state that has been harmed from the financial cost of the reporting requirements that the individual mandate aggravates. Judith Muck, the Executive Director of the Missouri Consolidated Health Care Plan, reported that Missouri's costs for preparing 1095-B forms, along with 1094-B forms, are projected to be $47,300 in fiscal year 2019 and $49,200 in fiscal year 2020. Similarly, Teresa MacCartney, the Chief Financial Officer of the State of Georgia and the Director of the Georgia Governor's Office of Planning and Budget, reported that Georgia's overall cost of compliance with the ACA's reporting requirements "is an estimated net $3.6 million to date." MacCartney also reported that after the ACA's implementation, Georgia's Department of Community Health "experienced increased enrollment of individuals already eligible for Medicaid benefits under pre-ACA eligibility standards." This enrollment increase required the Department to enhance its management systems, which was "very costly." Blaise Duran, who is the Manager for Underwriting, Data Analysis and Reporting for the Employees Retirement System of Texas, further documented Texas' costs of the reporting requirements. He declared that the Texas Employees Group Benefits Program "has made administrative process changes in connection with its ACA compliance, such as those related to the provision of Form 1095-Bs to plan participants and the Internal Revenue Service."28

The intervenor-defendant states and the U.S. House have not challenged the state plaintiffs' evidence or presented any evidence to the contrary. Instead, they argue that the reporting requirements set forth in Sections 6055(a) and 6056(a) "are separate from the mandate and serve independent purposes." U.S. House Reply Br. at 19. Therefore, they claim, "any resulting injury is thus neither traceable to Section 5000A nor redressable by its invalidation." U.S. House Reply Br. at 19. But this misreads the undisputed evidence in the record. The individual mandate commands individuals to get insurance. Every time an individual gets that insurance through a state employer, the state employer must send the individual a form certifying that he or she is covered and otherwise process that information through in-house management systems.29 Thus, the reporting requirements in Sections 6055(a) and 6056(a) flow from the individual mandate set forth in Section 5000A(a).

These costs to the state plaintiffs are well-established.30 Moreover, the continuing nature of these fiscal injuries is consistent with Fifth Circuit and Supreme Court precedent.

In DAPA, we held that the state of Texas had standing to challenge the federal government's DAPA program because it stood to "have a major effect on the states' fisc." Id. at 152. This was because, if DAPA were permitted to go into effect, it would have "enable[d] at least 500,000 illegal aliens in Texas" to satisfy Texas's requirements that the Department of Public Safety "`shall issue' a license to a qualified applicant," including noncitizens who present "documentation issued by the appropriate United States agency that authorizes the applicant to be in the United States." Id. at 155 (quoting Tex. Transp. Code §§ 521.142(a), 521.181). Evidence in the record showed that Texas, which subsidizes its licenses, would "lose a minimum of $130.89 on each one it issued to a DAPA beneficiary." Id. Even a "modest estimate" of predictable third-party behavior would rack up costs of "several million dollars." Id.

The Supreme Court recently applied a similar analysis in Department of Commerce v. New York, ___ U.S. ___, 139 S.Ct. 2551, 204 L.Ed.2d 978 (2019). In that case, a group of state and local governments sued to prevent the federal government from including a question about citizenship status on the 2020 census. Id. at 2563. The Supreme Court held that these plaintiffs had standing because they met their burden "of showing that third parties will likely react in predictable ways to the citizenship question." Id. at 2566. The census question would likely lead to "noncitizen households responding ... at lower rates than other groups, which in turn would cause them to be undercounted." Id. at 2565. This undercounting of third parties would injure the state and local governments by "diminishment of political representation, loss of federal funds, degradation of census data, and diversion of resources." Id.

In both DAPA and Department of Commerce, the state plaintiffs demonstrated injury by showing that the challenged law would cause third parties to behave in predictable ways, which would inflict a financial injury on the states. The instant case is no different. The individual mandate commands people to ensure that they have minimum health insurance coverage. That predictably causes more people to buy insurance, which increases the administrative costs of the states to report, manage, and track the insurance coverage of their employees and Medicaid recipients.31

V.

Having concluded that both groups of plaintiffs have standing to bring this lawsuit, we must next determine whether the individual mandate is a constitutional exercise of congressional power. We conclude that it is not. We first discuss the Supreme Court's holding in NFIB, and then we explain why, under that holding, the individual mandate is no longer constitutional.

A.

The NFIB opinion was extremely fractured. In that case, Chief Justice Roberts wrote an opinion addressing several issues. Parts of that opinion garnered a majority of votes and served as the opinion of the Court.32 In relevant part, Part III-A of the Chief Justice's opinion, joined by no other Justice, observed that "[t]he most straightforward reading of the [individual] mandate is that it commands individuals to purchase insurance," and that, using that reading of the statute, the individual mandate is not a valid exercise of Congress' power under the Interstate Commerce Clause. NFIB, 567 U.S. at 562, 546-61, 132 S.Ct. 2566 (Roberts, C.J.). The Constitution, he explained, "gave Congress the power to regulate commerce, not to compel it." Id. at 555, 132 S.Ct. 2566 (Roberts, C.J.). For similar reasons, the Chief Justice concluded that this command to purchase insurance could not be sustained under the Constitution's Necessary and Proper Clause. Id. The individual mandate was not "proper" because it expanded federal power, "vest[ing] Congress with the extraordinary ability to create the necessary predicate to the exercise of" its Interstate Commerce Clause powers. Id. at 560, 132 S.Ct. 2566.

Though no other Justices joined this part of the Chief Justice's opinion, the "joint dissent"—joined by Justices Scalia, Kennedy, Thomas, and Alito—reached the same conclusions on the Interstate Commerce Clause and Necessary and Proper Clause questions. Id. at 650-60, 132 S.Ct. 2566 (joint dissent). A majority of the court, therefore, concluded that the individual mandate is not constitutional under either the Interstate Commerce Clause or the Necessary and Proper Clause.

This limited reading of the Interstate Commerce Clause—and, by extension, of the Necessary and Proper Clause—was necessary to preserving "the country [that] the Framers of our Constitution envisioned." Id. at 554, 132 S.Ct. 2566 (Roberts, C.J.). As Chief Justice Roberts observed, if the individual mandate were a proper use of the power to regulate interstate commerce, that power would "justify a mandatory purchase to solve almost any problem." Id. at 553, 132 S.Ct. 2566 (Roberts, C.J.). If Congress can compel the purchase of health insurance today, it can, for example, micromanage Americans' day-to-day nutrition choices tomorrow. Id. (Roberts, C.J.); see also id. at 558, 132 S.Ct. 2566 (Roberts, C.J.) (reasoning that, under an expansive view of the Commerce Clause, nothing would stop the federal government from compelling the purchase of broccoli).

An expansive reading of the Interstate Commerce Clause would be foreign to the Framers, who saw the clause as "an addition which few oppose[d] and from which no apprehensions [were] entertained." Id. at 554, 132 S.Ct. 2566 (Roberts, C.J.) (quoting The Federalist No. 45, at 293 (J. Madison) (C. Rossiter ed., 1961)). Elevating Congress' power to "regulate commerce... among the several states," U.S. Const. art. I, § 8, cl. 3, to a power to create commerce among the several states would make a Leviathan of the federal government, "everywhere extending the sphere of its activity and drawing all power into its impetuous vortex." NFIB, 567 U.S. at 554, 132 S.Ct. 2566 (Roberts, C.J.) (quoting The Federalist No. 48, at 309 (J. Madison) (C. Rossiter ed., 1961)). Justice Scalia, writing for the joint dissenters, similarly noted that the more expansive reading of the Interstate Commerce Clause would render that provision a "font of unlimited power," id. at 653, 132 S.Ct. 2566 (joint dissent), or, in the words of Alexander Hamilton, a "hideous monster whose devouring jaws ... spare neither sex nor age, nor high nor low, nor sacred nor profane," id. (quoting The Federalist No. 33, at 202 (C. Rossiter ed., 1961)).

In Part III-B, again joined by no other Justice, Chief Justice Roberts concluded that because the individual mandate found no constitutional footing in the Interstate Commerce or Necessary and Proper Clauses, the Supreme Court was obligated to consider the federal government's argument that, as an exercise in constitutional avoidance, the mandate could be read not as a command but as an option to purchase insurance or pay a tax. This "option" interpretation of the statute could save the statute from being unconstitutional, as it would be justified under Congress' taxing power. Id. at 561-63, 132 S.Ct. 2566 (Roberts, C.J.); see also id. at 562, 132 S.Ct. 2566 (Roberts, C.J.) ("No court ought, unless the terms of an act rendered it unavoidable, to give a construction to it which should involve a violation, however unintentional, of the constitution.") (quoting Parsons v. Bedford, 28 U.S. (3 Pet.) 433, 448-49, 7 S.Ct. 732 (1830)); see also id. at 563, 132 S.Ct. 2566 (Roberts, C.J.) ("The question is not whether that is the most natural interpretation of the mandate, but only whether it is a `fairly possible' one.") (quoting Crowell v. Benson, 285 U.S. 22, 62, 52 S.Ct. 285, 76 S.Ct. 598 (1932)).

In Part III-C, the Chief Justice—writing for a majority of the Court, joined by Justices Ginsburg, Breyer, Sotomayor, and Kagan—undertook that inquiry of determining whether it was "fairly possible" to read the individual mandate as an option and thereby save its constitutionality. See id. at 563-74, 132 S.Ct. 2566 (majority opinion). Chief Justice Roberts reasoned that the individual mandate could be read in conjunction with the shared responsibility payment in order to save the individual mandate from unconstitutionality. Read together with the shared responsibility payment, the entire statutory provision could be read as a legitimate exercise of Congress' taxing power for four reasons.

First and most fundamentally, the shared-responsibility payment "yield[ed] the essential feature of any tax: It produce[d] at least some revenue for the Government." Id. at 564, 132 S.Ct. 2566. Second, the shared-responsibility payment was "paid into the Treasury by taxpayers when they file their tax returns." Id. at 563, 132 S.Ct. 2566 (alternations and internal quotation marks omitted). Third, the amount owed under the ACA was "determined by such familiar factors as taxable income, number of dependents, and joint filing status." Id. Fourth and finally, "[t]he requirement to pay [was] found in the Internal Revenue Code and enforced by the IRS, which ... collect[ed] it in the same manner as taxes." Id. at 563-64, 132 S.Ct. 2566 (internal quotation marks omitted).

Because of these four attributes of the shared responsibility payment, the Court reasoned that "[t]he Federal Government does have the power to impose a tax on those without health insurance." Id. at 575, 132 S.Ct. 2566. The Court concluded that "[s]ection 5000A is therefore constitutional, because it can reasonably be read as a tax."33 Id. We agree with the dissenting opinion that "this case begins and ought to end" with NFIB.

B.

Now that the shared responsibility payment amount is set at zero,34 the provision's saving construction is no longer available. The four central attributes that once saved the statute because it could be read as a tax no longer exist. Most fundamentally, the provision no longer yields the "essential feature of any tax" because it does not produce "at least some revenue for the Government." Id. at 564, 132 S.Ct. 2566. Because the provision no longer produces revenue, it necessarily lacks the three other characteristics that once rendered the provision a tax. The shared-responsibility payment is no longer "paid into the Treasury by taxpayer[s] when they file their tax returns" because the payment is no longer paid by anyone. Id. at 563, 132 S.Ct. 2566 (alteration in original and internal quotation marks omitted). The payment amount is no longer "determined by such familiar factors as taxable income, number of dependents, and joint filing status." Id. The amount is zero for everyone, without regard to any of these factors. The IRS no longer collects the payment "in the same manner as taxes" because the IRS cannot collect it at all. Id. at 563-64, 132 S.Ct. 2566 (internal quotation marks omitted).

Because these four critical attributes are now missing from the shared responsibility payment, it is, in the words of the state plaintiffs, "no longer `fairly possible' to save the mandate's constitutionality under Congress' taxing power." State Plaintiffs' Br. at 32. The proper application of NFIB to the new version of the statute is to interpret it according to what Chief Justice Roberts—and four other Justices of the Court—said was the "most straightforward" reading of that provision: a command to purchase insurance. Id. at 562, 132 S.Ct. 2566 (Roberts, C.J.). As the district court properly observed, "the only reading available is the most natural one." Under that reading, the individual mandate is unconstitutional because, under NFIB, it finds no constitutional footing in either the Interstate Commerce Clause or the Necessary and Proper Clause. Id. at 546-61, 132 S.Ct. 2566 (Roberts, C.J.); id. at 650-60, 132 S.Ct. 2566 (joint dissent).

