JUSTICE BLACKMUN delivered the opinion of the Court.

[1] We revisit in these cases an issue raised in National League of Cities v. Usery (1976). In that litigation, this Court, by a sharply divided vote, ruled that the Commerce Clause does not empower Congress to enforce the minimum-wage and overtime provisions of the Fair Labor Standards Act (FLSA) against the States “in areas of traditional governmental functions.” Id., at 852. Although National League of Cities supplied some examples of “traditional governmental functions,” it did not offer a general explanation of how a “traditional” function is to be distinguished from a “nontraditional” one. Since then, federal and state courts have struggled with the task, thus imposed, of identifying a traditional function for purposes of state immunity under the Commerce Clause.

[2] In the present cases, a Federal District Court concluded that municipal ownership and operation of a mass-transit system is a traditional governmental function and thus, under National League of Cities, is exempt from the obligations imposed by the FLSA. Faced with the identical question, three Federal Courts of Appeals and one state appellate court have reached the opposite conclusion.

[3] Our examination of this “function” standard applied in these and other cases over the last eight years now persuades us that the attempt to draw the boundaries of state regulatory immunity in terms of “traditional governmental function” is not only unworkable but is also inconsistent with established principles of federalism and, indeed, with those very federalism principles on which National League of Cities purported to rest. That case, accordingly, is overruled.

**

[4] The present controversy concerns the extent to which SAMTA may be subjected to the minimum-wage and overtime requirements of the FLSA.

**

[5] The prerequisites for governmental immunity under National League of Cities were summarized by this Court in Hodel, supra. Under that summary, four conditions must be satisfied before a state activity may be deemed immune from a particular federal regulation under the Commerce Clause. First, it is said that the federal statute at issue must regulate “the `States as States.’ ” Second, the statute must “address matters that are indisputably `attribute[s] of state sovereignty.’ ” Third, state compliance with the federal obligation must “directly impair [the States’] ability `to structure integral operations in areas of traditional governmental functions.’ ” Finally, the relation of state and federal interests must not be such that “the nature of the federal interest . . . justifies state submission.”

[6] The controversy in the present cases has focused on the third Hodel requirement — that the challenged federal statute trench on “traditional governmental functions.” The District Court voiced a common concern: “Despite the abundance of adjectives, identifying which particular state functions are immune remains difficult.” Just how troublesome the task has been is revealed by the results reached in other federal cases. Thus, courts have held that regulating ambulance services; licensing automobile drivers; operating a municipal airport; performing solid waste disposal; and operating a highway authority, are functions protected under National League of Cities. At the same time, courts have held that issuance of industrial development bonds; regulation of intrastate natural gas sales; regulation of traffic on public roads; regulation of air transportation; operation of a telephone system; leasing and sale of natural gas; operation of a mental health facility; and provision of in-house domestic services for the aged and handicapped, are not entitled to immunity. We find it difficult, if not impossible, to identify an organizing principle that places each of the cases in the first group on one side of a line and each of the cases in the second group on the other side. The constitutional distinction between licensing drivers and regulating traffic, for example, or between operating a highway authority and operating a mental health facility, is elusive at best.

[7] Thus far, this Court itself has made little headway in defining the scope of the governmental functions deemed protected under National League of Cities. In that case the Court set forth examples of protected and unprotected functions, but provided no explanation of how those examples were identified. The only other case in which the Court has had occasion to address the problem is Long Island. We there observed: “The determination of whether a federal law impairs a state’s authority with respect to `areas of traditional [state] functions’ may at times be a difficult one.” The accuracy of that statement is demonstrated by this Court’s own difficulties in Long Island is developing a workable standard for “traditional governmental functions.” We relied in large part there on “the historical reality that the operation of railroads is not among the functions traditionally performed by state and local governments,” but we simultaneously disavowed “a static historical view of state functions generally immune from federal regulation.” We held that the inquiry into a particular function’s “traditional” nature was merely a means of determining whether the federal statute at issue unduly handicaps “basic state prerogatives,” but we did not offer an explanation of what makes one state function a “basic prerogative” and another function not basic. Finally, having disclaimed a rigid reliance on the historical pedigree of state involvement in a particular area, we nonetheless found it appropriate to emphasize the extended historical record of federal involvement in the field of rail transportation.

**

[8] [After going through the court’s parallel historical struggle to find a workable distinction between proprietary and governmental functions for the purposes of tax immunity, the Court Made the following observations]

[9] The distinction the Court discarded as unworkable in the field of tax immunity has proved no more fruitful in the field of regulatory immunity under the Commerce Clause. Neither do any of the alternative standards that might be employed to distinguish between protected and unprotected governmental functions appear manageable. We rejected the possibility of making immunity turn on a purely historical standard of “tradition” in Long Island, and properly so. The most obvious defect of a historical approach to state immunity is that it prevents a court from accommodating changes in the historical functions of States, changes that have resulted in a number of once-private functions like education being assumed by the States and their subdivisions. At the same time, the only apparent virtue of a rigorous historical standard, namely, its promise of a reasonably objective measure for state immunity, is illusory. Reliance on history as an organizing principle results in line-drawing of the most arbitrary sort; the genesis of state governmental functions stretches over a historical continuum from before the Revolution to the present, and courts would have to decide by fiat precisely how longstanding a pattern of state involvement had to be for federal regulatory authority to be defeated.

[10] A nonhistorical standard for selecting immune governmental functions is likely to be just as unworkable as is a historical standard. The goal of identifying “uniquely” governmental functions, for example, has been rejected by the Court in the field of government tort liability in part because the notion of a “uniquely” governmental function is unmanageable. Another possibility would be to confine immunity to “necessary” governmental services, that is, services that would be provided inadequately or not at all unless the government provided them. The set of services that fits into this category, however, may well be negligible. The fact that an unregulated market produces less of some service than a State deems desirable does not mean that the State itself must provide the service; in most if not all cases, the State can “contract out” by hiring private firms to provide the service or simply by providing subsidies to existing suppliers. It also is open to question how well equipped courts are to make this kind of determination about the workings of economic markets.

[11] We believe, however, that there is a more fundamental problem at work here, a problem that explains why the Court was never able to provide a basis for the governmental/proprietary distinction in the intergovernmental tax-immunity cases and why an attempt to draw similar distinctions with respect to federal regulatory authority under National League of Cities is unlikely to succeed regardless of how the distinctions are phrased. The problem is that neither the governmental/proprietary distinction nor any other that purports to separate out important governmental functions can be faithful to the role of federalism in a democratic society. The essence of our federal system is that within the realm of authority left open to them under the Constitution, the States must be equally free to engage in any activity that their citizens choose for the common weal, no matter how unorthodox or unnecessary anyone else — including the judiciary — deems state involvement to be. Any rule of state immunity that looks to the “traditional,” “integral,” or “necessary” nature of governmental functions inevitably invites an unelected federal judiciary to make decisions about which state policies it favors and which ones it dislikes. “The science of government . . . is the science of experiment,” and the States cannot serve as laboratories for social and economic experiment, if they must pay an added price when they meet the changing needs of their citizenry by taking up functions that an earlier day and a different society left in private hands.

**

[12] We therefore now reject, as unsound in principle and unworkable in practice, a rule of state immunity from federal regulation that turns on a judicial appraisal of whether a particular governmental function is “integral” or “traditional.” Any such rule leads to inconsistent results at the same time that it disserves principles of democratic self-governance, and it breeds inconsistency precisely because it is divorced from those principles. If there are to be limits on the Federal Government’s power to interfere with state functions — as undoubtedly there are — we must look elsewhere to find them. We accordingly return to the underlying issue that confronted this Court in National League of Cities — the manner in which the Constitution insulates States from the reach of Congress’ power under the Commerce Clause.

**

[13] The States unquestionably do “retai[n] a significant measure of sovereign authority.” They do so, however, only to the extent that the Constitution has not divested them of their original powers and transferred those powers to the Federal Government. In the words of James Madison to the Members of the First Congress: “Interference with the power of the States was no constitutional criterion of the power of Congress. If the power was not given, Congress could not exercise it; if given, they might exercise it, although it should interfere with the laws, or even the Constitution of the States.” 2 Annals of Cong. 1897 (1791).

**

[14] When we look for the States’ “residuary and inviolable sovereignty,” in the shape of the constitutional scheme rather than in predetermined notions of sovereign power, a different measure of state sovereignty emerges. Apart from the limitation on federal authority inherent in the delegated nature of Congress’ Article I powers, the principal means chosen by the Framers to ensure the role of the States in the federal system lies in the structure of the Federal Government itself. It is no novelty to observe that the composition of the Federal Government was designed in large part to protect the States from overreaching by Congress. The Framers thus gave the States a role in the selection both of the Executive and the Legislative Branches of the Federal Government. The States were vested with indirect influence over the House of Representatives and the Presidency by their control of electoral qualifications and their role in Presidential elections. U. S. Const., Art. I, § 2, and Art. II, § 1. They were given more direct influence in the Senate, where each State received equal representation and each Senator was to be selected by the legislature of his State. Art. I, § 3. The significance attached to the States’ equal representation in the Senate is underscored by the prohibition of any constitutional amendment divesting a State of equal representation without the State’s consent. Art. V.

**

[15] Insofar as the present cases are concerned, then, we need go no further than to state that we perceive nothing in the overtime and minimum-wage requirements of the FLSA, as applied to SAMTA, that is destructive of state sovereignty or violative of any constitutional provision. SAMTA faces nothing more than the same minimum-wage and overtime obligations that hundreds of thousands of other employers, public as well as private, have to meet.

**

[16] Congress has not simply placed a financial burden on the shoulders of States and localities that operate mass-transit systems, but has provided substantial countervailing financial assistance as well, assistance that may leave individual mass-transit systems better off than they would have been had Congress never intervened at all in the area. Congress’ treatment of public mass transit reinforces our conviction that the national political process systematically protects States from the risk of having their functions in that area handicapped by Commerce Clause regulation.

[17] This analysis makes clear that Congress’ action in affording SAMTA employees the protections of the wage and hour provisions of the FLSA contravened no affirmative limit on Congress’ power under the Commerce Clause. The judgment of the District Court therefore must be reversed.

[18] Of course, we continue to recognize that the States occupy a special and specific position in our constitutional system and that the scope of Congress’ authority under the Commerce Clause must reflect that position. But the principal and basic limit on the federal commerce power is that inherent in all congressional action — the built-in restraints that our system provides through state participation in federal governmental action. The political process ensures that laws that unduly burden the States will not be promulgated. In the factual setting of these cases the internal safeguards of the political process have performed as intended.

[19] These cases do not require us to identify or define what affirmative limits the constitutional structure might impose on federal action affecting the States under the Commerce Clause. We note and accept Justice Frankfurter’s observation in (1946):

[20] “The process of Constitutional adjudication does not thrive on conjuring up horrible possibilities that never happen in the real world and devising doctrines sufficiently comprehensive in detail to cover the remotest contingency. Nor need we go beyond what is required for a reasoned disposition of the kind of controversy now before the Court.”

**

[21] We do not lightly overrule recent precedent. We have not hesitated, however, when it has become apparent that a prior decision has departed from a proper understanding of congressional power under the Commerce Clause. Due respect for the reach of congressional power within the federal system mandates that we do so now.