The intervenor-defendant states have several arguments against this conclusion, all of which fail. They first argue that the saving construction of the individual mandate, interpreting the provision as an option to buy insurance or pay a tax, is still "fairly possible." As the individual plaintiffs point out, the Court interpreted the individual mandate as an option only because doing so would save it from being unconstitutional. Accordingly, the intervenor-defendant states must show that the "option" would still be a constitutional exercise of Congress' taxing power. To make that showing, the intervenor-defendant states reject the plaintiffs' attempt to read a "some revenue" requirement into the Constitution's Taxing and Spending Clause, arguing instead for a potential-to-produce-revenue requirement. The individual mandate, they say, is still set out in the Internal Revenue Code. It still provides a "statutory structure through which" Congress could eventually tax people for failing to buy insurance. It still includes references to taxable income, number of dependents, and joint filing status. 26 U.S.C. §§ 5000A(b)(3), (c)(2), (c)(4). Further, it still does not apply to individuals who pay no federal income taxes. 26 U.S.C. § 5000A(e)(2).

The intervenor-defendant states have little support for this reading of the Taxing and Spending Clause. For starters, NFIB could not be clearer that the "produc[tion]" of "at least some revenue for the Government"—not the potential to produce that revenue—is "the essential feature of any tax." 567 U.S. at 564, 132 S.Ct. 2566 (majority opinion) (emphasis added). As the district court observed, when determining whether a statute is a tax, the actual production of revenue is "not indicative, not common—[but] essential."

The intervenor-defendant states also find no support in United States v. Ardoin, 19 F.3d 177, 179-80 (5th Cir. 1994). In that unusual case, Congress had imposed a tax on machine guns, but subsequently out-lawed machine guns altogether, which prompted the relevant agency to stop collecting the tax. Id. at 179-80. The defendant was convicted not only for possessing a machine gun but also for failing to pay the tax, which remained on the books. Id. at 178. The court upheld the conviction on the basis that the tax law at issue could "be upheld on the preserved, but unused, power to tax or on the power to regulate interstate commerce." Id. at 180. But the taxing power was "preserved" in Ardoin because it was non-revenue-producing only in practice whereas the "tax" here is actually $0.00 as written on the books.35 See Fed. Defendants' Br. at 32. Expanding Ardoin to apply here would, as the federal defendants point out, puzzlingly allow Congress to "prohibit conduct that exceeds its commerce power through a two-step process of first taxing it and then eliminating the tax while retaining the prohibition." Fed. Defendants' Br. at 32.

The intervenor-defendant states argue further that the individual mandate does not even need constitutional justification because it is merely a suggestion, not binding legislative action. The individual mandate, they contend, is no different from the Flag Code, which, though entered into the pages of the U.S. Code, "was not intended to proscribe conduct." Dimmitt v. City of Clearwater, 985 F.2d 1565, 1573 (11th Cir. 1993) (analyzing 36 U.S.C. §§ 174-76). This argument is just a repackaged version of their argument that the individual mandate can still be read as an option. But, as the state plaintiffs, the individual plaintiffs, and the federal defendants point out, the Supreme Court has already held that the "most straightforward" reading of the individual mandate— which emphatically demands that individuals "shall" buy insurance, 26 U.S.C. § 5000A(a)—is as a command to purchase health insurance. The Court then concluded that that command lacked constitutional justification. The zeroing out of the shared responsibility payment does not render the provision any less of a command. Quite the opposite: Chief Justice Roberts concluded that the greater-than-zero shared responsibility payment actually converted the individual mandate into an option. NFIB, 567 U.S. at 563-64, 132 S.Ct. 2566 (majority opinion). Now that the shared responsibility payment has been zeroed out, the only logical conclusion under NFIB is to read the individual mandate as a command, quite unlike the Flag Code. It is an individual mandate, not an individual suggestion.

Moreover, it is not true that when the Court adopts a limiting construction to avoid constitutional questions, that construction controls as to all applications of the statute, regardless of whether the original constitutional implications are present. The case on which the U.S. House relies involved different applications of an identical statute to different facts. Clark v. Martinez, 543 U.S. 371, 380, 125 S.Ct. 716, 160 L.Ed.2d 734 (2005) (rejecting the argument that "the constitutional concerns that influenced" a previous interpretation of a provision of the Immigration and Nationality Act were "not present for" the aliens at issue in that case). This case is readily distinguishable because the four characteristics that made the previous interpretation possible—the production of revenue and other tax-like features—have now been legislatively removed. The limiting construction is no longer available as a matter of statutory interpretation. The interpretation must accordingly change to comport with what five Justices of the Supreme Court have said is the "most straightforward reading" of that interpretation.36

The dissenting opinion justifies its continued reliance on the saving construction—even though it is no longer applicable—by citing Kimble v. Marvel Entm't, LLC, ___ U.S. ___, 135 S.Ct. 2401, 192 L.Ed.2d 463 (2015). This approach fares no better. The dissenting opinion quotes Kimble to say that "in whatever way reasoned," the Court's interpretation "effectively become[s] part of the statutory scheme, subject ... to congressional change." Id. at 2409. The dissenting opinion correctly acknowledges that the individual mandate was never changed. But what did change was the provision that actually mattered: the shared responsibility payment. When it was set above zero, it could be saved as a tax, even though five justices agreed this was an unnatural reading. It would be puzzling if Congress could change a statute at will, entirely insulated from constitutional infirmity, just because the Court had previously used constitutional avoidance to save a previous version of the statute.

The intervenor-defendant states argue furthermore that the individual mandate can now be constitutional under the Interstate Commerce Clause because it does not compel anyone into commerce. This is again a repackaged version of their argument that the individual mandate is an option even without a revenue-generating shared responsibility payment, an argument that, as the state plaintiffs point out, the Supreme Court has already rejected. This argument, as the district court observed, is also logically inconsistent. If the individual mandate no longer truly compels anything, then it can hardly be said to be a "regulat[ion]" of interstate commerce. In the words of the district court, the intervenor-defendant states "hope to have their cake and eat it too."37

Finally, we would be remiss if we did not engage with the dissenting opinion's contention that § 5000A is not an exercise of legislative power. This would likely come as a shock to the legislature that drafted it, the president who signed it, and the voters who celebrated or lamented it. It is not surprising that the dissenting opinion can cite no case in which a federal court deems a duly enacted statute not an exercise of legislative power, much less a statute that clearly commands that an individual "shall" do something.38 The dissenting opinion is inconsistent on this point: it argues that the provision's status as an exercise of legislative power fluctuates according to the amount of the shared responsibility payment while simultaneously contending that "if the text of the coverage requirement has not changed, its meaning could not have changed either." Our decision breaks no new ground. We simply observe that § 5000A was originally cognizable as either a command or a tax. Today, it is only cognizable as a command. It has always been an exercise of legislative power.

* * *

In NFIB, the individual mandate—most naturally read as a command to purchase insurance—was saved from unconstitutionality because it could be read together with the shared responsibility payment as an option to purchase insurance or pay a tax. It could be read this way because the shared responsibility payment produced revenue. It no longer does so. Therefore, the most straightforward reading applies: the mandate is a command. Using that meaning, the individual mandate is unconstitutional.

VI.

Having concluded that the individual mandate is unconstitutional, we must next determine whether, or how much of, the rest of the ACA is severable from that constitutional defect. On this question, we remand to the district court to undertake two tasks: to explain with more precision what provisions of the post-2017 ACA are indeed inseverable from the individual mandate; and to consider the federal defendants' newly-suggested relief of enjoining the enforcement only of those provisions that injure the plaintiffs or declaring the Act unconstitutional only as to the plaintiff states and the two individual plaintiffs. We address each issue in turn.

A.

The Supreme Court has said that the "standard for determining the severability of an unconstitutional provision is well established." Alaska Airlines, Inc. v. Brock, 480 U.S. 678, 684, 107 S.Ct. 1476, 94 L.Ed.2d 661 (1987). Unless it is "evident that the Legislature would not have enacted those provisions which are within its power, independently of that which is not, the invalid part may be dropped if what is left is fully operative as a law." Id. (quoting Buckley v. Valeo, 424 U.S. 1, 108, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976)).

This inquiry into counterfactual Congressional intent has been crystallized into a "two-part ... framework." NFIB, 567 U.S. at 692, 132 S.Ct. 2566 (joint dissent). First, if a court holds a statutory provision unconstitutional, it then determines whether the now-truncated statute will operate in "a manner consistent with the intent of Congress." Alaska Airlines, 480 U.S. at 685, 107 S.Ct. 1476 (emphasis omitted). This first step asks whether the constitutional provisions—standing on their own, without the unconstitutional provisions—are "fully operative as a law," not whether they would simply "operate in some coherent way" not designed by Congress. Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 509, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (quoting New York v. United States, 505 U.S. 144, 186, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992)); NFIB, 567 U.S. at 692, 132 S.Ct. 2566 (joint dissent). Second, even if the remaining provisions can operate as Congress designed them to, the court must determine if Congress would have enacted the remaining provisions without the unconstitutional portion. If Congress would not have done so, then those provisions must be deemed inseverable. Alaska Airlines, 480 U.S. at 685, 107 S.Ct. 1476 ("[T]he unconstitutional provision must be severed unless the statute created in its absence is legislation that Congress would not have enacted."); Free Enter. Fund, 561 U.S. at 509, 130 S.Ct. 3138 ("[N]othing in the statute's text or historical context makes it evident that Congress, faced with the limitations imposed by the Constitution, would have preferred no Board at all to a Board whose members are removable at will." (internal quotation marks omitted)).

Severability doctrine places courts between a rock and a hard place. On the one hand, courts strive to be faithful agents of Congress,39 which often means refusing to create a hole in a statute in a way that creates legislation Congress never would have agreed to or passed. See Murphy, 138 S. Ct. at 1482 ("[Courts] cannot rewrite a statute and give it an effect altogether different from that sought by the measure viewed as a whole." (quoting R.R. Ret. Bd. v. Alton R.R., 295 U.S. 330, 362, 55 S.Ct. 758, 79 S.Ct. 1468 (1935))). On the other hand, courts often try to abide by the medical practitioner's maxim of "first, do no harm," aiming "to limit the solution to the problem" by "refrain[ing] from invalidating more of the statute than is necessary." Ayotte v. Planned Parenthood of N. New England, 546 U.S. 320, 328, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006); Collins v. Mnuchin, 938 F.3d 553, 592 (5th Cir. 2019) (en banc) (Haynes, J.) (severing unconstitutional removal restriction from remainder of Federal Housing Finance Agency's enabling statute).40 In fact, courts have a "duty" to "maintain the act in so far as it is valid" if it "contains unobjectionable provisions separable from those found to be unconstitutional." Alaska Airlines, 480 U.S. at 684, 107 S.Ct. 1476 (quoting Regan v. Time, Inc., 468 U.S. 641, 652, 104 S.Ct. 3262, 82 L.Ed.2d 487 (1984) (plurality opinion)).

The Supreme Court emphasizes this duty so strongly that commentators have identified "a presumption [of severability] implicit in the Court's" severability jurisprudence. Adrian Vermeule, Saving Constructions, 85 Geo. L.J. 1945, 1950 n.28 (1997); see also Brian Charles Lea, Situational Severability, 103 Va. L. Rev. 735, 744 (2017) ("[C]ourts assume that a legislature intends for any unlawful part of its handiwork to be severable from all lawful parts in the absence of indicia of a contrary intention."). This presumption is strongest when Congress includes a severability clause in the statutory text; however, "[i]n the absence of a severability clause ... Congress's silence is just that— silence—and does not raise a presumption against severability." Alaska Airlines, 480 U.S. at 686, 107 S.Ct. 1476.

Nevertheless, the meticulous analysis required by severability doctrine defies reliance on presumptions or generalities. The Supreme Court's latest venture into severability territory, Murphy v. NCAA, ___ U.S. ___, 138 S.Ct. 1461, 200 L.Ed.2d 854 (2018), provides an example. There, the Court held that the entirety of the Professional and Amateur Sports Protection Act was unconstitutional because one of its provisions—authorizing private sports gambling—violated the anti-commandeering doctrine. Id. at 1484. Justice Alito's majority opinion separately explored each of the other operative provisions in the act, reasoning that all of the act's provisions were "obviously meant to work together" and be "deployed in tandem." Id. at 1483. Because Congress would not have wanted the otherwise-valid provisions "to stand alone," the Court declined to sever them. Id. This conclusion prompted a dissent from Justice Ginsburg, who characterized the majority as "wield[ing] an ax ... instead of using a scalpel to trim the statute" and reiterated that "the Court ordinarily engages in a salvage rather than a demolition operation." Id. at 1489-90 (Ginsburg, J., dissenting).

These Murphy opinions draw attention to one difficulty inherent in severability analysis: selecting the right tool for the job. Justice Thomas' concurring opinion goes further, providing two reasons why navigating between the Scylla of poking small but critical holes in complex, carefully crafted legislative bargains and the Charybdis of invalidating more duly enacted legislation than necessary stands "in tension with traditional limits on judicial authority." Murphy, 138 S. Ct. at 1485 (Thomas, J., concurring). "[T]he judicial power is, fundamentally, the power to render judgments in individual cases," and severability doctrine threatens to violate that vital separation-of-powers principle in more than one way. Id. (Thomas, J., concurring).