[21] National League of Cities v. Usery (1976), is overruled. The judgment of the District Court is reversed, and these cases are remanded to that court for further proceedings consistent with this opinion.

[22] It is so ordered.

JUSTICE POWELL, with whom THE CHIEF JUSTICE, JUSTICE REHNQUIST, and JUSTICE O’CONNOR join, dissenting.

**

[1] Despite some genuflecting in the Court’s opinion to the concept of federalism, today’s decision effectively reduces the Tenth Amendment to meaningless rhetoric when Congress acts pursuant to the Commerce Clause.

**

[2] Much of the Court’s opinion is devoted to arguing that it is difficult to define a priori “traditional governmental functions.” National League of Cities neither engaged in, nor required, such a task. The Court discusses and condemns as standards “traditional governmental functions,” “purely historical” functions, ” `uniquely’ governmental functions,” and ” `necessary’ governmental services.” But nowhere does it mention that National League of Cities adopted a familiar type of balancing test for determining whether Commerce Clause enactments transgress constitutional limitations imposed by the federal nature of our system of government. This omission is noteworthy, since the author of today’s opinion joined National League of Cities and concurred separately to point out that the Court’s opinion in that case “adopt[s] a balancing approach [that] does not outlaw federal power in areas . . . where the federal interest is demonstrably greater and where state . . . compliance with imposed federal standards would be essential.” (BLACKMUN, J., concurring).

**

[3] Today’s opinion does not explain how the States’ role in the electoral process guarantees that particular exercises of the Commerce Clause power will not infringe on residual state sovereignty. Members of Congress are elected from the various States, but once in office they are Members of the Federal Government. Although the States participate in the Electoral College, this is hardly a reason to view the President as a representative of the States’ interest against federal encroachment. We noted recently “[t]he hydraulic pressure inherent within each of the separate Branches to exceed the outer limits of its power . . . .” INS v. Chadh (1983). The Court offers no reason to think that this pressure will not operate when Congress seeks to invoke its powers under the Commerce Clause, notwithstanding the electoral role of the States.

**

[4] Much of the initial opposition to the Constitution was rooted in the fear that the National Government would be too powerful and eventually would eliminate the States as viable political entities. This concern was voiced repeatedly until proponents of the Constitution made assurances that a Bill of Rights, including a provision explicitly reserving powers in the States, would be among the first business of the new Congress.

**

[5] This history, which the Court simply ignores, documents the integral role of the Tenth Amendment in our constitutional theory. It exposes as well, I believe, the fundamental character of the Court’s error today. Far from being “unsound in principle,” ante, at 546, judicial enforcement of the Tenth Amendment is essential to maintaining the federal system so carefully designed by the Framers and adopted in the Constitution.

**

[6] [The majority opinion reflects] the Court’s unprecedented view that Congress is free under the Commerce Clause to assume a State’s traditional sovereign power, and to do so without judicial review of its action. Indeed, the Court’s view of federalism appears to relegate the States to precisely the trivial role that opponents of the Constitution feared they would occupy.

**

[7] Although the Court’s opinion purports to recognize that the States retain some sovereign power, it does not identify even a single aspect of state authority that would remain when the Commerce Clause is invoked to justify federal regulation.

**

[8] As I view the Court’s decision today as rejecting the basic precepts of our federal system and limiting the constitutional role of judicial review, I dissent.

**

JUSTICE O’CONNOR, with whom JUSTICE POWELL and JUSTICE REHNQUIST join, dissenting.

**

[1] In my view, federalism cannot be reduced to the weak “essence” distilled by the majority today. . . The true “essence” of federalism is that the States as States have legitimate interests which the National Government is bound to respect even though its laws are supreme. Younger v.Harris, (1971). If federalism so conceived and so carefully cultivated by the Framers of our Constitution is to remain meaningful, this Court cannot abdicate its constitutional responsibility to oversee the Federal Government’s compliance with its duty to respect the legitimate interests of the States.

**

[2] It is worth recalling the cited passage in McCulloch v. Maryland, 4 Wheat. 316, 421 (1819), that lies at the source of the recent expansion of the commerce power. “Let the end be legitimate, let it be within the scope of the constitution,” Chief Justice Marshall said, “and all means which are appropriate, which are plainly adapted to that end, which are not prohibited, but consist with the letter and spirit of the constitution, are constitutional” (emphasis added). The spirit of the Tenth Amendment, of course, is that the States will retain their integrity in a system in which the laws of the United States are nevertheless supreme. Fry v. United States, supra, at 547, n. 7.

[3] It is not enough that the “end be legitimate”; the means to that end chosen by Congress must not contravene the spirit of the Constitution. Thus many of this Court’s decisions acknowledge that the means by which national power is exercised must take into account concerns for state autonomy.

**

[4] The problems of federalism in an integrated national economy are capable of more responsible resolution than holding that the States as States retain no status apart from that which Congress chooses to let them retain. The proper resolution, I suggest, lies in weighing state autonomy as a factor in the balance when interpreting the means by which Congress can exercise its authority on the States as States. It is insufficient, in assessing the validity of congressional regulation of a State pursuant to the commerce power, to ask only whether the same regulation would be valid if enforced against a private party. That reasoning, embodied in the majority opinion, is inconsistent with the spirit of our Constitution. It remains relevant that a State is being regulated, as National League of Cities and every recent case have recognized.

**

[5] It has been difficult for this Court to craft bright lines defining the scope of the state autonomy protected by National League of Cities. Such difficulty is to be expected whenever constitutional concerns as important as federalism and the effectiveness of the commerce power come into conflict. Regardless of the difficulty, it is and will remain the duty of this Court to reconcile these concerns in the final instance. That the Court shuns the task today by appealing to the “essence of federalism” can provide scant comfort to those who believe our federal system requires something more than a unitary, centralized government. I would not shirk the duty acknowledged by National League of Cities and its progeny, and I share JUSTICE REHNQUIST’S belief that this Court will in time again assume its constitutional responsibility.

[6] I respectfully dissent.

JUSTICE O’CONNOR, delivered the opinion of the Court.

[1] These cases implicate one of our Nation’s newest problems of public policy and perhaps our oldest question of constitutional law. The public policy issue involves the disposal of radioactive waste: In these cases, we address the constitutionality of three provisions of the Low-Level Radioactive Waste Policy Amendments Act of 1985, Pub. L. 99-240, 99 Stat. 1842, 42 U. S. C. § 2021b et seq. The constitutional question is as old as the Constitution: It consists of discerning the proper division of authority between the Federal Government and the States. We conclude that while Congress has substantial power under the Constitution to encourage the States to provide for the disposal of the radioactive waste generated within their borders, the Constitution does not confer upon Congress the ability simply to compel the States to do so.

I

[2] We live in a world full of low level radioactive waste. Radioactive material is present in luminous watch dials, smoke alarms, measurement devices, medical fluids, research materials, and the protective gear and construction materials used by workers at nuclear power plants. Low level radioactive waste is generated by the Government, by hospitals, by research institutions, and by various industries. The waste must be isolated from humans for long periods of time, often for hundreds of years. Millions of cubic feet of low level radioactive waste must be disposed of each year.

**

[3] Since 1979 only three disposal sites— those in Nevada, Washington, and South Carolina—have been in operation. Waste generated in the rest of the country must be shipped to one of these three sites for disposal.

[4] In 1979, both the Washington and Nevada sites were forced to shut down temporarily, leaving South Carolina to shoulder the responsibility of storing low level radioactive waste produced in every part of the country. The Governor of South Carolina, understandably perturbed, ordered a 50% reduction in the quantity of waste accepted at the Barnwell site. The Governors of Washington and Nevada announced plans to shut their sites permanently.

[5] Faced with the possibility that the Nation would be left with no disposal sites for low level radioactive waste, Congress responded by enacting the Low-Level Radioactive Waste Policy Act. Relying largely on a report submitted by the National Governors’ Association, see App. 105a—141a, Congress declared a federal policy of holding each State “responsible for providing for the availability of capacity either within or outside the State for the disposal of low-level radioactive waste generated within its borders,” and found that such waste could be disposed of “most safely and efficiently . . . on a regional basis.”. The 1980 Act authorized States to enter into regional compacts that, once ratified by Congress, would have the authority beginning in 1986 to restrict the use of their disposal facilities to waste generated within member States. The 1980 Act included no penalties for States that failed to participate in this plan.

[6] By 1985, only three approved regional compacts had operational disposal facilities; not surprisingly, these were the compacts formed around South Carolina, Nevada, and Washington, the three sited States. The following year, the 1980 Act would have given these three compacts the ability to exclude waste from nonmembers, and the remaining 31 States would have had no assured outlet for their low level radioactive waste. With this prospect looming, Congress once again took up the issue of waste disposal. The result was the legislation challenged here, the Low-Level Radioactive Waste Policy Amendments Act of 1985.

**

[7] The Act provides three types of incentives to encourage the States to comply with their statutory obligation to provide for the disposal of waste generated within their borders.

1. Monetary incentives. One quarter of the surcharges collected by the sited States must be transferred to an escrow account held by the Secretary of Energy. § 2021e (d)(2)(A). The Secretary then makes payments from this account to each State that has complied with a series of deadlines.

**

1. Access incentives. The second type of incentive involves the denial of access to disposal sites. States that fail to meet the July 1986 deadline may be charged twice the ordinary surcharge for the remainder of 1986 and may be denied access to disposal facilities thereafter.

**

1. The take title provision. The third type of incentive is the most severe. The Act provides:

[8] “If a State (or, where applicable, a compact region) in which low-level radioactive waste is generated is unable to provide for the disposal of all such waste generated within such State or compact region by January 1, 1996, each State in which such waste is generated, upon the request of the generator or owner of the waste, shall take title to the waste, be obligated to take possession of the waste, and shall be liable for all damages directly or indirectly incurred by such generator or owner as a consequence of the failure of the State to take possession of the waste as soon after January 1, 1996, as the generator or owner notifies the State that the waste is available for shipment.” § 2021e(d)(2)(C).

**

[9] New York, . . . did not join a regional compact. Instead, the State complied with the Act’s requirements by enacting legislation providing for the siting and financing of a disposal facility in New York. The State has identified five potential sites, three in Allegany County and two in Cortland County. Residents of the two counties oppose the State’s choice of location.

[10] Petitioners—the State of New York and the two counties—filed this suit against the United States in 1990.

**

[11] In some cases the Court has inquired whether an Act of Congress is authorized by one of the powers delegated to Congress in Article I of the Constitution. In other cases the Court has sought to determine whether an Act of Congress invades the province of state sovereignty reserved by the Tenth Amendment. In a case like these, involving the division of authority between federal and state governments, the two inquiries are mirror images of each other. If a power is delegated to Congress in the Constitution, the Tenth Amendment expressly disclaims any reservation of that power to the States; if a power is an attribute of state sovereignty reserved by the Tenth Amendment, it is necessarily a power the Constitution has not conferred on Congress

**

[12] It is in this sense that the Tenth Amendment “states but a truism that all is retained which has not been surrendered.” United States v. Darby, 312 U. S. 100, 124 (1941).