First, severability doctrine requires "a nebulous inquiry into hypothetical congressional intent," as opposed to the usual judicial bread-and-butter of "determin[ing] what a statute means." Id. at 1486 (Thomas, J., concurring) (quoting United States v. Booker, 543 U.S. 220 at 321 n.7, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005) (Thomas, J., dissenting in part)). Because "Congress typically does not pass statutes with the expectation that some part will later be deemed unconstitutional," id. at 1487, this requirement often leaves courts to exercise their imagination or "intuitions regarding what the legislature would have desired had it considered the severability issue." Lea, supra, at 747. This, in turn, "enmeshes the judiciary in making policy choices" the Constitution reserves for the legislature, David H. Gans, Severability as Judicial Lawmaking, 76 Geo. Wash. L. Rev. 639, 663 (2008), providing unelected judicial officers with cover to simply implement their own policy preferences.

Second, severability doctrine forces courts to "weigh in on statutory provisions that no party has standing to challenge, bringing courts dangerously close to issuing advisory opinions." Murphy, 138 S. Ct. at 1487 (Thomas, J., concurring); see also Jonathan F. Mitchell, The Writ-of-Erasure Fallacy, 104 Va. L. Rev. 933, 936 (2018) ("The federal courts have no authority to erase a duly enacted law from the statute books, [but can only] decline to enforce a statute in a particular case or controversy."41). As Justice Thomas points out, when Chief Justice Marshall famously declared that "[i]t is emphatically the province and duty of the judicial department to say what the law is," he justified that assertion by explaining that "[t]hose who apply [a] rule to particular cases, must of necessity expound and interpret that rule." Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177, 2 S.Ct. 60 (1803). Yet severability doctrine directs courts to go beyond the necessary —that is, the application of a particular statutory provision to a particular case—to consider the viability of other provisions without even "ask[ing] whether the plaintiff has standing to challenge those other provisions." Murphy, 138 S. Ct. at 1487 (Thomas, J., concurring). "[S]everability doctrine is thus an unexplained exception to the normal rules of standing, as well as the separation-of-powers principles that those rules protect." Id.

Severability analysis is at its most demanding in the context of sprawling (and amended) statutory schemes like the one at issue here. The ACA's framework of economic regulations and incentives spans over 900 pages of legislative text and is divided into ten titles. Most of the provisions directly regulating health insurance, including the one challenged in this case, are found in Titles I and II. See, e.g., 26 U.S.C. § 5000A(a) (individual mandate); 42 U.S.C. § 300gg-14(a) (requiring insurers offering family plans to cover adult children until age 26), §§ 18031-18044 (creating health insurance exchanges). The other titles generally amend Medicare (Title III), fund preventative healthcare programs (Title IV), seek to expand the supply of healthcare workers (Title V), enact anti-fraud requirements for Medicare/Medicaid facilities (Title VI), establish or expand drug regulations (Title VII), create a voluntary long-term care insurance program (Title VIII), address taxation (Title IX), and improve health care for Native Americans (Title X42).

The plaintiffs group this host of provisions into three categories for ease of reference. State Plaintiffs' Br. at 38. The first category includes the three core ACA provisions the Supreme Court has called "closely intertwined": the individual mandate, 26 U.S.C. § 5000A(a), the guaranteed-issue requirement, 42 U.S.C. §§ 300gg, 300gg-1, and the community-rating requirement, 42 U.S.C. § 300gg-4. King, 135 S. Ct. at 2487. The second category includes the remaining "[m]ajor provisions of the Affordable Care Act," NFIB, 567 U.S. at 697, 132 S.Ct. 2566 (joint dissent), namely other provisions dealing with "insurance regulations and taxes," "reductions in federal reimbursements to hospitals and other Medicare spending reductions," the insurance "exchanges and their federal subsidies," and "the employer responsibility assessment." See, e.g., 25 U.S.C. § 4980H; 26 U.S.C. § 36B; 42 U.S.C. §§ 1395ww, 18021-22. The third category includes a variety of minor provisions, for example taxes on certain medical devices or provisions requiring the display of nutritional content at restaurants. See, e.g., 21 U.S.C. § 343(q)(5)(H); 26 U.S.C. § 4191(a).

Moreover, Congress has made a number of substantive amendments to the ACA, revising the statute in 2010, 2011, 2014, 2017, and 2018. See, e.g., Medicare and Medicaid Extenders Act of 2010, Pub. L. No. 111-309, 124 Stat. 3285 (2010) (modifying tax credit scale and Medicaid requirements); Department of Defense and Full-Year Continuing Appropriations Act, 2011, Pub. L. No. 112-10, 125 Stat. 38 (2011) (repealing program that required some employers to provide some employees with vouchers for purchasing insurance); Bipartisan Budget Act of 2015, Pub. L. No. 114-74, 129 Stat. 584 (2015) (repealing requirement that employers with more than 200 employees enroll new full-time employees in health insurance and continue coverage for current employees). Most of these amendments occurred prior to the 2017 legislation eliminating the shared responsibility payment, but some are more recent. See, e.g., Bipartisan Budget Act of 2018, Pub. L. No. 115-123, 132 Stat. 64 (2018) (repealing Independent Payment Advisory Board).

In summary, then, this issue involves a challenging legal doctrine applied to an extensive, complex, and oft-amended statutory scheme. All together, these observations highlight the need for a careful, granular approach to carrying out the inherently difficult task of severability analysis in the specific context of this case. We are not persuaded that the approach to the severability question set out in the district court opinion satisfies that need. The district court opinion does not explain with precision how particular portions of the ACA as it exists post-2017 rise or fall on the constitutionality of the individual mandate. Instead, the opinion focuses on the 2010 Congress' labeling of the individual mandate as "essential" to its goal of "creating effective health insurance markets," 42 U.S.C. § 18091(2)(I), and then proceeds to designate the entire ACA inseverable. In using this approach, the opinion does not address the ACA's provisions with specificity, nor does it discuss how the individual mandate fits within the post-2017 regulatory scheme of the ACA.

The district court opinion begins by addressing the 2010 version of the ACA. Starting with the text of the ACA, the district court opinion points out that the 2010 Congress incorporated into the text its view that "the absence of the [individual mandate] would undercut Federal regulation of the health insurance market." 42 U.S.C. § 18091(2)(H). The district court opinion notes that the 2010 Congress devised the individual mandate, "together with the other provisions" of the ACA, to "add millions of new customers to the health insurance market." 42 U.S.C. § 18091(2)(C). In this way, the 2010 Congress sought to "minimize th[e] adverse selection" that might otherwise occur if healthy individuals "wait[ed] to purchase health insurance until they needed care," 42 U.S.C. § 18091(2)(I)—a strategic choice that would otherwise be available given the ACA's guaranteed-issue and community-rating provisions. According to the district court opinion: because the 2010 Congress found the individuate mandate "essential" to this plan to reshape health insurance markets, the individual mandate is inseverable from the rest of the ACA "[o]n the unambiguous enacted text alone."

The district court opinion also addresses ACA caselaw. Citing the Supreme Court's decisions in NFIB and King, the district court opinion states that "[a]ll nine Justices... agreed the Individual Mandate is inseverable from at least the pre-existing-condition provisions." See NFIB, 567 U.S. at 548, 132 S.Ct. 2566 (Roberts, C.J.), 596-98, 132 S.Ct. 2566 (Ginsburg, J., joined by Breyer, Kagan, and Sotomayor, JJ.), 695-96, 132 S.Ct. 2566 (joint dissent of Scalia, Kennedy, Thomas, and Alito, JJ.); King, 135 S. Ct. at 2487 (stating that the individual mandate is "closely intertwined" with the guaranteed-issue and community-rating provisions). As to the ACA's other provisions, the district court opinion notes that the only group of Justices who fully considered whether the other major and minor provisions were severable was the joint dissent in NFIB—and those Justices would have held that "invalidation of the ACA's major provisions requires the Court to invalidate the ACA's other provisions." NFIB, 567 U.S. at 704, 132 S.Ct. 2566 (joint dissent).

Beyond these points, the district court opinion states that its "conclusion would only be reinforced" if it "parse[d] the ACA's provisions one by one." The district court opinion arrives at this conclusion by reasoning that declaring only the individual mandate unlawful would disrupt the Act's careful balance of "shared responsibility." The district court opinion lists a few examples of how it would expect this to happen with regard to the ACA's major provisions. First, the district court opinion reasons that "the Individual Mandate reduces the financial risk forced upon insurance companies and their customers by the ACA's major regulations and taxes." If the individual mandate fell and the regulations and taxes did not, insurance companies would suffer a burden without enjoying a countervailing benefit—"a choice no Congress made and one contrary to the text." Second, if a court were to declare just the individual mandate and the protections for preexisting conditions unlawful—but not the subsidies for health insurance—then the Act would be transformed into "a law that subsidizes the kinds of discriminatory products Congress sought to abolish at, presumably, the re-inflated prices it sought to suppress." Third, Congress never intended "a duty on employers, see 26 U.S.C. § 4980H, to cover the skyrocketing insurance premium costs" that would "inevitably result from removing" the individual mandate. Fourth, because "the Medicaid-expansion provisions were designed to serve and assist fulfillment of the Individual Mandate," removing the individual mandate would remove the need for that expansion.

As to the ACA's minor provisions, the district court opinion states that it is "impossible to know which minor provisions Congress would have passed absent the Individual Mandate," and that such an inquiry involves too much "legislative guesswork." Relying on the 2010 Congress' labeling of the individual mandate as "essential," the district court opinion ultimately determines that there is "no reason to believe that Congress would have enacted" the minor provisions independently. The district court opinion similarly disclaims the ability to divine the intent of the 2017 Congress—which had zeroed out the shared responsibility payment but left the rest of the ACA untouched—labeling such an inquiry "a fool's errand." To the extent it analyzed the intent of the 2017 Congress, the district court opinion determines that Congress' failure to repeal the individual mandate shows that it "knew that provision is essential to the ACA." In sum, the district court opinion concludes that the entire ACA is inseverable from the individual mandate.

The plaintiffs urge affirmance for essentially the same reasons stated in the district court opinion.43 As to the guaranteed-issue and community-rating provisions, they rely primarily on the 2010 Congress' express findings linking those provisions to the individual mandate. State Plaintiffs' Br. at 39-44; Individual Plaintiffs' Br. at 47-48. The 2010 Congress found that, without the individual mandate, "many individuals would wait to purchase health insurance until they needed care," creating an "adverse selection" problem. 42 U.S.C. § 18091(2)(I); see also id. (finding that the individual mandate is "essential to creating effective health insurance markets in which improved health insurance products that are guaranteed issue and do not exclude coverage of pre-existing conditions can be sold"). As to the remaining major and some of the minor provisions, the plaintiffs rely primarily on the joint dissent in NFIB for the proposition that leaving these provisions standing would "undermine Congress' scheme of shared responsibility," throwing off the balance interlocking insurance market reforms set out in the ACA. 567 U.S. at 698, 132 S.Ct. 2566 (joint dissent) (internal quotation marks omitted); State Plaintiffs' Br. at 44-49. As for the most minor provisions, they argue that these were "mere adjuncts" of the more important provisions and would not have been independently enacted. State Plaintiffs' Br. at 50.

On appeal, the federal defendants agree with the plaintiffs that the entirety of the ACA is inseverable from the individual mandate. Fed. Defendants' Br. at 36-49. This marks a significant change in litigation position, as the federal defendants had previously submitted to the district court that only the guaranteed-issue and community-rating provisions were inseverable. And that is not the only new argument the federal defendants make on appeal. For the first time on appeal, the federal defendants argue that the remedy in this case should be limited to enjoining enforcement of the ACA only to the extent it harms the plaintiffs. See Fed. Defendants' Br. at 26-29 (arguing that the individual "plaintiffs do not have standing to seek relief against provisions of the ACA that do not in any way affect them"); Fed. Defendants' Supp. Br. at 10 ("[T]he judgment itself, as opposed to its underlying legal reasoning, cannot be understood as extending beyond the plaintiff states to invalidate the ACA in the intervenor states.").

The intervenor-defendant states, meanwhile, argue that every provision of the ACA is severable from the individual mandate. They argue that the 2017 Congress' decision not to repeal or otherwise undermine any other provision of the ACA shows that it intended the rest of the ACA to remain operative—and that the court should not focus on the intent of the 2010 Congress. Intervenor-Defendant States' Br. at 34-35, 43. They point to the statements of several legislators in the 2017 Congress that seem to evince an assumption that other parts of the ACA would not be altered,44 and to Congress' knowledge of reports highlighting the severe consequences a total invalidation of the ACA would have. Intervenor-Defendant States' Br. at 40. Finally, they argue that the passage of time since the ACA's enactment has shown that the individual mandate is not all that crucial after all, and they provide examples of ACA provisions they say have nothing to do with insurance markets or became operative years before the individual mandate took effect. Intervenor-Defendant States' Br. at 45.