**

[13] Congress exercises its conferred powers subject to the limitations contained in the Constitution. Thus, for example, under the Commerce Clause Congress may regulate publishers engaged in interstate commerce, but Congress is constrained in the exercise of that power by the First Amendment. The Tenth Amendment likewise restrains the power of Congress, but this limit is not derived from the text of the Tenth Amendment itself, which, as we have discussed, is essentially a tautology. Instead, the Tenth Amendment confirms that the power of the Federal Government is subject to limits that may, in a given instance, reserve power to the States. The Tenth Amendment thus directs us to determine, as in this case, whether an incident of state sovereignty is protected by a limitation on an Article I power.

**

[14] The actual scope of the Federal Government’s authority with respect to the States has changed over the years, therefore, but the constitutional structure underlying and limiting that authority has not. In the end, just as a cup may be half empty or half full, it makes no difference whether one views the question at issue in these cases as one of ascertaining the limits of the power delegated to the Federal Government under the affirmative provisions of the Constitution or one of discerning the core of sovereignty retained by the States under the Tenth Amendment. Either way, we must determine whether any of the three challenged provisions of the Low-Level Radioactive Waste Policy Amendments Act of 1985 oversteps the boundary between federal and state authority.

B

[15] Petitioners do not contend that Congress lacks the power to regulate the disposal of low level radioactive waste. . . . Petitioners likewise do not dispute that under the Supremacy Clause Congress could, if it wished, pre-empt state radioactive waste regulation. Petitioners contend only that the Tenth Amendment limits the power of Congress to regulate in the way it has chosen. Rather than addressing the problem of waste disposal by directly regulating the generators and disposers of waste, petitioners argue, Congress has impermissibly directed the States to regulate in this field.

[16] Most of our recent cases interpreting the Tenth Amendment have concerned the authority of Congress to subject state governments to generally applicable laws. The Court’s jurisprudence in this area has traveled an unsteady path. This litigation presents no occasion to apply or revisit the holdings of any of these cases, as this is not a case in which Congress has subjected a State to the same legislation applicable to private parties. Cf. FERC v. Mississippi, 456 U. S. 742, 758-759 (1982).

[17] This litigation instead concerns the circumstances under which Congress may use the States as implements of regulation; that is, whether Congress may direct or otherwise motivate the States to regulate in a particular field or a particular way. Our cases have established a few principles that guide our resolution of the issue.

1

[18] As an initial matter, Congress may not simply “commandee[r] the legislative processes of the States by directly compelling them to enact and enforce a federal regulatory program.” Hodel v. Virginia Surface Mining & Reclamation Assn., Inc., 452 U. S. 264, 288 (1981).

**

[19] While Congress has substantial powers to govern the Nation directly, including in areas of intimate concern to the States, the Constitution has never been understood to confer upon Congress the ability to require the States to govern according to Congress’ instructions.

**

[20] Indeed, the question whether the Constitution should permit Congress to employ state governments as regulatory agencies was a topic of lively debate among the Framers. Under the Articles of Confederation, Congress lacked the authority in most respects to govern the people directly.

**

[21] The inadequacy of this governmental structure was responsible in part for the Constitutional Convention. Alexander Hamilton observed: “The great and radical vice in the construction of the existing Confederation is in the principle of legislation for states or governments, in their corporate or collective capacities, and as contradistinguished from the individuals of whom they consist.” The Federalist No. 15, p. 108 (C. Rossiter ed. 1961). As Hamilton saw it, “we must resolve to incorporate into our plan those ingredients which may be considered as forming the characteristic difference between a league and a government; we must extend the authority of the Union to the persons of the citizens—the only proper objects of government.”

**

[22] The Convention generated a great number of proposals for the structure of the new Government, but two quickly took center stage. Under the Virginia Plan, as first introduced by Edmund Randolph, Congress would exercise legislative authority directly upon individuals, without employing the States as intermediaries. 1 Records of the Federal Convention of 1787, p. 21 (M. Farrand ed. 1911). Under the New Jersey Plan, as first introduced by William Paterson, Congress would continue to require the approval of the States before legislating, as it had under the Articles of Confederation.

**

[23] In the end, the Convention opted for a Constitution in which Congress would exercise its legislative authority directly over individuals rather than over States.

**

[24] In providing for a stronger central government, therefore, the Framers explicitly chose a Constitution that confers upon Congress the power to regulate individuals, not States.

**

2

[25] This is not to say that Congress lacks the ability to encourage a State to regulate in a particular way, or that Congress may not hold out incentives to the States as a method of influencing a State’s policy choices. Our cases have identified a variety of methods, short of outright coercion, by which Congress may urge a State to adopt a legislative program consistent with federal interests. Two of these methods are of particular relevance here.

[26] First, under Congress’ spending power, “Congress may attach conditions on the receipt of federal funds.” South Dakota v. Dole, 483 U. S., at 206. Such conditions must (among other requirements) bear some relationship to the purpose of the federal spending.

**

[27] Second, where Congress has the authority to regulate private activity under the Commerce Clause, we have recognized Congress’ power to offer States the choice of regulating that activity according to federal standards or having state law pre-empted by federal regulation.. These include the Clean Water Act, the Occupational Safety and Health Act of 1970, the Resource Conservation and Recovery Act of 1976,and the Alaska National Interest Lands Conservation Act.

[28] By either of these methods, as by any other permissible method of encouraging a State to conform to federal policy choices, the residents of the State retain the ultimate decision as to whether or not the State will comply. If a State’s citizens view federal policy as sufficiently contrary to local interests, they may elect to decline a federal grant. If state residents would prefer their government to devote its attention and resources to problems other than those deemed important by Congress, they may choose to have the Federal Government rather than the State bear the expense of a federally mandated regulatory program, and they may continue to supplement that program to the extent state law is not pre-empted. Where Congress encourages state regulation rather than compelling it, state governments remain responsive to the local electorate’s preferences; state officials remain accountable to the people.

[29] By contrast, where the Federal Government compels States to regulate, the accountability of both state and federal officials is diminished. If the citizens of New York, for example, do not consider that making provision for the disposal of radioactive waste is in their best interest, they may elect state officials who share their view. That view can always be pre-empted under the Supremacy Clause if it is contrary to the national view, but in such a case it is the Federal Government that makes the decision in full view of the public, and it will be federal officials that suffer the consequences if the decision turns out to be detrimental or unpopular. But where the Federal Government directs the States to regulate, it may be state officials who will bear the brunt of public disapproval, while the federal officials who devised the regulatory program may remain insulated from the electoral ramifications of their decision. Accountability is thus diminished when, due to federal coercion, elected state officials cannot regulate in accordance with the views of the local electorate in matters not pre-empted by federal regulation.

[30] With these principles in mind, we turn to the three challenged provisions of the Low-Level Radioactive Waste Policy Amendments Act of 1985.

**

[31] The first of these steps is an unexceptionable exercise of Congress’ power to authorize the States to burden interstate commerce.

[32] The second step, the Secretary’s collection of a percentage of the surcharge, is no more than a federal tax on interstate commerce, which petitioners do not claim to be an invalid exercise of either Congress’ commerce or taxing power.

**

[33] The third step is a conditional exercise of Congress’ authority under the Spending Clause:

[34] In the second set of incentives, Congress has authorized States and regional compacts with disposal sites gradually to increase the cost of access to the sites, and then to deny access altogether, to radioactive waste generated in States that do not meet federal deadlines. As a simple regulation, this provision would be within the power of Congress to authorize the States to discriminate against interstate commerce. Where federal regulation of private activity is within the scope of the Commerce Clause, we have recognized the ability of Congress to offer States the choice of regulating that activity according to federal standards or having state law pre-empted by federal regulation.

**

C

[35] The take title provision is of a different character. This third so-called “incentive” offers States, as an alternative to regulating pursuant to Congress’ direction, the option of taking title to and possession of the low level radioactive waste generated within their borders and becoming liable for all damages waste generators suffer as a result of the States’ failure to do so promptly. In this provision, Congress has crossed the line distinguishing encouragement from coercion.

**

[36] The take title provision offers state governments a “choice” of either accepting ownership of waste or regulating according to the instructions of Congress. Respondents do not claim that the Constitution would authorize Congress to impose either option as a freestanding requirement. On one hand, the Constitution would not permit Congress simply to transfer radioactive waste from generators to state governments. Such a forced transfer, standing alone, would in principle be no different than a congressionally compelled subsidy from state governments to radioactive waste producers. The same is true of the provision requiring the States to become liable for the generators’ damages. Standing alone, this provision would be indistinguishable from an Act of Congress directing the States to assume the liabilities of certain state residents. Either type of federal action would “commandeer” state governments into the service of federal regulatory purposes, and would for this reason be inconsistent with the Constitution’s division of authority between federal and state governments. On the other hand, the second alternative held out to state governments—regulating pursuant to Congress’ direction—would, standing alone, present a simple command to state governments to implement legislation enacted by Congress. As we have seen, the Constitution does not empower Congress to subject state governments to this type of instruction.

[37] Because an instruction to state governments to take title to waste, standing alone, would be beyond the authority of Congress, and because a direct order to regulate, standing alone, would also be beyond the authority of Congress, it follows that Congress lacks the power to offer the States a choice between the two.

**

[38] A choice between two unconstitutionally coercive regulatory techniques is no choice at all. Either way, “the Act commandeers the legislative processes of the States by directly compelling them to enact and enforce a federal regulatory program,” Hodel v. Virginia Surface Mining & Reclamation Assn., Inc., supra, at 288, an outcome that has never been understood to lie within the authority conferred upon Congress by the Constitution.

[39] Respondents emphasize the latitude given to the States to implement Congress’ plan. The Act enables the States to regulate pursuant to Congress’ instructions in any number of different ways. States may avoid taking title by contracting with sited regional compacts, by building a disposal site alone or as part of a compact, or by permitting private parties to build a disposal site. States that host sites may employ a wide range of designs and disposal methods, subject only to broad federal regulatory limits. This line of reasoning, however, only underscores the critical alternative a State lacks: A State may not decline to administer the federal program. No matter which path the State chooses, it must follow the direction of Congress.

[40] The take title provision appears to be unique. No other federal statute has been cited which offers a state government no option other than that of implementing legislation enacted by Congress. Whether one views the take title provision as lying outside Congress’ enumerated powers, or as infringing upon the core of state sovereignty reserved by the Tenth Amendment, the provision is inconsistent with the federal structure of our Government established by the Constitution.

IV

**

[41] [T]he United States argues that the Constitution’s prohibition of congressional directives to state governments can be overcome where the federal interest is sufficiently important to justify state submission.

**

[42] But whether or not a particularly strong federal interest enables Congress to bring state governments within the orbit of generally applicable federal regulation, no Member of the Court has ever suggested that such a federal interest would enable Congress to command a state government to enact state regulation. No matter how powerful the federal interest involved, the Constitution simply does not give Congress the authority to require the States to regulate. The Constitution instead gives Congress the authority to regulate matters directly and to pre-empt contrary state regulation. Where a federal interest is sufficiently strong to cause Congress to legislate, it must do so directly; it may not conscript state governments as its agents.

**

B

[43] The sited state respondents focus their attention on the process by which the Act was formulated. They correctly observe that public officials representing the State of New York lent their support to the Act’s enactment.

**

[44] Respondents note that the Act embodies a bargain among the sited and unsited States, a compromise to which New York was a willing participant and from which New York has reaped much benefit. Respondents then pose what appears at first to be a troubling question: How can a federal statute be found an unconstitutional infringement of state sovereignty when state officials consented to the statute’s enactment?

[45] The answer follows from an understanding of the fundamental purpose served by our Government’s federal structure. The Constitution does not protect the sovereignty of States for the benefit of the States or state governments as abstract political entities, or even for the benefit of the public officials governing the States. To the contrary, the Constitution divides authority between federal and state governments for the protection of individuals. State sovereignty is not just an end in itself: “Rather, federalism secures to citizens the liberties that derive from the diffusion of sovereign power.” Coleman v. Thompson, 501 U. S. 722, 759 (1991) (Blackmun, J., dissenting). “Just as the separation and independence of the coordinate branches of the Federal Government serve to prevent the accumulation of excessive power in any one branch, a healthy balance of power between the States and the Federal Government will reduce the risk of tyranny and abuse from either front.” Gregory v. Ashcroft, 501 U. S., at 458. See The Federalist No. 51, p. 323 (C. Rossiter ed. 1961).

[46] Where Congress exceeds its authority relative to the States, therefore, the departure from the constitutional plan cannot be ratified by the “consent” of state officials. An analogy to the separation of powers among the branches of the Federal Government clarifies this point. The Constitution’s division of power among the three branches is violated where one branch invades the territory of another, whether or not the encroached-upon branch approves the encroachment.

**

[47] State officials thus cannot consent to the enlargement of the powers of Congress beyond those enumerated in the Constitution.

**

[48] States are not mere political subdivisions of the United States. State governments are neither regional offices nor administrative agencies of the Federal Government. The positions occupied by state officials appear nowhere on the Federal Government’s most detailed organizational chart. The Constitution instead “leaves to the several States a residuary and inviolable sovereignty,” The Federalist No. 39, p. 245 (C. Rossiter ed. 1961), reserved explicitly to the States by the Tenth Amendment.

[49] Whatever the outer limits of that sovereignty may be, one thing is clear: The Federal Government may not compel the States to enact or administer a federal regulatory program. The Constitution permits both the Federal Government and the States to enact legislation regarding the disposal of low level radioactive waste. The Constitution enables the Federal Government to pre-empt state regulation contrary to federal interests, and it permits the Federal Government to hold out incentives to the States as a means of encouraging them to adopt suggested regulatory schemes. It does not, however, authorize Congress simply to direct the States to provide for the disposal of the radioactive waste generated within their borders.

**

[50] Justice White, with whom Justice Blackmun and Justice Stevens join, concurring in part and dissenting in part.

[51] The Court today affirms the constitutionality of two facets of the Low-Level Radioactive Waste Policy Amendments Act of 1985 (1985 Act). . . The Court strikes down and severs a third component of the 1985 Act, the “take title” provision, which requires a noncomplying State to take title to or to assume liability for its low-level radioactive waste if it fails to provide for the disposal of such waste by January 1, 1996. § 2021e(d)(2)(C). The Court deems this last provision unconstitutional under principles of federalism. Because I believe the Court has mischaracterized the essential inquiry, misanalyzed the inquiry it has chosen to undertake, and undervalued the effect the seriousness of this public policy problem should have on the constitutionality of the take title provision, I can only join Parts III—A and III—B, and I respectfully dissent from the rest of its opinion.

I

[52] My disagreement with the Court’s analysis begins at the basic descriptive level of how the legislation at issue in these cases came to be enacted. The Court goes some way toward setting out the bare facts, but its omissions cast the statutory context of the take title provision in the wrong light. To read the Court’s version of events, see ante, at 150-151, one would think that Congress was the sole proponent of a solution to the Nation’s low-level radioactive waste problem. Not so. The Low-Level Radioactive Waste Policy Act of 1980 (1980 Act), Pub. L. 96-573, 94 Stat. 3347, and its amendatory 1985 Act, resulted from the efforts of state leaders to achieve a state-based set of remedies to the waste problem. They sought not federal pre-emption or intervention, but rather congressional sanction of interstate compromises they had reached.

[53] The two signal events in 1979 that precipitated movement toward legislation were the temporary closing of the Nevada disposal site in July 1979, after several serious transportation-related incidents, and the temporary shutting of the Washington disposal site because of similar transportation and packaging problems in October 1979. At that time the facility in Barnwell, South Carolina, received approximately three-quarters of the Nation’s low-level radioactive waste, and the Governor ordered a 50 percent reduction in the amount his State’s plant would accept for disposal. National Governors’ Association Task Force on Low-Level Radioactive Waste Disposal, Low-Level Waste: A Program for Action 3 (Nov. 1980) (lodged with the Clerk of this Court) (hereinafter A Program for Action). The Governor of Washington threatened to shut down the Hanford, Washington, facility entirely by 1982 unless “some meaningful progress occurs toward” development of regional solutions to the waste disposal problem. Id., at 4, n. Only three sites existed in the country for the disposal of low-level radioactive waste, and the “sited” States confronted the undesirable alternatives either of continuing to be the dumping grounds for the entire Nation’s low-level waste or of eliminating or reducing in a constitutional manner the amount of waste accepted for disposal.

[54] The imminence of a crisis in low-level radioactive waste management cannot be overstated.

**

[55] Accordingly, the Governors’ Task Force urged that “each state should accept primary responsibility for the safe disposal of low-level radioactive waste generated within its borders” and that “the states should pursue a regional approach to the low-level waste disposal problem.”

**

[56] A movement thus arose to achieve a compromise between the sited and the unsited States, in which the sited States agreed to continue accepting waste in exchange for the imposition of stronger measures to guarantee compliance with the unsited States’ assurances that they would develop alternative disposal facilities.

**

[57] The bill that in large measure became the 1985 Act “represent[ed] the diligent negotiating undertaken by” the National Governors’ Association and “embodied” the “fundamentals of their settlement.”

**

[58] Unlike legislation that directs action from the Federal Government to the States, the 1980 and 1985 Acts reflected hard-fought agreements among States as refereed by Congress. The distinction is key, and the Court’s failure properly to characterize this legislation ultimately affects its analysis of the take title provision’s constitutionality.

**

B

[59] Even were New York not to be estopped from challenging the take title provision’s constitutionality, I am convinced that, seen as a term of an agreement entered into between the several States, this measure proves to be less constitutionally odious than the Court opines. First, the practical effect of New York’s position is that because it is unwilling to honor its obligations to provide in-state storage facilities for its low-level radioactive waste, other States with such plants must accept New York’s waste, whether they wish to or not. Otherwise, the many economically and socially beneficial producers of such waste in the State would have to cease their operations. The Court’s refusal to force New York to accept responsibility for its own problem inevitably means that some other State’s sovereignty will be impinged by it being forced, for public health reasons, to accept New York’s low-level radioactive waste. I do not understand the principle of federalism to impede the National Government from acting as referee among the States to prohibit one from bullying another.

[60] Moreover, it is utterly reasonable that, in crafting a delicate compromise between the three overburdened States that provided low-level radioactive waste disposal facilities and the rest of the States, Congress would have to ratify some punitive measure as the ultimate sanction for noncompliance. The take title provision, though surely onerous, does not take effect if the generator of the waste does not request such action, or if the State lives up to its bargain of providing a waste disposal facility either within the State or in another State pursuant to a regional compact arrangement or a separate contract. See 42 U. S. C. § 2021e(d)(2)(C).

**

[61] Ultimately, I suppose, the entire structure of our federal constitutional government can be traced to an interest in establishing checks and balances to prevent the exercise of tyranny against individuals. But these fears seem extremely far distant to me in a situation such as this. We face a crisis of national proportions in the disposal of low-level radioactive waste, and Congress has acceded to the wishes of the States by permitting local decisionmaking rather than imposing a solution from Washington. New York itself participated and supported passage of this legislation at both the gubernatorial and federal representative levels, and then enacted state laws specifically to comply with the deadlines and timetables agreed upon by the States in the 1985 Act. For me, the Court’s civics lecture has a decidedly hollow ring at a time when action, rather than rhetoric, is needed to solve a national problem.

**

V

[62] The ultimate irony of the decision today is that in its formalistically rigid obeisance to “federalism,” the Court gives Congress fewer incentives to defer to the wishes of state officials in achieving local solutions to local problems. This legislation was a classic example of Congress acting as arbiter among the States in their attempts to accept responsibility for managing a problem of grave import. The States urged the National Legislature not to impose from Washington a solution to the country’s low-level radioactive waste management problems. Instead, they sought a reasonable level of local and regional autonomy consistent with Art. I, § 10, cl. 3, of the Constitution. By invalidating the measure designed to ensure compliance for recalcitrant States, such as New York, the Court upsets the delicate compromise achieved among the States and forces Congress to erect several additional formalistic hurdles to clear before achieving exactly the same objective. Because the Court’s justifications for undertaking this step are unpersuasive to me, I respectfully dissent.

Justice STEVENS, concurring in part and dissenting in part.

[1] Under the Articles of Confederation, the Federal Government had the power to issue commands to the States. See Arts. VIII, IX. Because that indirect exercise of federal power proved ineffective, the Framers of the Constitution empowered the Federal Government to exercise legislative authority directly over individuals within the States, even though that direct authority constituted a greater intrusion on state sovereignty. Nothing in that history suggests that the Federal Government may not also impose its will upon the several States as it did under the Articles. The Constitution enhanced, rather than diminished, the power of the Federal Government.

[2] The notion that Congress does not have the power to issue “a simple command to state governments to implement legislation enacted by Congress,” is incorrect and unsound. There is no such limitation in the Constitution. The Tenth Amendment surely does not impose any limit on Congress’ exercise of the powers delegated to it by Article I. Nor does the structure of the constitutional order or the values of federalism mandate such a formal rule. To the contrary, the Federal Government directs state governments in many realms. The Government regulates state-operated railroads, state school systems, state prisons, state elections, and a host of other state functions. Similarly, there can be no doubt that, in time of war, Congress could either draft soldiers itself or command the States to supply their quotas of troops. I see no reason why Congress may not also command the States to enforce federal water and air quality standards or federal standards for the disposition of low-level radioactive wastes.

**

JUSTICE SCALIA, delivered the opinion of the Court.

[1] The question presented in these cases is whether certain interim provisions of the Brady Handgun Violence Prevention Act, Pub. L. 103-159, 107 Stat. 1536, commanding state and local law enforcement officers to conduct background checks on prospective handgun purchasers and to perform certain related tasks, violate the Constitution.

I

[2] The Gun Control Act of 1968 (GCA), establishes a detailed federal scheme governing the distribution of firearms. It prohibits firearms dealers from transferring handguns to any person under 21, not resident in the dealer’s State, or prohibited by state or local law from purchasing or possessing firearms, § 922(b). It also forbids possession of a firearm by, and transfer of a firearm to, convicted felons, fugitives from justice, unlawful users of controlled substances, persons adjudicated as mentally defective or committed to mental institutions, aliens unlawfully present in the United States, persons dishonorably discharged from the Armed Forces, persons who have renounced their citizenship, and persons who have been subjected to certain restraining orders or been convicted of a misdemeanor offense involving domestic violence.