Although we understand and share the district court's general disinclination to engage in what it refers to as "legislative guesswork"—and what a Supreme Court Justice has described as "a nebulous inquiry into hypothetical congressional intent," Murphy, 138 S. Ct. at 1486 (Thomas, J., concurring) (quoting Booker, 543 U.S. at 321 n.7, 125 S.Ct. 738 (Thomas, J., dissenting in part))—we nevertheless conclude that the severability analysis in the district court opinion is incomplete in two ways.

First, the opinion gives relatively little attention to the intent of the 2017 Congress, which appears in the analysis only as an afterthought despite the fact that the 2017 Congress had the benefit of hindsight over the 2010 Congress: it was able to observe the ACA's actual implementation. Although the district court opinion states that burdening insurance companies with taxes and regulations without giving them the benefit of compelling the purchase of their product is "a choice no Congress made," it only links this observation to the 2010 Congress. It does not explain its statement that the 2017 Congress' failure to repeal the individual mandate is evidence of an understanding that no part of the ACA could survive without it.

Second, the district court opinion does not do the necessary legwork of parsing through the over 900 pages of the post-2017 ACA, explaining how particular segments are inextricably linked to the individual mandate. The opinion lists a few examples of major provisions and cogently explains their link to the individual mandate, at least as it existed in 2010. For example, the opinion discusses the individual mandate's interplay with the guaranteed-issue and community-rating provisions—all of which are found in Title I of the ACA—analyzing how Congress intended those provisions to work and how they might be expected to work without the individual mandate. But in order to strike the delicate balance that severability analysis requires, the district court must undertake a similar inquiry for each segment of the post-2017 law that it ultimately declares unlawful—and it has not done so. Instead, the district court opinion focuses on the 2010 Congress' designation of the individual mandate as "essential to creating effective health insurance markets" and intention that, for at least one set of legislative goals, the individual mandate was intended to work "together with the other provisions" of the ACA. E.g., 42 U.S.C. § 18091(2)(I). On this basis, and on the views of the dissenting Justices in NFIB addressing the ACA as it stood in 2012, the district court opinion renders the entire ACA inoperative. More is needed to justify the district court's remedy.

Take, for example, the ACA provisions in Title IV requiring certain chain restaurants to disclose to consumers nutritional information like "the number of calories contained in the standard menu item." Patient Protection and Affordable Care Act, Pub. L. No. 111-148, § 4206, 124 Stat. 119, 573-74 (2012) (codified at 21 U.S.C. § 343). Or consider the provisions in Title X establishing the level of scienter necessary to be convicted of healthcare fraud. Patient Protection and Affordable Care Act § 10606, 124 Stat. 119, 1006-09, (codified at 18 U.S.C. § 1347). Without more detailed analysis from the district court opinion, it is unclear how provisions like these— which certainly do not directly regulate the health insurance marketplace—were intended to work "together" with the individual mandate. Similarly, the district court opinion's assertion that "most of the minor provisions" of the ACA "are mere adjuncts of" or "aids to the[ ] effective execution" of the project of the individual mandate is not supported by the actual analysis in the district court opinion, which does not dive into those provisions. Finally, some insurance-related reforms became law years before the effective date of the individual mandate; the district court opinion does not explain how provisions like these are inextricably linked to the individual mandate. See, e.g., 42 U.S.C. §§ 300gg-11, 300gg-14(a). Whatever the solution to the problem of "legislative guesswork" the district court opinion identifies in severability doctrine as it currently stands, it must include a careful parsing of the statutory scheme at issue to address questions like these.

We have long "require[d] that a district court explain its reasons for granting a motion for summary judgment in sufficient detail for us to determine whether the court correctly applied the appropriate legal test." Wildbur v. ARCO Chem. Co., 974 F.2d 631, 644 (5th Cir. 1992). This is because we have "little opportunity for effective review" when the district court opinion leaves some reasoning "vague" or "unsaid." Myers v. Gulf Oil Corp., 731 F.2d 281, 284 (1984). "In such cases, we have not hesitated to remand...." Id. In this case, the analysis the district court opinion provides is substantial and far exceeds the sort of cursory reasoning that normally prompts us to remand. Yet, the vast, wide-ranging statutory scheme at issue in this case also far exceeds the comparatively small number of provisions at issue in other severability cases, see, e.g., Chadha, 462 U.S. at 931-35, 103 S.Ct. 2764 (considering whether 8 U.S.C. § 244(c)(2) could be severed from the rest of § 244)— especially cases in which entire legislative acts are determined to be inseverable, see, e.g., Murphy, 138 S. Ct. at 1481-84 (considering whether part of 28 U.S.C. § 3702(1) could be severed from §§ 3701-04).

Moreover, the Supreme Court has remanded in the severability context upon a determination that additional analysis was necessary. In Ayotte v. Planned Parenthood of Northern New England, 546 U.S. 320, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006), the Supreme Court took up the issue of what relief was appropriate upon a determination that a New Hampshire provision requiring parental notification prior to abortion was unconstitutional in some applications. Id. at 328-32, 126 S.Ct. 961. The Supreme Court determined that, although the district court's choice to use "the most blunt remedy"—total inseverability—was "understandable" under its own precedent, more analysis was needed to determine "whether New Hampshire's legislature intended the statute to be susceptible to" severability. Id. at 330-31, 126 S.Ct. 961. As a result, the Supreme Court remanded for "lower courts to determine legislative intent in the first instance." Id.

We do the same here, directing the district court to employ a finer-toothed comb on remand and conduct a more searching inquiry into which provisions of the ACA Congress intended to be inseverable from the individual mandate. We do not hold forth on just how fine-toothed that comb should be—the district court may use its best judgment to determine how best to break the ACA down into constituent groupings, segments, or provisions to be analyzed. Nor do we make any comment on whether the district court should take into account the government's new posture on appeal or what the ultimate outcome of the severability analysis should be.45 Although "we cannot affirm the order as it is presently supported," we do not suggest what result will be merited "[a]fter a more thorough inquiry." Unger v. Amedisys Inc., 401 F.3d 316, 325 (5th Cir. 2005). We only note that the inquiry must be made, and that the district court—which has many tools at its disposal —is best positioned to determine in the first instance whether the ACA "remains `fully operative as a law'" and whether it is evident from "the statute's text or historical context" that Congress would have preferred no ACA at all to an ACA without the individual mandate. Free Enter. Fund, 561 U.S. at 509, 130 S.Ct. 3138 (quoting New York, 505 U.S. at 186, 112 S.Ct. 2408).

It may still be that none of the ACA is severable from the individual mandate, even after this inquiry is concluded. It may be that all of the ACA is severable from the individual mandate. It may also be that some of the ACA is severable from the individual mandate, and some is not.46 But it is no small thing for unelected, life-tenured judges to declare duly enacted legislation passed by the elected representatives of the American people unconstitutional. The rule of law demands a careful, precise explanation of whether the provisions of the ACA are affected by the unconstitutionality of the individual mandate as it exists today.

B.

Remand is appropriate in this case for a second reason: so that the district court may consider the federal defendants' new arguments as to the proper scope of relief in this case. The relief the plaintiffs sought in the district court was a universal nationwide injunction: an order that totally "enjoin[ed] Defendants from enforcing the Affordable Care Act and its associated regulations." Before the district court, the federal defendants urged entry of a declaratory judgment stating that the guaranteed-issue and community-rating provisions —at that time, the only provisions the federal defendants argued were inseverable —were "invalid[ated]" by the zeroing out of the shared responsibility payment. This would be "sufficient relief against the Government," the federal defendants argued, because a declaratory judgment would "operate[ ] in a similar manner as an injunction" against the federal government, which would be "presumed to comply with the law" once the court provides "a definitive interpretation of the statute."

Ultimately, of course, the district court opinion determined that no ACA provision was severable and resulted in a judgment declaring the entire ACA "invalid." On appeal, the federal defendants first changed their litigation position to agree that no ACA provision was severable. Now they have changed their litigation position to argue that relief in this case should be tailored to enjoin enforcement of the ACA in only the plaintiff states—and not just that, but that the declaratory judgment should only reach ACA provisions that injure the plaintiffs. They argue that the Supreme Court has made clear that "[a] plaintiff's remedy must be tailored to redress the plaintiff's particular injury." Gill v. Whitford, ___ U.S. ___, 138 S.Ct. 1916, 1934, 201 L.Ed.2d 313 (2018); see also Printz v. United States, 521 U.S. 898, 935, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997) (reasoning that the Court has "no business answering" questions dealing with enforcement of provisions that "burden ... no plaintiff"); see also Murphy, 138 S. Ct. at 1485-86 (Thomas, J., concurring). This argument came as a surprise to the plaintiffs, who explained at oral argument that they saw the government's new position as a possible "bait and switch." The federal defendants admitted at oral argument that they had raised the scope-of-relief issue on appeal "for the first time," but argued that it was necessary to address, as it went to the district court's Article III jurisdiction. The federal defendants therefore suggested that it "would be appropriate to remand to consider the scope of the judgment."

The court agrees that remand is appropriate for the district court to consider these new arguments in the first instance. The district court did not have the benefit of considering them when it crafted the relief now on appeal.47 On remand, the district court—which is in a far better position than this court to determine which ACA provisions actually injure the plaintiffs —may consider the federal defendants' position on the proper relief to be afforded. As part of this inquiry, the district court may consider whether the federal defendants' arguments were timely raised, and whether limiting the remedy in this case is supported by Supreme Court precedent. Once again, we place no thumb on the scale as to the ultimate outcome; the district court is free to weigh the federal defendants' changed arguments as it sees fit.

VII.

For these reasons, the judgment of the district court is AFFIRMED in part and VACATED in part. We REMAND for proceedings consistent with this opinion.

KING, Circuit Judge, dissenting:

Any American can choose not to purchase health insurance without legal consequence. Before January 1, 2018, individuals had to choose between complying with the Affordable Care Act's coverage requirement or making a payment to the IRS. For better or worse, Congress has now set that payment at $0. Without any enforcement mechanism to speak of, questions about the legality of the individual "mandate" are purely academic, and people can purchase insurance—or not—as they please. No more need be said; it has long been settled that the federal courts deal in cases and controversies, not academic curiosities.

The majority sees things differently and today holds that an unenforceable law is also unconstitutional. If the majority had stopped there, I would be confident its extrajurisdictional musings would ultimately prove harmless. What does it matter if the coverage requirement is unenforceable by congressional design or constitutional demand? Either way, that law does not do anything or bind anyone.

But again, the majority disagrees. It feels bound to ask whether Congress would want the rest of the Affordable Care Act to remain in force now that the coverage requirement is unenforceable. Answering that question should be easy, since Congress removed the coverage requirement's only enforcement mechanism but left the rest of the Affordable Care Act in place. It is difficult to imagine a plainer indication that Congress considered the coverage requirement entirely dispensable and, hence, severable. And yet, the majority is unwilling to resolve the severability issue. Instead, it merely identifies serious flaws in the district court's analysis and remands for a do-over, which will unnecessarily prolong this litigation and the concomitant uncertainty over the future of the healthcare sector.

I would vacate the district court's order because none of the plaintiffs have standing to challenge the coverage requirement. And although I would not reach the merits or remedial issues, if I did, I would conclude that the coverage requirement is constitutional, albeit unenforceable, and entirely severable from the remainder of the Affordable Care Act.

I.

To my mind, this case begins and ought to end with the Supreme Court's decision in National Federation of Independent Business v. Sebelius, 567 U.S. 519, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012). In that case, the Court held that the coverage requirement would be unconstitutional if it were a legal command, because neither the Commerce Clause nor the Necessary and Proper Clause allows Congress to compel individuals to engage in commerce by purchasing health insurance. See NFIB, 567 U.S. at 552, 560, 132 S.Ct. 2566 (opinion of Roberts, C.J.); id. at 652-53, 132 S.Ct. 2566 (joint dissent). The Court concluded, however, that the coverage requirement was constitutional, because—notwithstanding the most natural reading of the provision's text—the coverage requirement was not actually a legal command to purchase insurance.

Instead, according to the NFIB Court, the coverage requirement "leaves an individual with a lawful choice to do or not do a certain act," i.e., purchase health insurance. Id. at 574, 132 S.Ct. 2566 (Roberts, C.J., majority opinion). All that is required, under this reading, is "a payment to the IRS" if one chooses not to purchase health insurance. Id. at 567, 132 S.Ct. 2566. Beyond this shared-responsibility payment, there are no further "negative legal consequences to not buying health insurance," and individuals who forgo insurance do not violate the law as long as they make the required payment. Id. at 567, 132 S.Ct. 2566. "Those subject to the [coverage requirement] may lawfully forgo health insurance and pay higher taxes, or buy health insurance and pay lower taxes. The only thing they may not lawfully do is not buy health insurance and not pay the resulting tax." Id. at 574, 132 S.Ct. 2566 n.11. Forcing individuals to make that choice was constitutional, per NFIB, because Congress could "impose a tax on not obtaining health insurance" by exercising its enumerated power to lay and collect taxes, duties, imposts, and excises. Id. at 570, 132 S.Ct. 2566.