[3] In 1993, Congress amended the GCA by enacting the Brady Act. The Act requires the Attorney General to establish a national instant background-check system by November 30, 1998, and immediately puts in place certain interim provisions until that system becomes operative. Under the interim provisions, a firearms dealer who proposes to transfer a handgun must first: (1) receive from the transferee a statement (the Brady Form), § 922(s)(1)(A)(i)(I), containing the name, address, and date of birth of the proposed transferee along with a sworn statement that the transferee is not among any of the classes of prohibited purchasers; (2) verify the identity of the transferee by examining an identification document; and (3) provide the “chief law enforcement officer” (CLEO) of the transferee’s residence with notice of the contents (and a copy) of the Brady Form. With some exceptions, the dealer must then wait five business days before consummating the sale, unless the CLEO earlier notifies the dealer that he has no reason to believe the transfer would be illegal.

**

[4] When a CLEO receives the required notice of a proposed transfer from the firearms dealer, the CLEO must “make a reasonable effort to ascertain within 5 business days whether receipt or possession would be in violation of the law, including research in whatever State and local recordkeeping systems are available and in a national system designated by the Attorney General.” The Act does not require the CLEO to take any particular action if he determines that a pending transaction would be unlawful; he may notify the firearms dealer to that effect, but is not required to do so. If, however, the CLEO notifies a gun dealer that a prospective purchaser is ineligible to receive a handgun, he must, upon request, provide the would-be purchaser with a written statement of the reasons for that determination. § 922(s)(6)(C). Moreover, if the CLEO does not discover any basis for objecting to the sale, he must destroy any records in his possession relating to the transfer, including his copy of the Brady Form. Under a separate provision of the GCA, any person who “knowingly violates [the section of the GCA amended by the Brady Act] shall be fined under this title, imprisoned for not more than 1 year, or both.”

[5] Petitioners Jay Printz and Richard Mack, the CLEOs for Ravalli County, Montana, and Graham County, Arizona, respectively, filed separate actions challenging the constitutionality of the Brady Act’s interim provisions.

**

[6] From the description set forth above, it is apparent that the Brady Act purports to direct state law enforcement officers to participate, albeit only temporarily, in the administration of a federally enacted regulatory scheme.

**

[7] Petitioners here object to being pressed into federal service, and contend that congressional action compelling state officers to execute federal laws is unconstitutional. Because there is no constitutional text speaking to this precise question, the answer to the CLEOs’ challenge must be sought in historical understanding and practice, in the structure of the Constitution, and in the jurisprudence of this Court. We treat those three sources, in that order, in this and the next two sections of this opinion.

[8] Petitioners contend that compelled enlistment of state executive officers for the administration of federal programs is, until very recent years at least, unprecedented. The Government contends, to the contrary, that “the earliest Congresses enacted statutes that required the participation of state officials in the implementation of federal laws.” The Government’s contention demands our careful consideration, since early congressional enactments “provid[e] `contemporaneous and weighty evidence’ of the Constitution’s meaning.” . . . Conversely if, as petitioners contend, earlier Congresses avoided use of this highly attractive power, we would have reason to believe that the power was thought not to exist.

[9] The Government observes that statutes enacted by the first Congresses required state courts to record applications for citizenship, to transmit abstracts of citizenship applications and other naturalization records to the Secretary of State, and to register aliens seeking naturalization and issue certificates of registry. It may well be, however, that these requirements applied only in States that authorized their courts to conduct naturalization proceedings.

[10] Other statutes of that era apparently or at least arguably required state courts to perform functions unrelated to naturalization, such as resolving controversies between a captain and the crew of his ship concerning the seaworthiness of the vessel, hearing the claims of slave owners who had apprehended fugitive slaves and issuing certificates authorizing the slave’s forced removal to the State from which he had fled, taking proof of the claims of Canadian refugees who had assisted the United States during the Revolutionary War, and ordering the deportation of alien enemies in times of war.

[11] These early laws establish, at most, that the Constitution was originally understood to permit imposition of an obligation on state judges to enforce federal prescriptions, insofar as those prescriptions related to matters appropriate for the judicial power. That assumption was perhaps implicit in one of the provisions of the Constitution, and was explicit in another. In accord with the so-called Madisonian Compromise, Article III, § 1, established only a Supreme Court, and made the creation of lower federal courts optional with the Congress—even though it was obvious that the Supreme Court alone could not hear all federal cases throughout the United States. And the Supremacy Clause, Art. VI, cl. 2, announced that “the Laws of the United States . . . shall be the supreme Law of the Land; and the Judges in every State shall be bound thereby.” It is understandable why courts should have been viewed distinctively in this regard; unlike legislatures and executives, they applied the law of other sovereigns all the time.

**

[12] For these reasons, we do not think the early statutes imposing obligations on state courts imply a power of Congress to impress the state executive into its service. Indeed, it can be argued that the numerousness of these statutes, contrasted with the utter lack of statutes imposing obligations on the States’ executive (notwithstanding the attractiveness of that course to Congress), suggests an assumed absence of such power. The only early federal law the Government has brought to our attention that imposed duties on state executive officers is the Extradition Act of 1793, which required the “executive authority” of a State to cause the arrest and delivery of a fugitive from justice upon the request of the executive authority of the State from which the fugitive had fled. That was in direct implementation, however, of the Extradition Clause of the Constitution itself, see Art. IV, § 2.

**

[13] To complete the historical record, we must note that there is not only an absence of executive-commandeering statutes in the early Congresses, but there is an absence of them in our later history as well, at least until very recent years.

**

[14] We turn next to consideration of the structure of the Constitution, to see if we can discern among its “essential postulate[s],” a principle that controls the present cases.

[15] It is incontestible that the Constitution established a system of “dual sovereignty.”Gregory v. Ashcroft (1991). Although the States surrendered many of their powers to the new Federal Government, they retained “a residuary and inviolable sovereignty,” The Federalist No. 39, at 245 (J. Madison).

**

[16] [T]he Framers rejected the concept of a central government that would act upon and through the States, and instead designed a system in which the State and Federal Governments would exercise concurrent authority over the people—who were, in Hamilton’s words, “the only proper objects of government,” The Federalist No. 15, at 109.

**

[17] “[T]he Framers explicitly chose a Constitution that confers upon Congress the power to regulate individuals, not States.” The great innovation of this design was that “our citizens would have two political capacities, one state and one federal, each protected from incursion by the other”—”a legal system unprecedented in form and design, establishing two orders of government, each with its own direct relationship, its own privity, its own set of mutual rights and obligations to the people who sustain it and are governed by it.” U. S. Term Limits, Inc. v. Thornton, 514 U. S. 779, 838 (1995) (Kennedy, J., concurring). The Constitution thus contemplates that a State’s government will represent and remain accountable to its own citizens.

**

[18] This separation of the two spheres is one of the Constitution’s structural protections of liberty. “Just as the separation and independence of the coordinate branches of the Federal Government serve to prevent the accumulation of excessive power in any one branch, a healthy balance of power between the States and the Federal Government will reduce the risk of tyranny and abuse from either front.” Gregory,501 U. S., at 458.

**

[19] The power of the Federal Government would be augmented immeasurably if it were able to impress into its service—and at no cost to itself—the police officers of the 50 States.

**

[20] The Constitution does not leave to speculation who is to administer the laws enacted by Congress; the President, it says, “shall take Care that the Laws be faithfully executed,” Art. II, § 3, personally and through officers whom he appoints (save for such inferior officers as Congress may authorize to be appointed by the “Courts of Law” or by “the Heads of Departments” who are themselves Presidential appointees), Art. II, § 2. The Brady Act effectively transfers this responsibility to thousands of CLEOs in the 50 States, who are left to implement the program without meaningful Presidential control (if indeed meaningful Presidential control is possible without the power to appoint and remove). The insistence of the Framers upon unity in the Federal Executive—to ensure both vigor and accountability—is well known. That unity would be shattered, and the power of the President would be subject to reduction, if Congress could act as effectively without the President as with him, by simply requiring state officers to execute its laws.

[21] The dissent of course resorts to the last, best hope of those who defend ultra vires congressional action, the Necessary and Proper Clause. It reasons that the power to regulate the sale of handguns under the Commerce Clause, coupled with the power to “make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers,” Art. I, § 8, conclusively establishes the Brady Act’s constitutional validity, because the Tenth Amendment imposes no limitations on the exercise of delegated powers but merely prohibits the exercise of powers “not delegated to the United States.” What destroys the dissent’s Necessary and Proper Clause argument, however, is not the Tenth Amendment but the Necessary and Proper Clause itself. When a “La[w] . . . for carrying into Execution” the Commerce Clause violates the principle of state sovereignty reflected in the various constitutional provisions we mentioned earlier, it is not a “La[w] . . . proper for carrying into Execution the Commerce Clause,” and is thus, in the words of The Federalist, “merely [an] ac[t] of usurpation” which “deserve[s] to be treated as such.” We in fact answered the dissent’s Necessary and Proper Clause argument in New York: “[E]ven where Congress has the authority under the Constitution to pass laws requiring or prohibiting certain acts, it lacks the power directly to compel the States to require or prohibit those acts. . . .

**

[22] Federal commandeering of state governments is such a novel phenomenon that this Court’s first experience with it did not occur until the 1970’s, when the Environmental Protection Agency promulgated regulations requiring States to prescribe auto emissions testing, monitoring and retrofit programs, and to designate preferential bus and carpool lanes. The Courts of Appeals for the Fourth and Ninth Circuits invalidated the regulations on statutory grounds in order to avoid what they perceived to be grave constitutional issues; and the District of Columbia Circuit invalidated the regulations on both constitutional and statutory grounds. After we granted certiorari to review the statutory and constitutional validity of the regulations, the Government declined even to defend them, and instead rescinded some and conceded the invalidity of those that remained, leading us to vacate the opinions below and remand for consideration of mootness. EPA v. Brown (1977) (per curiam).

[23] Although we had no occasion to pass upon the subject in Brown, later opinions of ours have made clear that the Federal Government may not compel the States to implement, by legislation or executive action, federal regulatory programs. In Hodel v.Virginia Surface Mining & Reclamation Assn., Inc., (1981), and FERC v. Mississippi (1982), we sustained statutes against constitutional challenge only after assuring ourselves that they did not require the States to enforce federal law.

**

[24] At issue in New York v. United States (1992), were the so-called “take title” provisions of the Low-Level Radioactive Waste Policy Amendments Act of 1985, which required States either to enact legislation providing for the disposal of radioactive waste generated within their borders, or to take title to, and possession of, the waste—effectively requiring the States either to legislate pursuant to Congress’s directions, or to implement an administrative solution. We concluded that Congress could constitutionally require the States to do neither. “The Federal Government,” we held, “may not compel the States to enact or administer a federal regulatory program.”

[25] The Government contends that New York is distinguishable on the following ground: Unlike the “take title” provisions invalidated there, the background-check provision of the Brady Act does not require state legislative or executive officials to make policy, but instead issues a final directive to state CLEOs. It is permissible, the Government asserts, for Congress to command state or local officials to assist in the implementation of federal law so long as “Congress itself devises a clear legislative solution that regulates private conduct” and requires state or local officers to provide only “limited, non-policy making help in enforcing that law.” “[T]he constitutional line is crossed only when Congress compels the States to make law in their sovereign capacities.”