Contrary to the suggestion of the majority, which I address specifically infra at Part III, Congress did not alter the coverage requirement's operation when it amended the ACA in 2017. See Tax Cuts and Jobs Act of 2017, Pub. L. No. 115-97, § 11081, 131 Stat. 2054, 2092 ("TCJA"). All the TCJA did, with respect to healthcare, was change the amount of the shared-responsibility payment to zero dollars. Thus, despite textual appearances, the post-TCJA coverage requirement does nothing more than require individuals to pay zero dollars to the IRS if they do not purchase health insurance, which is to say it does nothing at all.

This insight, that the coverage requirement now does nothing, should be the end of this case. Nobody has standing to challenge a law that does nothing. When Congress does nothing, no matter the form that nothing takes, it does not exceed its enumerated powers. And since courts do not change anything when they invalidate a law that does nothing, every other law retains, or at least should retain, its full force and effect.

II.

But as the majority goes well past NFIB, I respond. To begin, I emphasize the importance of the rule that a plaintiff must have standing to invoke a federal court's power. This is not an anachronism lingering from some era in which empty formalities abounded in legal practice. Quite the opposite: "[T]he requirement that a claimant have `standing is an essential and unchanging part of the case-or-controversy requirement of Article III.'" Davis v. FEC, 554 U.S. 724, 733, 128 S.Ct. 2759, 171 L.Ed.2d 737 (2008) (quoting Lujan v. Defs. of Wildlife, 504 U.S. 555, 560, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992)); see also Susan B. Anthony List v. Driehaus, 573 U.S. 149, 157, 134 S.Ct. 2334, 189 L.Ed.2d 246 (2014) ("Article III of the Constitution limits the jurisdiction of federal courts to `Cases' and `Controversies.'" (quoting U.S. Const. art. III, § 2)). And "[n]o principle is more fundamental to the judiciary's proper role in our system of government than the constitutional limitation of federal-court jurisdiction to actual cases or controversies." Clapper v. Amnesty Int'l USA, 568 U.S. 398, 408, 133 S.Ct. 1138, 185 L.Ed.2d 264 (2013) (alteration in original) (quoting DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 341, 126 S.Ct. 1854, 164 L.Ed.2d 589 (2006)); accord Raines v. Byrd, 521 U.S. 811, 818, 117 S.Ct. 2312, 138 L.Ed.2d 849 (1997).

The Constitution's case-or-controversy requirement reflects the Framers' view of the judiciary's place among the coequal branches of the federal government: to fulfill "the traditional role of Anglo-American courts, which is to redress or prevent actual or imminently threatened injury to persons caused by private or official violation of law." Summers v. Earth Island Inst., 555 U.S. 488, 492, 129 S.Ct. 1142, 173 L.Ed.2d 1 (2009). Strict adherence to the case-or-controversy requirement—and to standing in particular—thus "serves to prevent the judicial process from being used to usurp the powers of the political branches." Clapper, 568 U.S. at 408, 133 S.Ct. 1138; see also Town of Chester v. Laroe Estates, Inc., ___ U.S. ___, 137 S.Ct. 1645, 1650, 198 L.Ed.2d 64 (2017) ("This fundamental limitation preserves the `tripartite structure' of our Federal Government, prevents the Federal Judiciary from `intrud[ing] upon the powers given to the other branches,' and `confines the federal courts to a properly judicial role.'" (alteration in original) (quoting Spokeo, Inc. v. Robins, ___ U.S. ___, 136 S.Ct. 1540, 1547, 194 L.Ed.2d 635 (2016))). Thus, "federal courts may exercise power only `in the last resort, and as a necessity,' and only when adjudication is `consistent with a system of separated powers and [the dispute is one] traditionally thought to be capable of resolution through the judicial process.'" Allen v. Wright, 468 U.S. 737, 752, 104 S.Ct. 3315, 82 L.Ed.2d 556 (1984) (alteration in original) (citation omitted) (first quoting Chi. & Grand Trunk Ry. Co. v. Wellman, 143 U.S. 339, 345, 12 S.Ct. 400, 36 S.Ct. 176 (1892); then quoting Flast v. Cohen, 392 U.S. 83, 97, 88 S.Ct. 1942, 20 L.Ed.2d 947 (1968)), abrogated on other grounds, Lexmark Int'l, Inc. v. Static Control Components, Inc., 572 U.S. 118, 134 S.Ct. 1377, 188 L.Ed.2d 392 (2014). And needless to say, a federal court must conduct an "especially rigorous" standing inquiry "when reaching the merits of the dispute would force [it] to decide whether an action taken by one of the other two branches of the Federal Government was unconstitutional." Amnesty Int'l, 568 U.S. at 408, 133 S.Ct. 1138 (quoting Raines, 521 U.S. at 819-20, 117 S.Ct. 2312). "The importance of this precondition should not be underestimated as a means of `defin[ing] the role assigned to the judiciary in a tripartite allocation of power.'" Valley Forge Christian Coll. v. Ams. United for Separation of Church & State, 454 U.S. 464, 474, 102 S.Ct. 752, 70 L.Ed.2d 700 (1982) (alteration in original) (quoting Flast, 392 U.S. at 95, 88 S.Ct. 1942).

The standing doctrine polices this constitutional limit on the judiciary's power "by `identify[ing] those disputes which are appropriately resolved through the judicial process.'" Susan B. Anthony List, 573 U.S. at 157, 134 S.Ct. 2334 (alteration in original) (quoting Lujan, 504 U.S. at 560, 112 S.Ct. 2130). The party seeking redress in the courts has the burden to establish standing. See Spokeo, 136 S. Ct. at 1547. To do so, the plaintiff must show it has "(1) suffered an injury in fact, (2) that is fairly traceable to the challenged conduct of the defendant, and (3) that is likely to be redressed by a favorable judicial decision." Id. "To establish injury in fact, a plaintiff must show that he or she suffered `an invasion of a legally protected interest' that is `concrete and particularized' and `actual or imminent, not conjectural or hypothetical.'" Id. at 1548 (quoting Lujan, 504 U.S. at 560, 112 S.Ct. 2130). This means the injury must be "personal" to the plaintiff and, although the injury does not need to be "tangible," "it must actually exist." Id. at 1548-49.

The plaintiffs' evidentiary burden depends on the stage of the litigation. At each stage, the plaintiffs must demonstrate standing "with the manner and degree of evidence" otherwise required to establish the plaintiffs' merits case. Lujan, 504 U.S. at 561, 112 S.Ct. 2130. Thus, because this case comes to us on the plaintiffs' own motion for summary judgment, the plaintiffs must conclusively prove all three elements of standing with evidence that "would `entitle [them] to a directed verdict if the evidence went uncontroverted at trial.'" Int'l Shortstop, Inc. v. Rally's, Inc., 939 F.2d 1257, 1264-65 (5th Cir. 1991) (quoting Golden Rule Ins. Co. v. Lease, 755 F.Supp. 948, 951 (D. Colo. 1991)). If a plaintiff meets its burden, the defendant can nevertheless defeat summary judgment "by merely demonstrating the existence of a genuine dispute of material fact." Id. at 1265. In other words, the plaintiffs here must show that, considering the summary-judgment record, all reasonable factfinders would agree that the plaintiffs demonstrate an injury traceable to the coverage requirement and redressable by a favorable decision. See Alonso v. Westcoast Corp., 920 F.3d 878, 885-86 (5th Cir. 2019).

These general principles alone should make the majority's error apparent. More specific authority illuminates it. I explain first why the majority errs in concluding the individual plaintiffs have standing, then I explain why the majority errs in concluding the state plaintiffs have standing.

A.

The majority concludes that the individual plaintiffs have standing to challenge the coverage requirement in the Patient Protection and Affordable Care Act (the "ACA"), 26 U.S.C. § 5000A(a),1 because it forces them to purchase health insurance that they would not purchase otherwise. The majority overlooks what will happen if the individual plaintiffs fail to purchase insurance: absolutely nothing. The individual plaintiffs will be no worse off by any conceivable measure if they choose not to purchase health insurance. Thus, whatever injury the individual plaintiffs have incurred by purchasing health insurance is entirely self-inflicted.

A long line of cases establishes that self-inflicted injuries cannot establish standing because a self-inflicted injury, by definition, is not traceable to the challenged action. See, e.g., Amnesty Int'l, 568 U.S. at 416, 133 S.Ct. 1138 ("[R]espondents cannot manufacture standing merely by inflicting harm on themselves...."); Pennsylvania v. New Jersey, 426 U.S. 660, 664, 96 S.Ct. 2333, 49 L.Ed.2d 124 (1976) ("The injuries to the plaintiffs' fiscs were self-inflicted, resulting from decisions by their respective state legislatures.... No State can be heard to complain about damage inflicted by its own hand."); Zimmerman v. City of Austin, 881 F.3d 378, 389 (5th Cir.) ("[S]tanding cannot be conferred by a self-inflicted injury."), cert. denied, ___ U.S. ___, 139 S.Ct. 639, 202 L.Ed.2d 492 (2018). When a plaintiff chooses to incur an expense, the plaintiff must show that the challenged law forced the plaintiff to incur that expense to avoid some other concrete injury. See Amnesty Int'l, 568 U.S. at 415-16, 133 S.Ct. 1138 (concluding costs plaintiffs incurred trying to avoid surveillance were self-inflicted because plaintiffs' fear of surveillance was speculative); Contender Farms, L.L.P. v. USDA, 779 F.3d 258, 266 (5th Cir. 2015) (finding plaintiff had standing to challenge regulations that required plaintiff to either "take additional measures" to comply with regulation or "face harsher, mandatory penalties" and prosecution). In other words, a plaintiff can show standing if the challenged act placed him between the proverbial rock and hard place. But without showing such a dilemma, a plaintiff "cannot manufacture standing" by expending costs to avoid an otherwise noncognizable injury, which is exactly what the individual plaintiffs did here. Amnesty Int'l, 568 U.S. at 416, 133 S.Ct. 1138.

The majority brushes off this authority by insisting—without explanation—that labeling the plaintiffs' injuries self-inflicted "assumes" that the coverage requirement does not act as a legal command to purchase insurance, which the majority refuses to question at the standing stage. The majority misunderstands the argument. Even accepting that the coverage requirement acts as a legal command, the individual plaintiffs are still free to disregard that command without legal consequence. Therefore, any injury they incur by freely choosing to obtain insurance is still self-inflicted.

Nor does it matter that to avoid inflicting injury upon themselves, the plaintiffs would have to violate an unenforceable statute. Plaintiffs may challenge a statute that requires them "to take significant and costly compliance measures or risk criminal prosecution." Virginia v. Am. Booksellers Ass'n, 484 U.S. 383, 392, 108 S.Ct. 636, 98 L.Ed.2d 782 (1988) (emphasis added); see also, e.g., Int'l Tape Mfrs. Ass'n v. Gerstein, 494 F.2d 25, 28 (5th Cir. 1974) (explaining that standing to challenge a statute requires a "realistic possibility that the challenged statute will be enforced to [the plaintiff's] detriment"). But "[w]hen plaintiffs `do not claim that they have ever been threatened with prosecution, that a prosecution is likely, or even that a prosecution is remotely possible,' they do not allege a dispute susceptible to resolution by a federal court." Babbitt v. United Farm Workers Nat'l Union, 442 U.S. 289, 298-99, 99 S.Ct. 2301, 60 L.Ed.2d 895 (1979) (quoting Younger v. Harris, 401 U.S. 37, 42, 91 S.Ct. 746, 27 L.Ed.2d 669 (1971)); see also Poe v. Ullman, 367 U.S. 497, 507, 81 S.Ct. 1752, 6 L.Ed.2d 989 (1961) (Frankfurter, J., plurality) ("It is clear that the mere existence of a state penal statute would constitute insufficient grounds to support a federal court's adjudication of its constitutionality in proceedings brought against the State's prosecuting officials if real threat of enforcement is wanting."); cf. Zimmerman, 881 F.3d at 389-90 ("[T]o confer standing, allegations of chilled speech or `self-censorship must arise from a fear of prosecution that is not "imaginary or wholly speculative."'" (quoting Ctr. for Individual Freedom v. Carmouche, 449 F.3d 655, 660 (5th Cir. 2006))).

Ullman illustrates this principle well.2 The plaintiffs there sought to challenge Connecticut's criminal prohibition on contraception. Ullman, 367 U.S. at 498, 81 S.Ct. 1752 (Frankfurter, J., plurality). But in the more than 75 years that the statute had been on the books, only one violation had been prosecuted—and even that was a collusive prosecution brought to challenge the law. Id. at 501-02, 81 S.Ct. 1752. The Court dismissed the challenge for lack of standing, holding that "[t]he fact that Connecticut has not chosen to press the enforcement of this statute deprives these controversies of the immediacy which is an indispensable condition of constitutional adjudication." Id. at 508, 81 S.Ct. 1752. The Court explained that it could not "be umpire to debates concerning harmless, empty shadows." Id.3

Ullman makes this an easy case. Connecticut's contraception law at least allowed the possibility of enforcement, even if it was speculative and unlikely to ever occur. Here, as I cannot say often enough, the coverage requirement has no enforcement mechanism. It is impossible for the individual plaintiffs to ever be prosecuted (or face any other consequences) for violating it. In "find[ing] it necessary to pass on" the coverage requirement, the majority "close[s] [its] eyes to reality." Id.4

The majority does not engage with the lessons of Ullman and its progeny. The closest it comes is in its citation to Texas v. EEOC, 933 F.3d 433 (5th Cir. 2019). That case does not abrogate Ullman, Younger, Babbitt, American Booksellers, or Tape Manufacturers—nor could it. In Texas v. EEOC, Texas challenged EEOC administrative guidance stating that employers who screen out job applicants with criminal records could be held liable for disparate-impact discrimination. Id. at 437-38. The EEOC argued that Texas did not have standing to challenge the guidance because the guidance reflected only the EEOC's interpretation of Title VII, and the Attorney General, not the EEOC, has the sole power to enforce Title VII against states. See Brief for Appellants Cross-Appellees at 18-19, Texas v. EEOC, 933 F.3d 433 (5th Cir. 2019). In rejecting that argument, this court explained that Title VII's enforcement scheme is not so simple. Although the EEOC may not itself bring enforcement actions against states, it may investigate states and refer cases to the Attorney General for enforcement actions. EEOC, 933 F.3d at 447. Therefore, "the possibility of investigation by EEOC and referral to the Attorney General for enforcement proceedings if it fails to align its laws and policies with the Guidance" put pressure on Texas to conform to the EEOC's guidance. Id.

In other words, even absent a direct threat of a formal enforcement action from the EEOC, Texas faced other consequences for disobeying the guidance—including the possibility that the Attorney General would enforce Title VII against it. In fact, we noted that "[o]ne Texas agency ha[d] already been required to respond to a charge of discrimination filed with EEOC based on its no-felon hiring policy." Id. at 447 n.26. The majority here cites no similar concrete consequences that will (or even plausibly could) follow if the plaintiffs violate the coverage requirement.

My conclusion that individual plaintiffs lack standing is only bolstered by a unanimous opinion issued mere weeks ago by a panel that included the author of today's majority opinion. In that case, the court held that Austin, Texas could not use a suit against the Texas Attorney General to challenge a state statute, which the Attorney General was authorized to enforce, that barred the city from enforcing one of its ordinances. City of Austin v. Paxton, 943 F.3d 993, 1001-02 (5th Cir. 2019). Although the Paxton court based its holding on sovereign immunity, it looked to "our standing jurisprudence," and "note[d] that it's unlikely the City had standing," because it did not show that the Attorney General would likely "inflict `future harm'" by enforcing the statute against Austin. Id. at 1001-04. If standing was absent in Paxton because enforcement was insufficiently probable, I have no idea why standing should be present in this case, where enforcement of the challenged portion of the ACA is altogether impossible.

In sum, even if the unenforceable coverage requirement must be read as a command to purchase health insurance, it does not harm the individual plaintiffs because they can disregard it without consequence. Binding precedent squarely establishes that plaintiffs may not sue in such circumstances —and with good reason. The great power of the judiciary should not be invoked to disrupt the work of the democratic branches when the plaintiffs can easily avoid injury on their own.5

B.

The majority's conclusion that the state plaintiffs have standing to challenge the coverage requirement fares no better. I would deny the state plaintiffs standing because there is no evidence in the record, much less conclusive evidence, to support the state plaintiffs' alleged injuries.

1.

The majority first concludes that the state plaintiffs have standing because it believes that the coverage requirement increases the number of state employees who enroll in the states' employee healthcare programs. And with more enrollees, the logic goes, the states as employers must file more forms with the IRS at a higher cost to the states.

The majority's biggest mistake is that it ignores the posture of this case: the defendants appeal from the district court's order granting summary judgment to the plaintiffs. Accordingly, the state plaintiffs face a tremendous evidentiary burden—they must produce evidence so conclusive of the coverage requirement's effect on their healthcare-administration costs that the evidence "would `entitle [them] to a directed verdict if the evidence went uncontroverted at trial.'" Int'l Shortstop, 939 F.2d at 1264-65 (quoting Golden Rule Ins., 755 F. Supp. at 951).6 And the state plaintiffs provided no evidence at all, never mind conclusive evidence, to support the dubious notion that even a single state employee enrolled in one of state plaintiffs' health insurance programs solely because of the unenforceable coverage requirement.7

The majority relies on affidavits from several of the state plaintiffs' healthcare administrators. But these affidavits only establish that the state plaintiffs incur costs complying with the IRS reporting requirements found in 26 U.S.C. §§ 6055(a) and 6056(a). And as the majority recognizes, these requirements are distinct from the coverage requirement. Accordingly, to trace the state plaintiffs' reporting burden to the coverage requirement, the majority must additionally show that at least some state employees have enrolled in employer-sponsored health insurance solely because of the unenforceable coverage requirement. The majority comes up empty at this step, pointing only to a conclusory statement from a South Dakota human-resources director claiming that the coverage requirement, not §§ 6055(a) and 6056(a), caused South Dakota to incur its reporting expenses. This will not do. See, e.g., Lujan v. Nat'l Wildlife Fed'n, 497 U.S. 871, 888, 110 S.Ct. 3177, 111 L.Ed.2d 695 (1990) ("The object of [summary judgment] is not to replace conclusory allegations of the complaint or answer with conclusory allegations of an affidavit."); Shaboon v. Duncan, 252 F.3d 722, 737 (5th Cir. 2001) ("[U]nsupported affidavits setting forth `ultimate or conclusory facts and conclusions of law' are insufficient to either support or defeat a motion for summary judgment." (alteration in original) (quoting Orthopedic & Sports Injury Clinic v. Wang Labs., Inc., 922 F.2d 220, 225 (5th Cir. 1991))).8

Citing Department of Commerce v. New York, ___ U.S. ___, 139 S.Ct. 2551, 204 L.Ed.2d 978 (2019), the majority argues the state plaintiffs can establish standing by "showing that third parties will likely react in predictable ways" to the coverage requirement. Id. at 2566. But the majority fails to explain why state employees who do not want health insurance would nevertheless predictably enroll in health insurance solely because an unenforceable statute, here the coverage requirement, directs them to do so. What the majority fails to mention in its discussion of Department of Commerce is that the "predictable" behavior at issue there was individuals "choosing to violate their legal duty to respond to the census." Id. at 2565 (emphasis added). Thus, Department of Commerce shows that people will predictably violate the law when sufficiently incentivized to do so. This directly contradicts the assumption undergirding much of the majority's analysis—that people tend to follow the law regardless of the incentives. And state employees who do not want to enroll in insurance have every incentive to violate the coverage requirement.9

2.

The majority similarly argues that the coverage requirement increases the number of individuals enrolled in the state plaintiffs' Medicaid programs. This argument fails for the same reason: the state plaintiffs produce no evidence—let alone conclusive evidence—showing that anyone has enrolled in their Medicaid programs solely because of the unenforceable coverage requirement. To this end, the best the majority can scrape up is a statement from Teresa MacCartney, a Georgia budget official, stating that "[a]fter the implementation of the ACA, [Georgia] experienced increased enrollment of individuals already eligible for Medicaid benefits under pre-ACA eligibility standards." The majority's takeaway is that the coverage requirement caused this increase. Maybe so. But MacCartney's statement refers specifically to the coverage requirement at the time of the ACA's enactment, when the coverage requirement interacted with the shared-responsibility payment. This statement provides no insight into how the coverage requirement affects Medicaid rolls after the shared-responsibility payment's repeal. In fact, MacCartney signed her declaration on May 14, 2018, more than seven months before the shared-responsibility payment's repeal went into effect. See Budget Fiscal Year, 2018, Pub. L. No. 115-97, § 11081(b), 131 Stat. 2054, 2092 (2017).

Accordingly, the majority's analysis again rests on the necessary assumption that people will obey the coverage requirement regardless of the incentives, in direct contradiction to Department of Commerce. And because Medicaid is available to eligible recipients at little to no cost, it is especially unlikely that the unenforceable coverage requirement would play any significant part in anyone's decision to enroll. It belies common sense to conclude that anyone who would otherwise pass on the significant benefits of Medicaid would be motivated to enroll solely because of an unenforceable law.

In sum, the majority cites no actual evidence tying any costs the state plaintiffs have incurred to the unenforceable coverage requirement. The state plaintiffs accordingly cannot show an injury traceable to the coverage requirement, so they do not have standing to challenge the coverage requirement.

III.

I would not reach the merits of this case because, as explained in Part II, I would vacate the district court's order for lack of standing. But as the majority errs on the merits too, I voice my disagreement.

"Neither the Act nor any other law attaches negative legal consequences to not buying health insurance, beyond requiring a payment to the IRS." NFIB, 567 U.S. at 568, 132 S.Ct. 2566 (Roberts, C.J., majority opinion). Now that Congress has zeroed out that payment, the coverage requirement affords individuals the same choice individuals have had since the dawn of private health insurance, either purchase insurance or else pay zero dollars. Thus, to my mind, the majority's focus on whether Congress's taxing power or the Necessary and Proper Clause authorizes Congress to pass a $0 tax is a red herring; the real question is whether Congress exceeds its enumerated powers when it passes a law that does nothing.10 And of course it does not.11 Congress exercises its legislative power when it "alter[s] the legal rights, duties and relations of persons." INS v. Chadha, 462 U.S. 919, 952, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983); cf. id. ("Not every action taken by either House is subject to the bicameralism and presentment requirements of Art. I. Whether actions taken by either House are, in law and fact, an exercise of legislative power depends not on their form but upon `whether they contain matter which is properly to be regarded as legislative in its character and effect.'" (citation omitted) (quoting S. Rep. No. 1335, 54th Cong., 2d Sess., 8 (1897))).

Lest the majority mistake my position and end up shadowboxing with "bizarre metaphysical conclusions," "quantum musings," or ersatz inconsistencies, Maj. Op. at 395 & n.40, I need to make something explicit at the outset. The TCJA did not change the text or the meaning of the coverage requirement, but it did change the real-world effects it produces. Before the TCJA, the two options afforded by the coverage requirement—purchasing insurance or making a shared-responsibility payment—were both burdensome, but Congress could force individuals to choose one of those options by exercising its Taxing Power. Today, the shared-responsibility payment's meaning has not changed—it still gives individuals the choice to purchase insurance or make a shared-responsibility payment—but the amount of that payment is zero dollars, which means that the coverage requirement now does nothing. The majority's contrary conclusion rests on the premise that the coverage requirement compels individuals to purchase health insurance. With this understanding, the majority says that the coverage requirement does exactly what the Supreme Court said it cannot do: compel participation in commerce. See NFIB, 567 U.S. at 552, 132 S.Ct. 2566 (opinion of Roberts, C.J.); id. at 652-53, 132 S.Ct. 2566 (joint dissent). This conclusion follows fine from the premise, but the premise is wrong. Despite its seemingly mandatory language, the coverage requirement does not compel anyone to purchase health insurance.

In NFIB, although five Justices agreed that "[t]he most straightforward reading of the [coverage requirement] is that it commands individuals to purchase insurance," id. at 562, 132 S.Ct. 2566 (opinion of Roberts, C.J.); accord id. at 663, 132 S.Ct. 2566 (joint dissent), applying the canon of constitutional avoidance, the Court rejected this interpretation. Instead, the Court interpreted the coverage requirement to offer applicable individuals a "lawful choice" between purchasing health insurance and paying the shared-responsibility payment, which the Court interpreted as a valid exercise of Congress's taxing power. Id. at 574, 132 S.Ct. 2566 (Roberts, C.J., majority opinion). This is a permissible construction, the Court concluded, because "[w]hile the [coverage requirement] clearly aims to induce the purchase of health insurance, it need not be read to declare that failing to do so is unlawful." Id. at 567-68, 132 S.Ct. 2566. The Court observed that "[n]either the [ACA] nor any other law attaches negative legal consequences to not buying health insurance, beyond requiring a payment to the IRS." Id. at 568, 132 S.Ct. 2566. And the Court further explained:

Id. (citation omitted).

The NFIB Court's application of constitutional avoidance as an interpretive tool does not mean that the Court rewrote the statute. Only Congress can do that. Rather, the Court was "choosing between competing plausible interpretations of a statutory text, resting on the reasonable presumption that Congress did not intend the alternative which raises serious constitutional doubts." Clark v. Martinez, 543 U.S. 371, 381, 125 S.Ct. 716, 160 L.Ed.2d 734 (2005). "The canon is thus a means of giving effect to congressional intent, not of subverting it." Id. at 382, 125 S.Ct. 716. Accordingly, when the Court ruled in NFIB that "[t]hose subject to the [coverage requirement] may lawfully forgo health insurance," NFIB, 567 U.S. at 574 n.11, 132 S.Ct. 2566, that was an authoritative determination regarding what the text of the coverage requirement meant and what Congress intended.