[26] The Government’s distinction between “making” law and merely “enforcing” it, between “policy making” and mere “implementation,” is an interesting one. It is perhaps not meant to be the same as, but it is surely reminiscent of, the line that separates proper congressional conferral of Executive power from unconstitutional delegation of legislative authority for federal separation-of-powers purposes. This Court has not been notably successful in describing the latter line; indeed, some think we have abandoned the effort to do so. We are doubtful that the new line the Government proposes would be any more distinct. Executive action that has utterly no policy making component is rare, particularly at an executive level as high as a jurisdiction’s chief law enforcement officer. Is it really true that there is no policymaking involved in deciding, for example, what “reasonable efforts” shall be expended to conduct a background check?

**

[27] Even assuming, moreover, that the Brady Act leaves no “policymaking” discretion with the States, we fail to see how that improves rather than worsens the intrusion upon state sovereignty. Preservation of the States as independent and autonomous political entities is arguably less undermined by requiring them to make policy in certain fields than (as Judge Sneed aptly described it over two decades ago) by “reduc[ing] [them] to puppets of a ventriloquist Congress,” Brown v. EPA. It is an essential attribute of the States’ retained sovereignty that they remain independent and autonomous within their proper sphere of authority. It is no more compatible with this independence and autonomy that their officers be “dragooned” (as Judge Fernandez put it in his dissent below, 66 F. 3d, at 1035) into administering federal law, than it would be compatible with the independence and autonomy of the United States that its officers be impressed into service for the execution of state laws.

**

[28] The Government also maintains that requiring state officers to perform discrete, ministerial tasks specified by Congress does not violate the principle of New York because it does not diminish the accountability of state or federal officials. This argument fails even on its own terms. By forcing state governments to absorb the financial burden of implementing a federal regulatory program, Members of Congress can take credit for “solving” problems without having to ask their constituents to pay for the solutions with higher federal taxes. And even when the States are not forced to absorb the costs of implementing a federal program, they are still put in the position of taking the blame for its burdensomeness and for its defects.

**

[29] The Brady Act, the dissent asserts, is different from the “take title” provisions invalidated in New York because the former is addressed to individuals—namely, CLEOs—while the latter were directed to the State itself. That is certainly a difference, but it cannot be a constitutionally significant one. While the Brady Act is directed to “individuals,” it is directed to them in their official capacities as state officers; it controls their actions, not as private citizens, but as the agents of the State.

**

[30] Finally, the Government puts forward a cluster of arguments that can be grouped under the heading: “The Brady Act serves very important purposes, is most efficiently administered by CLEOs during the interim period, and places a minimal and only temporary burden upon state officers.” There is considerable disagreement over the extent of the burden, but we need not pause over that detail. Assuming all the mentioned factors were true, they might be relevant if we were evaluating whether the incidental application to the States of a federal law of general applicability excessively interfered with the functioning of state governments. But where, as here, it is the whole object of the law to direct the functioning of the state executive, and hence to compromise the structural framework of dual sovereignty, such a “balancing” analysis is inappropriate.

**

[31] We adhere to that principle today, and conclude categorically, as we concluded categorically in New York: “The Federal Government may not compel the States to enact or administer a federal regulatory program.” The mandatory obligation imposed on CLEOs to perform background checks on prospective handgun purchasers plainly runs afoul of that rule.

**

[32] We held in New York that Congress cannot compel the States to enact or enforce a federal regulatory program. Today we hold that Congress cannot circumvent that prohibition by conscripting the States’ officers directly. The Federal Government may neither issue directives requiring the States to address particular problems, nor command the States’ officers, or those of their political subdivisions, to administer or enforce a federal regulatory program. It matters not whether policymaking is involved, and no case-bycase weighing of the burdens or benefits is necessary; such commands are fundamentally incompatible with our constitutional system of dual sovereignty. Accordingly, the judgment of the Court of Appeals for the Ninth Circuit is reversed.

[33] It is so ordered.

JUSTICE O’CONNOR, concurring.

**

[1] Our holding, of course, does not spell the end of the objectives of the Brady Act. States and chief law enforcement officers may voluntarily continue to participate in the federal program. Moreover, the directives to the States are merely interim provisions scheduled to terminate November 30, 1998. Congress is also free to amend the interim program to provide for its continuance on a contractual basis with the States if it wishes, as it does with a number of other federal programs.

[2] In addition, the Court appropriately refrains from deciding whether other purely ministerial reporting requirements imposed by Congress on state and local authorities pursuant to its Commerce Clause powers are similarly invalid. See, e. g.,42 U. S. C. § 5779(a) (requiring state and local law enforcement agencies to report cases of missing children to the Department of Justice). The provisions invalidated here, however, which directly compel state officials to administer a federal regulatory program, utterly fail to adhere to the design and structure of our constitutional scheme.

**

JUSTICE STEVENS, with whom JUSTICE SOUTER, JUSTICE GINSBURG, and JUSTICE BRYER join, dissenting.

[1] When Congress exercises the powers delegated to it by the Constitution, it may impose affirmative obligations on executive and judicial officers of state and local governments as well as ordinary citizens. This conclusion is firmly supported by the text of the Constitution, the early history of the Nation, decisions of this Court, and a correct understanding of the basic structure of the Federal Government.

[2] These cases do not implicate the more difficult questions associated with congressional coercion of state legislatures addressed in New York v. UnitedStates, 505 U. S. 144 (1992).

[3] Indeed, since the ultimate issue is one of power, we must consider its implications in times of national emergency. Matters such as the enlistment of air raid wardens, the administration of a military draft, the mass inoculation of children to forestall an epidemic, or perhaps the threat of an international terrorist, may require a national response before federal personnel can be made available to respond

**

[4] Perhaps today’s majority would suggest that no such emergency is presented by the facts of these cases. But such a suggestion is itself an expression of a policy judgment. And Congress’ view of the matter is quite different from that implied by the Court today.

**

[5] The text of the Constitution provides a sufficient basis for a correct disposition of these cases.

[6] Article I, § 8, grants Congress the power to regulate commerce among the States. Putting to one side the revisionist views expressed by Justice Thomas in his concurring opinion in United States v. Lopez (1995), there can be no question that that provision adequately supports the regulation of commerce in handguns effected by the Brady Act. Moreover, the additional grant of authority in that section of the Constitution “[t]o make all Laws which shall be necessary and proper for carrying into Execution the foregoing Powers” is surely adequate to support the temporary enlistment of local police officers in the process of identifying persons who should not be entrusted with the possession of handguns. In short, the affirmative delegation of power in Article I provides ample authority for the congressional enactment.

**

[7] There is not a clause, sentence, or paragraph in the entire text of the Constitution of the United States that supports the proposition that a local police officer can ignore a command contained in a statute enacted by Congress pursuant to an express delegation of power enumerated in Article I.

**

[8] Indeed, the historical materials strongly suggest that the founders intended to enhance the capacity of the Federal Government by empowering it—as a part of the new authority to make demands directly on individual citizens—to act through local officials. Hamilton made clear that the new Constitution, “by extending the authority of the federal head to the individual citizens of the several States, will enable the government to employ the ordinary magistracy of each in the execution of its laws.” The Federalist No. 27, at 180. Hamilton’s meaning was unambiguous; the Federal Government was to have the power to demand that local officials implement national policy programs. As he went on to explain: “It is easy to perceive that this will tend to destroy, in the common apprehension, all distinction between the sources from which [the State and Federal Governments] might proceed; and will give the federal government the same advantage for securing a due obedience to its authority which is enjoyed by the government of each State.”

**

[9] [T]he majority’s opinion consists almost entirely of arguments against the substantial evidence weighing in opposition to its view; the Court’s ruling is strikingly lacking in affirmative support. Absent even a modicum of textual foundation for its judicially crafted constitutional rule, there should be a presumption that if the Framers had actually intended such a rule, at least one of them would have mentioned it.

**

[10] The Court’s “structural” arguments are not sufficient to rebut that presumption. The fact that the Framers intended to preserve the sovereignty of the several States simply does not speak to the question whether individual state employees may be required to perform federal obligations, such as registering young adults for the draft, 40 Stat. 80-81, creating state emergency response commissions designed to manage the release of hazardous substances, collecting and reporting data on underground storage tanks that may pose an environmental hazard, and reporting traffic fatalities, and missing children, to a federal agency.

[11] As we explained in Garcia v. San Antonio Metropolitan Transit Authority (1985): “[T]he principal means chosen by the Framers to ensure the role of the States in the federal system lies in the structure of the Federal Government itself. It is no novelty to observe that the composition of the Federal Government was designed in large part to protect the States from overreaching by Congress.” Given the fact that the Members of Congress are elected by the people of the several States, with each State receiving an equivalent number of Senators in order to ensure that even the smallest States have a powerful voice in the Legislature, it is quite unrealistic to assume that they will ignore the sovereignty concerns of their constituents. It is far more reasonable to presume that their decisions to impose modest burdens on state officials from time to time reflect a considered judgment that the people in each of the States will benefit therefrom.

**

[12] With colorful hyperbole, the Court suggests that the unity in the Executive Branch of the Federal Government “would be shattered, and the power of the President would be subject to reduction, if Congress could . . . requir[e] state officers to execute its laws.” Ante, at 923. Putting to one side the obvious tension between the majority’s claim that impressing state police officers will unduly tip the balance of power in favor of the federal sovereign and this suggestion that it will emasculate the Presidency, the Court’s reasoning contradicts New York v. United States.

[13] That decision squarely approved of cooperative federalism programs, designed at the national level but implemented principally by state governments. New York disapproved of a particular method of putting such programs into place, not the existence of federal programs implemented locally. See 505 U. S., at 166 (“Our cases have identified a variety of methods . . . by which Congress may urge a State to adopt a legislative program consistent with federal interests”). Indeed, nothing in the majority’s holding calls into question the three mechanisms for constructing such programs that New York expressly approved. Congress may require the States to implement its programs as a condition of federal spending, in order to avoid the threat of unilateral federal action in the area, or as a part of a program that affects States and private parties alike.

**

[14] [U]nlike any precedent in which the Court has held that Congress exceeded its powers, [this case] merely involves the imposition of modest duties on individual officers. The Court seems to accept the fact that Congress could require private persons, such as hospital executives or school administrators, to provide arms merchants with relevant information about a prospective purchaser’s fitness to own a weapon; indeed, the Court does not disturb the conclusion that flows directly from our prior holdings that the burden on police officers would be permissible if a similar burden were also imposed on private parties with access to relevant data.

**

[15] The provision of the Brady Act that crosses the Court’s newly defined constitutional threshold is more comparable to a statute requiring local police officers to report the identity of missing children to the Crime Control Center of the Department of Justice than to an offensive federal command to a sovereign State. If Congress believes that such a statute will benefit the people of the Nation, and serve the interests of cooperative federalism better than an enlarged federal bureaucracy, we should respect both its policy judgment and its appraisal of its constitutional power.

[16] Accordingly, I respectfully dissent.

JUSTICE SOUTER, dissenting.

**

[1] In deciding these cases, which I have found closer than I had anticipated, it is The Federalist that finally determines my position. I believe that the most straightforward reading of No. 27 is authority for the Government’s position here, and that this reading is both supported by No. 44 and consistent with Nos. 36 and 45.

[2] Hamilton in No. 27 first notes that because the new Constitution would authorize the National Government to bind individuals directly through national law, it could “employ the ordinary magistracy of each [State] in the execution of its laws.” The Federalist No. 27, p. 174 (J. Cooke ed. 1961) (A. Hamilton). Were he to stop here, he would not necessarily be speaking of anything beyond the possibility of cooperative arrangements by agreement. But he then addresses the combined effect of the proposed Supremacy Clause, U. S. Const., Art. VI, cl. 2, and state officers’ oath requirement, U. S. Const., Art. VI, cl. 3, and he states that “the Legislatures, Courts and Magistrates of the respective members will be incorporated into the operations of the national government, as far as its just and constitutional authority extends; and will be rendered auxiliary to the enforcement of its laws.” The Federalist No. 27, at 174-175 (emphasis in original). The natural reading of this language is not merely that the officers of the various branches of state governments may be employed in the performance of national functions; Hamilton says that the state governmental machinery “will be incorporated” into the Nation’s operation, and because the “auxiliary” status of the state officials will occur because they are “bound by the sanctity of an oath,” id., at 175, I take him to mean that their auxiliary functions will be the products of their obligations thus undertaken to support federal law, not of their own, or the States’, unfettered choices. Madison in No. 44 supports this reading in his commentary on the oath requirement. He asks why state magistrates should have to swear to support the National Constitution, when national officials will not be required to oblige themselves to support the state counterparts. His answer is that national officials “will have no agency in carrying the State Constitutions into effect. The members and officers of the State Governments, on the contrary, will have an essential agency in giving effect to the Federal Constitution.”Id., No. 44, at 307 (J. Madison). He then describes the state legislative “agency” as action necessary for selecting the President, see U. S. Const., Art. II, § 1, and the choice of Senators, see U. S. Const., Art. I, § 3 (repealed by Amdt. 17). The Federalist No. 44, at 307. The Supremacy Clause itself, of course, expressly refers to the state judges’ obligations under federal law, and other numbers of The Federalist give examples of state executive “agency” in the enforcement of national revenue laws.

**

[3] In the light of all these passages, I cannot persuade myself that the statements from No. 27 speak of anything less than the authority of the National Government, when exercising an otherwise legitimate power (the commerce power, say), to require state “auxiliaries” to take appropriate action.

**

JUSTICE BREYER, with whom Justice Stevens joins, dissenting.

[1] I would add to the reasons Justice Stevens sets forth the fact that the United States is not the only nation that seeks to reconcile the practical need for a central authority with the democratic virtues of more local control. At least some other countries, facing the same basic problem, have found that local control is better maintained through application of a principle that is the direct opposite of the principle the majority derives from the silence of our Constitution. The federal systems of Switzerland, Germany, and the European Union, for example, all provide that constituent states, not federal bureaucracies, will themselves implement many of the laws, rules, regulations, or decrees enacted by the central “federal” body.

[2] Of course, we are interpreting our own Constitution, not those of other nations, and there may be relevant political and structural differences between their systems and our own. But their experience may nonetheless cast an empirical light on the consequences of different solutions to a common legal problem—in this case the problem of reconciling central authority with the need to preserve the liberty-enhancing autonomy of a smaller constituent governmental entity. And that experience here offers empirical confirmation of the implied answer to a question Justice Stevens asks: Why, or how, would what the majority sees as a constitutional alternative—the creation of a new federal gun-law bureaucracy, or the expansion of an existing federal bureaucracy—better promote either state sovereignty or individual liberty? See ante, at 945, 959 (Stevens, J., dissenting).

[3] As comparative experience suggests, there is no need to interpret the Constitution as containing an absolute principle—forbidding the assignment of virtually any federal duty to any state official. Nor is there a need to read the Brady Act as permitting the Federal Government to overwhelm a state civil service. The statute uses the words “reasonable effort,” 18 U. S. C. § 922(s)(2)—words that easily can encompass the considerations of, say, time or cost necessary to avoid any such result.

[4] Regardless, as Justice Stevens points out, the Constitution itself is silent on the matter. Precedent supports the Government’s position here. And the fact that there is not more precedent—that direct federal assignment of duties to state officers is not common—likely reflects, not a widely shared belief that any such assignment is incompatible with basic principles of federalism, but rather a widely shared practice of assigning such duties in other ways. Thus, there is neither need nor reason to find in the Constitution an absolute principle, the inflexibility of which poses a surprising and technical obstacle to the enactment of a law that Congress believed necessary to solve an important national problem.

Justice ALITO delivered the opinion of the Court.
[1] The State of New Jersey wants to legalize sports gambling at casinos and horseracing tracks, but a federal law, the Professional and Amateur Sports Protection Act, generally makes it unlawful for a State to "authorize" sports gambling schemes. 28 U.S.C. § 3702(1). We must decide whether this provision is compatible with the system of "dual sovereignty" embodied in the Constitution.

I
A
[2] Americans have never been of one mind about gambling, and attitudes have swung back and forth. By the end of the 19th *1469 century, gambling was largely banned throughout the country,[1] but beginning in the 1920s and 1930s, laws prohibiting gambling were gradually loosened.

[3] New Jersey's experience is illustrative. In 1897, New Jersey adopted a constitutional amendment that barred all gambling in the State.[2] But during the Depression, the State permitted parimutuel betting on horse races as a way of increasing state revenue,[3] and in 1953, churches and other nonprofit organizations were allowed to host bingo games.[4] In 1970, New Jersey became the third State to run a state lottery,[5] and within five years, 10 other States followed suit.[6]

[4] By the 1960s, Atlantic City, "once the most fashionable resort of the Atlantic Coast," had fallen on hard times,[7] and casino gambling came to be seen as a way to revitalize the city.[8] In 1974, a referendum on statewide legalization failed,[9] but two years later, voters approved a narrower measure allowing casino gambling in Atlantic City alone.[10] At that time, Nevada was the only other State with legal casinos,[11] and thus for a while the Atlantic City casinos had an east coast monopoly. "With 60 million people living within a one-tank car trip away," Atlantic City became "the most popular tourist destination in the United States."[12] But that favorable situation eventually came to an end.

[5] With the enactment of the Indian Gaming Regulatory Act in 1988, 25 U.S.C. § 2701 et seq., casinos opened on Indian land throughout the country. Some were located within driving distance of Atlantic City,[13] and nearby States (and many others) legalized casino gambling.[14] But Nevada remained the only state venue for legal sports gambling in casinos, and sports gambling is immensely popular.[15]

[6] Sports gambling, however, has long had strong opposition. Opponents argue that it is particularly addictive and especially attractive to young people with a strong interest in sports,[16] and in the past gamblers *1470 corrupted and seriously damaged the reputation of professional and amateur sports.[17] Apprehensive about the potential effects of sports gambling, professional sports leagues and the National Collegiate Athletic Association (NCAA) long opposed legalization.[18]

B
[7] By the 1990s, there were signs that the trend that had brought about the legalization of many other forms of gambling might extend to sports gambling,[19] and this sparked federal efforts to stem the tide. Opponents of sports gambling turned to the legislation now before us, the Professional and Amateur Sports Protection Act (PASPA). 28 U.S.C. § 3701 et seq. PASPA's proponents argued that it would protect young people, and one of the bill's sponsors, Senator Bill Bradley of New Jersey, a former college and professional basketball star, stressed that the law was needed to safeguard the integrity of sports.[20] The Department of Justice opposed the bill,[21] but it was passed and signed into law.

[8] PASPA's most important provision, part of which is directly at issue in these cases, makes it "unlawful" for a State or any of its subdivisions[22] "to sponsor, operate, advertise, promote, license, or authorize by law or compact ... a lottery, sweepstakes, or other betting, gambling, or wagering scheme based ... on" competitive sporting events. § 3702(1). In parallel, § 3702(2) makes it "unlawful" for "a person to sponsor, operate, advertise, or promote" those same gambling schemes[23] — but only if this is done "pursuant to the law or compact of a governmental entity." PASPA does not make sports gambling a federal crime (and thus was not anticipated to impose a significant law enforcement burden on the Federal Government).[24] Instead, PASPA *1471 allows the Attorney General, as well as professional and amateur sports organizations, to bring civil actions to enjoin violations. § 3703.

[9] At the time of PASPA's adoption, a few jurisdictions allowed some form of sports gambling. In Nevada, sports gambling was legal in casinos,[25] and three States hosted sports lotteries or allowed sports pools.[26] PASPA contains "grandfather" provisions allowing these activities to continue. § 3704(a)(1)-(2). Another provision gave New Jersey the option of legalizing sports gambling in Atlantic City — provided that it did so within one year of the law's effective date. § 3704(a)(3).[27]

[10] New Jersey did not take advantage of this special option, but by 2011, with Atlantic City facing stiff competition, the State had a change of heart. New Jersey voters approved an amendment to the State Constitution making it lawful for the legislature to authorize sports gambling, Art. IV, § 7, ¶ 2(D), (F), and in 2012 the legislature enacted a law doing just that, 2011 N.J. Laws p. 1723 (2012 Act).

[11] The 2012 Act quickly came under attack. The major professional sports leagues and the NCAA brought an action in federal court against the New Jersey Governor and other state officials (hereinafter New Jersey), seeking to enjoin the new law on the ground that it violated PASPA. In response, the State argued, among other things, that PASPA unconstitutionally infringed the State's sovereign authority to end its sports gambling ban. See National Collegiate Athletic Assn. v. Christie, 926 F.Supp.2d 551, 561 (D.N.J.2013).

[12] In making this argument, the State relied primarily on two cases, New York v. United States, 505 U.S. 144, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992), and Printz v. United States, 521 U.S. 898, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997), in which we struck down federal laws based on what has been dubbed the "anticommandeering" principle. In New York, we held that a federal law unconstitutionally ordered the State to regulate in accordance with federal standards, and in Printz, we found that another federal statute unconstitutionally compelled state officers to enforce federal law.

[13] Relying on these cases, New Jersey argued that PASPA is similarly flawed because it regulates a State's exercise of its lawmaking power by prohibiting it from modifying or repealing its laws prohibiting sports gambling. See National Collegiate Athletic Assn. v. Christie, 926 F.Supp.2d, at 561-562. The plaintiffs countered that PASPA is critically different from the commandeering cases because it does not command the States to take any affirmative act. Id., at 562. Without an affirmative federal command to do something, the plaintiffs insisted, there can be no claim of commandeering. Ibid.

[14] The District Court found no anticommandeering violation, id., at 569-573, and a divided panel of the Third Circuit affirmed, National Collegiate Athletic Assn. v. Christie, 730 F.3d 208 (2013) (Christie I). The panel thought it significant that PASPA does not impose any affirmative command. Id., at 231. In the words of the panel, "PASPA does not require or coerce the states to lift a finger." Ibid. (emphasis deleted). The panel recognized *1472 that an affirmative command (for example, "Do not repeal") can often be phrased as a prohibition ("Repeal is prohibited"), but the panel did not interpret PASPA as prohibiting the repeal of laws outlawing sports gambling. Id., at 232. A repeal, it thought, would not amount to "authoriz[ation]" and thus would fall outside the scope of § 3702(1). "[T]he lack of an affirmative prohibition of an activity," the panel wrote, "does not mean it is affirmatively authorized by law. The right to do that which is not prohibited derives not from the authority of the state but from the inherent rights of the people." Id., at 232 (emphasis deleted).