The majority pushes aside NFIB's construction, acting as though the fact that the NFIB Court applied the canon of constitutional avoidance means that its interpretation no longer governs following the repeal of the shared-responsibility payment. But when the Court construes statutes, its "interpretive decisions, in whatever way reasoned, effectively become part of the statutory scheme, subject (just like the rest) to congressional change." Kimble v. Marvel Entm't, LLC, ___ U.S. ___, 135 S.Ct. 2401, 2409, 192 L.Ed.2d 463 (2015) (emphasis added). While Congress can change its mind and could have amended the coverage requirement to turn the "lawful choice" described by NFIB, 567 U.S. at 574, 132 S.Ct. 2566, into an unwavering command, the majority does not suggest that Congress ever made such a choice. Sure, Congress amended the shared-responsibility payment in 2017. Yet as the district court went to great lengths to establish and the majority is elsewhere eager to point out, the coverage requirement and the shared-responsibility payment are distinct provisions. See Maj. Op. at 378 ("To bring a claim against the [coverage requirement], therefore, the plaintiffs needed to show injury from the individual mandate —not from the shared responsibility payment."); Texas v. United States, 340 F.Supp.3d 579, 596 (N.D. Tex. 2018) ("It is critical to clarify something at the outset: the shared-responsibility payment, 26 U.S.C. § 5000A(b), is distinct from the [coverage requirement], id. § 5000A(a)."). And Congress did not touch the text of the coverage requirement when it amended the shared-responsibility payment. See Budget Fiscal Year, 2018, Pub. L. No. 115-97, § 11081. Compare § 5000A(a), with 26 U.S.C. § 5000A(a) (2011). At risk of stating the obvious, if the text of the coverage requirement has not changed, its meaning could not have changed either. By "giv[ing] these same words a different meaning," the majority "invent[s] a statute rather than interpret[s] one." Clark, 543 U.S. at 378, 125 S.Ct. 716.

The majority is thus left on unsteady ground: amendment by implication, which "will not be presumed unless the legislature's intent is `clear and manifest.'" In re Lively, 717 F.3d 406, 410 (5th Cir. 2013) (quoting Nat'l Ass'n of Home Builders v. Defs. of Wildlife, 551 U.S. 644, 662, 127 S.Ct. 2518, 168 L.Ed.2d 467 (2007)); see also, e.g., Epic Sys. Corp v. Lewis, ___ U.S. ___, 138 S.Ct. 1612, 1624, 200 L.Ed.2d 889 (2018) ("[I]n approaching a claimed conflict, we come armed with the `stron[g] presum[ption]' that repeals by implication are `disfavored' and that `Congress will specifically address' preexisting law when it wishes to suspend its normal operations in a later statute." (second and third alterations in original) (quoting United States v. Fausto, 484 U.S. 439, 452-53, 108 S.Ct. 668, 98 L.Ed.2d 830 (1988))). This rule operates with equal force when a judicial construction previously illuminated the meaning of the purportedly amended statute. See TC Heartland LLC v. Kraft Foods Grp. Brands LLC, ___ U.S. ___, 137 S.Ct. 1514, 1520, 197 L.Ed.2d 816 (2017) ("When Congress intends to effect a change of [a statute's earlier judicial interpretation], it ordinarily provides a relatively clear indication of its intent in the text of the amended provision."); Midlantic Nat'l Bank v. N.J. Dep't of Envtl. Prot., 474 U.S. 494, 501, 106 S.Ct. 755, 88 L.Ed.2d 859 (1986) ("The normal rule of statutory construction is that if Congress intends for legislation to change the interpretation of a judicially created concept, it makes that intent specific."); cf. Whitman v. Am. Trucking Ass'n, 531 U.S. 457, 468, 121 S.Ct. 903, 149 L.Ed.2d 1 (2001) ("Congress, we have held, does not alter the fundamental details of a regulatory scheme in vague terms or ancillary provisions—it does not, one might say, hide elephants in mouseholes."). Congress's silence on the matter is thus conclusive.

Yet even if one probes further, it boggles the mind to suggest that Congress intended to turn a nonmandatory provision into a mandatory provision by doing away with the only means of incentivizing compliance with that provision. Congress quite plainly intended to relieve individuals of the burden the coverage requirement put on them; it did not intend to increase that burden. And if it did, it certainly did not make that intent "clear and manifest." Lively, 717 F.3d at 410. Moreover, the considerations that led the NFIB Court to conclude that Congress did not intend the coverage requirement to impose a legal command to purchase health insurance are even more compelling in the absence of the shared-responsibility payment. Whereas before the only "negative legal consequence[ ] to not buying health insurance" was the payment of a tax, NFIB, 567 U.S. at 567-68, 132 S.Ct. 2566, now there are no consequences at all. And as the Congressional Budget Office ("CBO") has predicted, without the shared-responsibility payment, most applicable individuals will not maintain health insurance solely for the purpose of obeying the coverage requirement. See Cong. Budget Office, Repealing the Individual Health Insurance Mandate: An Updated Estimate at 1 (2017). "That Congress apparently regards such extensive failure to comply with the [coverage requirement] as tolerable suggests that Congress did not think it was creating [millions of] outlaws." NFIB, 567 U.S. at 568, 132 S.Ct. 2566.

Ergo, when Congress zeroed-out the shared-responsibility payment without amending the coverage requirement, it did not do away with the lawful choice it previously offered applicable individuals; it simply changed the parameters of that choice. Under the old scheme, applicable individuals could lawfully choose between maintaining health insurance and paying a tax. Under the new scheme, applicable individuals can lawfully choose between maintaining health insurance and doing nothing. In other words, the coverage requirement is a dead letter—it functions as an expression of national policy or words of encouragement, at most. Accordingly, although I would not reach the merits, I would reverse if I did.

IV.

I agree with much of what the majority has to say about the district court's severability ruling. But I fail to understand the logic behind remanding this case for a do-over. Severability is a question of law that this court can review de novo. And the answer here is quite simple—indeed, a severability analysis will rarely be easier. After all, "[o]ne determines what Congress would have done by examining what it did," and Congress declawed the coverage requirement without repealing any other part of the ACA. Legal Servs. Corp v. Velazquez, 531 U.S. 533, 560, 121 S.Ct. 1043, 149 L.Ed.2d 63 (2001) (Scalia, J., dissenting); see also Ayotte v. Planned Parenthood of N. New Eng., 546 U.S. 320, 330, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006) ("[T]he touchstone for [severability analysis] is legislative intent."). Consequently, little guesswork is needed to determine that Congress believed the ACA could stand in its entirety without the unenforceable coverage requirement.

The majority suggests that remand is necessary because the district court "has many tools at its disposal" and is thus "best positioned to undertake" the severability inquiry. Maj. Op. at 402. It is true that the district court is better able to assess factual issues than appellate judges, because it can hold evidentiary hearings, but I cannot see how that could be relevant, since severability is a question of law that we review de novo. Further, it is not clear what sort of evidence the district court could receive that would be useful when deciding severability questions except perhaps legislative history, a source which the majority derides. See Maj. Op. at 399-400 n.44 ("[W]e caution against relying on individual statements by legislators to determine the meaning of the law."). When it comes to analyzing the statute's text and historical context, see id., we are just as competent as the district court. There is thus no reason to prolong the uncertainty this litigation has caused to the future of this indubitably significant statute.12

A.

Before I address the more specific problems with the district court's inseverability ruling, some background on the ACA is in order. Congress passed the ACA in 2010 to address a growing crisis of Americans living without health insurance. Prior to the ACA, nearly 50 million Americans (about 15 percent of the population at the time) were uninsured. Florida ex rel. Att'y Gen. v. U.S. Dep't of Health & Human Servs., 648 F.3d 1235, 1244 (11th Cir. 2011), rev'd on other grounds, NFIB, 567 U.S. 519, 132 S.Ct. 2566. Although many large employers provided health insurance, coverage was often cost prohibitive for small businesses and consumers seeking insurance through the individual market (i.e., directly instead of through an employer). See U.S. Gov't Accountability Office, GAO-12-166R, Health Care Coverage: Job Lock and the Potential Impact of the Patient Protection and Affordable Care Act 3-4 (2011). Moreover, insurance companies could—and regularly would—deny coverage to high-risk consumers, especially those with preexisting medical conditions. Id. at 4.

The pre-ACA status quo created numerous economic and social problems. Most obviously, America's uninsured population could not afford spiraling healthcare costs, thus exacerbating health problems, leading to an estimated 45,000 premature deaths annually, Andrew P. Wilper et al., Health Insurance and Mortality in US Adults, 99 Am. J. Pub. Health 2289, 2292 (2009), and causing "62 percent of all personal bankruptcies," 42 U.S.C. § 18091(2)(G). The uninsured crisis caused some subtler problems too. For one thing, hospitals would have to absorb the costs of treating uninsured patients and would inevitably pass those costs along to insurance companies, which would then pass them along to consumers. See § 18091(2)(F) ("The cost of providing uncompensated care to the uninsured was $43,000,000,000 in 2008. To pay for this cost, health care providers pass on the cost to private insurers, which pass on the cost to families."). See generally Amicus Br. of HCA Healthcare, Inc. at 9-13. And dependency on employer-based healthcare decreased labor mobility, discouraged entrepreneurship, and kept potential caregivers away from the home. See GAO-12-166R, supra, at 5-6.

In enacting the ACA, Congress sought to address these and other problems with the national healthcare system by drastically reducing the number of uninsured and underinsured Americans. To achieve this goal, the ACA undertook a series of reforms, most notably to the individual insurance market. See generally Patient Protection and Affordable Care Act, Pub. L. No. 111-148, tit. I, 124 Stat. 119 (2010). Among the ACA's most important (and visible) reforms are two related provisions: guaranteed issue and community rate. See 42 U.S.C. §§ 300gg, 300gg-1. The guaranteed-issue provision requires health-insurance providers to accept every individual who applies for coverage, thus preventing insurers from denying coverage based on a consumer's preexisting medical condition. See § 300gg-1(a). The community-rate provision prevents insurers from charging a higher rate because of a policyholder's medical condition. See § 300gg(a).

Left without some counterbalance, the guaranteed-issue and community-rate provisions threatened to overload insurers' risk pools with high-risk policyholders. Beyond allowing more high-risk consumers to purchase health insurance (as intended), these provisions disincentivized healthy (i.e., low risk) consumers from purchasing health insurance because it allowed them to wait until they developed costly health problems to purchase insurance.13 This would have caused premiums to skyrocket, exacerbating many of the problems Congress sought to solve. See generally Amicus Br. of Blue Cross Blue Shield Ass'n at 3-4. Thus, the ACA included several provisions to incentivize low-risk consumers to purchase health insurance. It offered tax credits to offset much of the cost of health insurance for middle-income consumers. See 26 U.S.C. § 36B(b). It created healthcare exchanges to facilitate competition among health plans and to lower transaction costs. See 42 U.S.C. §§ 18031, 18041. It limited new enrollments to an open-enrollment period set by the Secretary of Health and Human Services, which mitigates the adverse-selection problem by preventing consumers from purchasing health insurance only when they need it. See § 18031(c)(6). And it included the coverage requirement at issue in this lawsuit. See § 5000A(a).

Although the coverage requirement has been among the ACA's best-known provisions, the ACA's reforms to the private insurance market extend well beyond it. As just mentioned, Congress created other mechanisms to achieve the same goal as the coverage requirement: incentivize low-risk consumers to purchase health insurance. The ACA also included other provisions expanding access to the private insurance market, including a requirement that employers with 50 or more employees offer health insurance, see 26 U.S.C. § 4980H, and a requirement that health-insurance providers allow young adults to remain on their parents' insurance until they turn 26, see 42 U.S.C. § 300gg-14. And it included provisions designed to make health-insurance policies more attractive, such as those directly regulating premiums, see, e.g., id. § 300gg-18(b), limiting benefits caps, see id. § 300gg-11, and prescribing certain minimum-coverage requirements for health plans, see, e.g., id. § 300gg-13. Moreover, the ACA contains countless other provisions that are unrelated to the private insurance market—and many that are only tangentially related to health insurance at all.14 The following are only some of many possible examples:

Given the breadth of the ACA and the importance of the problems that Congress set out to address, it is simply unfathomable to me that Congress hinged the future of the entire statute on the viability of a single, deliberately unenforceable provision.15

B.

In Planned Parenthood of Northern New England, the Court announced the three principles that must guide our severability analysis. "First, we try not to nullify more of a legislature's work than is necessary, for we know that `[a] ruling of unconstitutionality frustrates the intent of the elected representatives of the people.'" Planned Parenthood of N. New Eng., 546 U.S. at 329, 126 S.Ct. 961 (alteration in original) (quoting Regan v. Time, Inc., 468 U.S. 641, 652, 104 S.Ct. 3262, 82 L.Ed.2d 487 (1984) (plurality opinion)). "Second, mindful that our constitutional mandate and institutional competence are limited, we restrain ourselves from `rewrit[ing] [a] law to conform it to constitutional requirements' even as we strive to salvage it." Id. (first alteration in original) (quoting Am. Booksellers, 484 U.S. at 397, 108 S.Ct. 636). "Third, the touchstone for any decision about remedy is legislative intent, for a court cannot `use its remedial powers to circumvent the intent of the legislature.'" Id. at 330, 126 S.Ct. 961 (quoting Califano v. Westcott, 443 U.S. 76, 94, 99 S.Ct. 2655, 61 L.Ed.2d 382 (1979) (Powell, J., concurring in part and dissenting in part)).

In accordance with these principles, the Court's cases suggest a two-part inquiry. First, we must ask "whether the law remains `fully operative' without the invalid provisions." Murphy v. NCAA, ___ U.S. ___, 138 S.Ct. 1461, 1482, 200 L.Ed.2d 854 (2018); see also United States v. Booker, 543 U.S. 220, 258-59, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005); Alaska Airlines, Inc. v. Brock, 480 U.S. 678, 684, 107 S.Ct. 1476, 94 L.Ed.2d 661 (1987). If so, the remaining provisions are "presumed severable" from the invalid provision. Chadha, 462 U.S. at 934, 103 S.Ct. 2764 (quoting Champlin Ref. Co. v. Corp. Comm'n, 286 U.S. 210, 234, 52 S.Ct. 559, 76 S.Ct. 1062 (1932)). This presumption is rebutted only if "the statute's text or historical context makes it `evident' that Congress, faced with the limitations imposed by the Constitution, would have preferred" no statute over the statute with only the permissible provisions. Free Enter. Fund v. Pub. Co. Accounting Oversight Bd., 561 U.S. 477, 509, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010). And as should be clear by now, "the `normal rule' is `that partial, rather than facial, invalidation is the required course.'" Id. at 508, 130 S.Ct. 3138 (quoting Brockett v. Spokane Arcades, Inc., 472 U.S. 491, 504, 105 S.Ct. 2794, 86 L.Ed.2d 394 (1985)).

1.

The majority has identified the most glaring flaw in the district court's severability analysis: the district court "gives relatively little attention to the intent of the 2017 Congress, which appears in the analysis only as an afterthought." When one takes this fact into account, there can be little doubt as to Congress's intent.

We have unusual insight into Congress's thinking because Congress was given a chance to weigh in on the ACA's future without an effective coverage requirement and it decided the ACA should remain in place. By zeroing out the shared-responsibility payment, the 2017 Congress left the coverage requirement unenforceable. If Congress viewed the coverage requirement as so essential to the rest of the ACA that it intended the entire statute to rise and fall with the coverage requirement, it is inconceivable that Congress would have declawed the coverage requirement as it did. And make no mistake: Congress declawed the coverage requirement. As the CBO found only a month before Congress passed the TCJA, "[i]f the [coverage requirement] penalty was eliminated but the [coverage requirement] itself was not repealed, the results would be very similar to" if the coverage requirement itself were repealed. 2017 CBO Report, supra, at 1. Regardless of lofty civic notions about people who follow the law for the sake of following the law, the objective evidence before Congress was that "only a small number of people" would obey the coverage requirement without the shared-responsibility payment. Id.; cf. Dep't of Commerce, 139 S. Ct. at 2565-66 (concluding people will "predictabl[y]" "violate their legal duty" when incentivized to do so). Congress accordingly knew that repealing the shared-responsibility payment would have the same essential effect on the ACA's statutory scheme as would repealing the coverage requirement.

Furthermore, as various amici highlight, judicial repeal of the ACA would have potentially devastating effects on the national healthcare system and the economy at large. See, e.g., Amicus Br. of Am.'s Health Ins. Plans (discussing impact on health-insurance industry); Amicus Br. of 35 Counties, Cities, and Towns (discussing impact on municipalities); Amicus Br. of Bipartisan Econ. Scholars (discussing impact on economy); Amicus Br. of Am. Hosp. Ass'n et al. (discussing impact on hospitals). Regardless of whether the ACA is good or bad policy, it is undoubtedly significant policy. It is unlikely that Congress would want a statute on which millions of people rely for their healthcare and livelihoods to disappear overnight with the wave of a judicial wand. If Congress wanted to repeal the ACA through the deliberative legislative process, it could have done so. But with the stakes so high, it is difficult to imagine that this is a matter Congress intended to turn over to the judiciary.

2.

A second flaw in the district court's analysis is the great weight it places on the fact that Congress in 2017 did not repeal its statutory findings emphasizing the coverage requirement's importance to the guaranteed-issue and community-rate provisions. See 42 U.S.C. § 18091. The district court overread the significance of § 18091. Congress enacted the findings in § 18091 to demonstrate the coverage requirement's role in regulating interstate commerce. When it invokes its commerce power, Congress routinely makes such findings to facilitate judicial review. See United States v. Morrison, 529 U.S. 598, 612, 120 S.Ct. 1740, 146 L.Ed.2d 658 (2000) ("While `Congress normally is not required to make formal findings as to the substantial burdens that an activity has on interstate commerce,' the existence of such findings may `enable us to evaluate the legislative judgment that the activity in question substantially affect[s] interstate commerce, even though no such substantial effect [is] visible to the naked eye.'" (alterations in original) (citation omitted) (quoting United States v. Lopez, 514 U.S. 549, 562-63, 115 S.Ct. 1624, 131 L.Ed.2d 626 (1995))). Indeed, § 18091(2), the subsection the district court focused its attention on, is entitled "Effects on the national economy and interstate commerce."

Section 18091 is not an inseverability clause, and nothing in its text suggests that Congress intended to make the coverage requirement inseverable from the remainder of the ACA. If Congress intended to draft an inseverability clause, it knew how to do so. See Office of Legislative Counsel, U.S. Senate, Senate Legislative Drafting Manual § 131(b) (1997) (explaining purpose of inseverability clause). Compare id. § 131(c) (providing as example of proper form for inseverability clause: "EFFECT OF INVALIDITY ON OTHER PROVISIONS OF ACT.—If section 501, 502, or 503 of the Federal Election Campaign Act of 1971 (as added by this section) or any part of those sections is held to be invalid, all provisions of and amendments made by this Act shall be invalid"), with § 18091(2)(H) ("The requirement is an essential part of this larger regulation of economic activity, and the absence of the requirement would undercut Federal regulation of the health insurance market."). In fact, both the House and the Senate legislative drafting guides suggest that Congress should include an inseverability clause if it wants to make a statute inseverable because "[t]he Supreme Court has made it quite clear that invalid portions of statutes are to be severed `unless it is evident that the Legislature would not have enacted those provisions which are within its powers, independently of that which is not.'" Office of Legislative Counsel, U.S. House of Representatives, House Legislative Counsel's Manual on Drafting Style § 328 (1995) (quoting Chadha, 462 U.S. at 931, 103 S.Ct. 2764); accord Senate Legislative Drafting Manual, supra, at § 131(a). The absence of a genuine inseverability clause should be all but conclusive in assessing the legislature's intent.

Moreover, the argument that § 18091 is meant to signal Congress's intent that the coverage requirement be inseverable proves far too much. Section 18091 discusses the coverage requirement's importance to the entire federal healthcare regulatory scheme, including—along with the ACA—the Public Health Service Act ("PHSA") and the Employee Retirement Income Security Act ("ERISA"). See § 18091(2)(H) ("Under the Employee Retirement Income Security Act of 1974 (29 U.S.C. 1001 et seq.), the Public Health Service Act (42 U.S.C. 201 et seq.), and this Act, the Federal Government has a significant role in regulating health insurance. The [coverage] requirement is an essential part of this larger regulation of economic activity, and the absence of the requirement would undercut Federal regulation of the health insurance market." (emphasis added)). It is not suggested that Congress intended a court to strike down the PHSA and ERISA if it found the coverage requirement unconstitutional. This would be especially implausible given the intensity of the debate over the coverage requirement's constitutionality from the get-go. See NFIB, 567 U.S. at 540, 132 S.Ct. 2566 ("On the day the President signed the [ACA] into law, Florida and 12 other States filed a complaint in the Federal District Court for the Northern District of Florida."). Yet in signaling that the coverage requirement is "an essential part of this larger regulation," Congress did not distinguish between the ACA and these prior statutes. Thus, § 18091 cannot reasonably be read to bear on the coverage requirement's severability.

3.

Another flaw in the district court's analysis is its suggestion that the Supreme Court concluded in NFIB and King v. Burwell, ___ U.S. ___, 135 S.Ct. 2480, 192 L.Ed.2d 483 (2015), that the coverage requirement is inseverable from the ACA's guaranteed-issue and community-rate provisions. The district court misconstrued these opinions. And even if the district court read them correctly, these opinions address the coverage requirement as enforced by the shared-responsibility payment. They give little valuable insight into the coverage requirement's role in the post-TCJA ACA.

In NFIB, only the dissenters addressed the coverage requirement's severability. The district court did not suggest it is bound by a Supreme Court dissent, and of course it is not. The district court instead took language from the other five Justices out of context to conclude that each of them viewed the coverage requirement as inseverable. But none of the language the district court cited addresses severability. See NFIB, 567 U.S. at 547-48, 132 S.Ct. 2566 (opinion of Roberts, C.J.) (discussing Government's argument that coverage requirement plays a role in regulating interstate commerce); id. at 597, 132 S.Ct. 2566 (Ginsburg, J., dissenting in part) (same). Although the Justices' reasoning certainly suggests that they saw the coverage requirement as an important part of the statutory scheme as it existed in 2012, this does not mean the Justices found it "evident" that Congress would have preferred the entire statute to fall without the coverage requirement. Alaska Airlines, 480 U.S. at 684, 107 S.Ct. 1476.

King likewise contains some helpful commentary about the ACA's original statutory scheme, but it does not discuss severability or otherwise control the severability analysis. The Court ruled in King that the ACA's tax credits were available to every eligible consumer regardless of whether the state in which a consumer lived established its own exchange or relied on the federally operated exchange. 135 S. Ct. at 2496. The coverage requirement came up because many more individuals would have been exempt from the shared-responsibility payment if tax credits were not available to them. Id. at 2493-95; see also § 5000A(e)(1)(A) ("No penalty shall be imposed ... with respect to ... [a]ny applicable individual for any month if the applicable individual's required contribution (determined on an annual basis) for coverage for the month exceeds 8 percent of such individual's household income....").16 Noting the importance of the tax credits and coverage requirement (as enforced by the shared-responsibility payment) to the statutory structure, the Court concluded as a matter of statutory interpretation that Congress did not intend a scheme in which neither tax credits nor the coverage requirement were operating to bring low-risk consumers into the insurance pools. See King, 135 S. Ct. at 2492-94 ("The combination of no tax credits and an ineffective coverage requirement could well push a State's individual insurance market into a death spiral.... It is implausible that Congress meant the [ACA] to operate in this manner.").

The district court framed King as saying that Congress intrinsically tied the community-rate and guaranteed-issue provisions to the coverage requirement, meaning that those provisions must be inseverable from the coverage requirement. But the district court ignored a crucial aspect of the King Court's analysis: it explicitly discussed the coverage requirement as enforced by the shared-responsibility payment. See id. at 2493 (referring to the coverage requirement as "a requirement that individuals maintain health insurance coverage or make a payment to the IRS" (emphasis added)). Indeed, as the Court identified it, the crux of the problem with denying consumers tax credits in federal-exchange states was that doing so would make a large number of individuals unable to afford insurance, thus exempting them from the shared-responsibility payment. See id. These widespread exemptions would, in turn, make the coverage requirement "ineffective." Id. King thus speaks far more to the shared-responsibility payment's role in the ACA's pre-TCJA statutory scheme than it does the coverage requirement's role in the statutory scheme.

Even to the extent the Court in NFIB or King meant to opine on the coverage requirement's severability, these cases were both decided before the TCJA. They thus give no insight into how the coverage requirement fits into the post-TCJA scheme. Whatever reservations the Court previously harbored about severing the coverage requirement, Congress plainly did not share those concerns when it zeroed out the shared-responsibility payment. Congress either concluded that healthcare markets under the ACA had reached a point of stability at which they no longer needed an effective coverage requirement,17 or it chose to accept the negative side effects of effectively repealing the coverage requirement as a cost of relieving the burden it placed on applicable individuals. Either way, the legislative considerations have necessarily shifted.

In sum, there was no reason for the district court to conclude that any provision in the ACA was inseverable from the coverage requirement. The majority does not necessarily disagree. I thus do not understand its decision to remand when, even on the majority's analysis of the case, it could instead reverse and render a judgment declaring only the coverage requirement unconstitutional.

V.

Limits on judicial power demand special respect in a case like this. For one thing, careless judicial interference has the potential to be especially pernicious when it involves a complex statute like the ACA, which carries such significant implications for the welfare of the economy and the American populace at large. For another, the legitimacy of the judicial branch as a counter-majoritarian institution in an otherwise democratic system depends on its ability to operate with restraint—and especially so in a high-profile case such as the one at bar. The district court's opinion is textbook judicial overreach. The majority perpetuates that overreach and, in remanding, ensures that no end for this litigation is in sight.

I respectfully dissent.