[15] New Jersey filed a petition for a writ of certiorari, raising the anticommandeering issue. Opposing certiorari, the United States told this Court that PASPA does not require New Jersey "to leave in place the state-law prohibitions against sports gambling that it had chosen to adopt prior to PASPA's enactment. To the contrary, New Jersey is free to repeal those prohibitions in whole or in part." Brief for United States in Opposition in Christie v. National Collegiate Athletic Assn., O.T. 2013, No. 13-967 etc., p. 11. See also Brief for Respondents in Opposition in No. 13-967 etc., p. 23 ("Nothing in that unambiguous language compels states to prohibit or maintain any existing prohibition on sports gambling"). We denied review. Christie v. National Collegiate Athletic Assn., 573 U.S. ___, 134 S.Ct. 2866, 189 L.Ed.2d 806 (2014).

[16] Picking up on the suggestion that a partial repeal would be allowed, the New Jersey Legislature enacted the law now before us. 2014 N.J. Laws p. 602 (2014 Act). The 2014 Act declares that it is not to be interpreted as causing the State to authorize, license, sponsor, operate, advertise, or promote sports gambling. Ibid. Instead, it is framed as a repealer. Specifically, it repeals the provisions of state law prohibiting sports gambling insofar as they concerned the "placement and acceptance of wagers" on sporting events by persons 21 years of age or older at a horseracing track or a casino or gambling house in Atlantic City. Ibid. The new law also specified that the repeal was effective only as to wagers on sporting events not involving a New Jersey college team or a collegiate event taking place in the State. Ibid.

[17] Predictably, the same plaintiffs promptly commenced a new action in federal court. They won in the District Court, National Collegiate Athletic Assn. v. Christie, 61 F.Supp.3d 488 (N.J.2014), and the case was eventually heard by the Third Circuit sitting en banc. The en banc court affirmed, finding that the new law, no less than the old one, violated PASPA by "author[izing]" sports gambling. National Collegiate Athletic Assn. v. Governor of N. J., 832 F.3d 389 (2016) (case below). The court was unmoved by the New Jersey Legislature's "artful[ ]" attempt to frame the 2014 Act as a repealer. Id., at 397. Looking at what the law "actually does," the court concluded that it constitutes an authorization because it "selectively remove[s] a prohibition on sports wagering in a manner that permissively channels wagering activity to particular locations or operators." Id., at 397, 401. The court disavowed some of the reasoning in the Christie I opinion, finding its discussion of "the relationship between a `repeal' and an `authorization' to have been too facile." 832 F.3d, at 401. But the court declined to say whether a repeal that was more complete than the 2014 Act would still amount to an authorization. The court observed that a partial repeal that allowed only "de minimis wagers between friends and family would not have nearly the type of authorizing effect" that it found in the 2014 Act, and it added: "We need not ... articulate a line whereby a partial repeal of a sports *1473 wagering ban amounts to an authorization under PASPA, if indeed such a line could be drawn." Id., at 402 (emphasis added).
Having found that the 2014 Act violates PASPA's prohibition of state authorization of sports gambling schemes, the court went on to hold that this prohibition does not contravene the anticommandeering principle because it "does not command states to take affirmative actions." Id., at 401.

[18] We granted review to decide the important constitutional question presented by these cases, sub nom. Christie v. National Collegiate Athletic Assn., 582 U.S. ___, 137 S.Ct. 2327, 198 L.Ed.2d 754 (2017).

II
[19] Before considering the constitutionality of the PASPA provision prohibiting States from "author[izing]" sports gambling, we first examine its meaning. The parties advance dueling interpretations, and this dispute has an important bearing on the constitutional issue that we must decide. Neither respondents nor the United States, appearing as an amicus in support of respondents, contends that the provision at issue would be constitutional if petitioners' interpretation is correct. Indeed, the United States expressly concedes that the provision is unconstitutional if it means what petitioners claim. Brief for United States 8, 19.

A
[20] Petitioners argue that the anti-authorization provision requires States to maintain their existing laws against sports gambling without alteration. One of the accepted meanings of the term "authorize," they point out, is "permit." Brief for Petitioners in No. 16-476, p. 42 (citing Black's Law Dictionary 133 (6th ed. 1990); Webster's Third New International Dictionary 146 (1992)). They therefore contend that any state law that has the effect of permitting sports gambling, including a law totally or partially repealing a prior prohibition, amounts to an authorization. Brief for Petitioners in No. 16-476, at 42.

[21] Respondents interpret the provision more narrowly. They claim that the primary definition of "authorize" requires affirmative action. Brief for Respondents 39. To authorize, they maintain, means "`[t]o empower; to give a right or authority to act; to endow with authority.'" Ibid. (quoting Black's Law Dictionary, at 133). And this, they say, is precisely what the 2014 Act does: It empowers a defined group of entities, and it endows them with the authority to conduct sports gambling operations.

[22] Respondents do not take the position that PASPA bans all modifications of old laws against sports gambling, Brief for Respondents 20, but just how far they think a modification could go is not clear. They write that a State "can also repeal or enhance [laws prohibiting sports gambling] without running afoul of PASPA" but that it "cannot `partially repeal' a general prohibition for only one or two preferred providers, or only as to sports-gambling schemes conducted by the state." Ibid. Later in their brief, they elaborate on this point:

[23] "If, for example, a state had an existing felony prohibition on all lotteries, it could maintain the law, it could repeal the law, it could downgrade the crime to a misdemeanor or increase the penalty.... But if the state modified its law, whether through a new authorization or through an amendment partially repealing the existing prohibition, to authorize the state to conduct a sports lottery, that modified law would be preempted." Id., at 31.

[24] *1474 The United States makes a similar argument. PASPA, it contends, does not prohibit a State from enacting a complete repeal because "one would not ordinarily say that private conduct is `authorized by law' simply because the government has not prohibited it." Brief for United States 17. But the United States claims that "[t]he 2014 Act's selective and conditional permission to engage in conduct that is generally prohibited certainly qualifies" as an authorization. Ibid. The United States does not argue that PASPA outlaws all partial repeals, but it does not set out any clear rule for distinguishing between partial repeals that constitute the "authorization" of sports gambling and those that are permissible. The most that it is willing to say is that a State could "eliminat[e] prohibitions on sports gambling involving wagers by adults or wagers below a certain dollar threshold." Id., at 29.

B
[25] In our view, petitioners' interpretation is correct: When a State completely or partially repeals old laws banning sports gambling, it "authorize[s]" that activity. This is clear when the state-law landscape at the time of PASPA's enactment is taken into account. At that time, all forms of sports gambling were illegal in the great majority of States, and in that context, the competing definitions offered by the parties lead to the same conclusion. The repeal of a state law banning sports gambling not only "permits" sports gambling (petitioners' favored definition); it also gives those now free to conduct a sports betting operation the "right or authority to act"; it "empowers" them (respondents' and the United States's definition).

[26] The concept of state "authorization" makes sense only against a backdrop of prohibition or regulation. A State is not rearded as authorizing everything that it does not prohibit or regulate. No one would use the term in that way. For example, no one would say that a State "authorizes" its residents to brush their teeth or eat apples or sing in the shower. We commonly speak of state authorization only if the activity in question would otherwise be restricted.[28]

[27] The United States counters that, even if the term "authorize," standing alone, is interpreted as petitioners claim, PASPA contains additional language that precludes that reading. The provision at issue refers to "authoriz[ation] by law," § 3702(1) (emphasis added), and the parallel provision governing private conduct, § 3702(2), applies to conduct done "pursuant to the law ... of a governmental entity." The United States maintains that one "would not naturally describe a person conducting a sports-gambling operation that is merely left unregulated as acting `pursuant to' state law." Brief for United States 18. But one might well say exactly that if the person previously was prohibited from engaging in the activity. ("Now that the State has legalized the sale of marijuana, Joe is able to sell the drug pursuant to state law.")

[28] The United States also claims to find support for its interpretation in the fact that the authorization ban applies to all "governmental entities." It is implausible, the United States submits, to think that Congress "commanded every county, district, *1475 and municipality in the Nation to prohibit sports betting." Ibid. But in making this argument, the United States again ignores the legal landscape at the time of PASPA's enactment. At that time, sports gambling was generally prohibited by state law, and therefore a State's political subdivisions were powerless to legalize the activity. But what if a State enacted a law enabling, but not requiring, one or more of its subdivisions to decide whether to authorize sports gambling? Such a state law would not itself authorize sports gambling. The ban on legalization at the local level addresses this problem.

[29] The interpretation adopted by the Third Circuit and advocated by respondents and the United States not only ignores the situation that Congress faced when it enacted PASPA but also leads to results that Congress is most unlikely to have wanted. This is illustrated by the implausible conclusions that all of those favoring alternative interpretations have been forced to reach about the extent to which the provision permits the repeal of laws banning sports gambling.

[30] The Third Circuit could not say which, if any, partial repeals are allowed. 832 F.3d, at 402. Respondents and the United States tell us that the PASPA ban on state authorization allows complete repeals, but beyond that they identify no clear line. It is improbable that Congress meant to enact such a nebulous regime.

C
[31] The respondents and United States argue that even if there is some doubt about the correctness of their interpretation of the anti-authorization provision, that interpretation should be adopted in order to avoid any anti-commandeering problem that would arise if the provision were construed to require States to maintain their laws prohibiting sports gambling. Brief for Respondents 38; Brief for United States 19. They invoke the canon of interpretation that a statute should not be held to be unconstitutional if there is any reasonable interpretation that can save it. See Jennings v. Rodriguez, 583 U.S. ___, ___, 138 S.Ct. 830, 841-842, 200 L.Ed.2d 122 (2018). The plausibility of the alternative interpretations is debatable, but even if the law could be interpreted as respondents and the United States suggest, it would still violate the anticommandeering principle, as we now explain.

III
A
[32] The anticommandeering doctrine may sound arcane, but it is simply the expression of a fundamental structural decision incorporated into the Constitution, i.e., the decision to withhold from Congress the power to issue orders directly to the States. When the original States declared their independence, they claimed the powers inherent in sovereignty — in the words of the Declaration of Independence, the authority "to do all ... Acts and Things which Independent States may of right do." ¶ 32. The Constitution limited but did not abolish the sovereign powers of the States, which retained "a residuary and inviolable sovereignty." The Federalist No. 39, p. 245 (C. Rossiter ed. 1961). Thus, both the Federal Government and the States wield sovereign powers, and that is why our system of government is said to be one of "dual sovereignty." Gregory v. Ashcroft, 501 U.S. 452, 457, 111 S.Ct. 2395, 115 L.Ed.2d 410 (1991).

[33] The Constitution limits state sovereignty in several ways. It directly prohibits the States from exercising some attributes of sovereignty. See, e.g., Art. I, § 10. Some grants of power to the Federal Government have been held to impose implicit *1476 restrictions on the States. See, e.g., Department of Revenue of Ky. v. Davis, 553 U.S. 328, 128 S.Ct. 1801, 170 L.Ed.2d 685 (2008); American Ins. Assn. v. Garamendi, 539 U.S. 396, 123 S.Ct. 2374, 156 L.Ed.2d 376 (2003). And the Constitution indirectly restricts the States by granting certain legislative powers to Congress, see Art. I, § 8, while providing in the Supremacy Clause that federal law is the "supreme Law of the Land ... any Thing in the Constitution or Laws of any State to the Contrary notwithstanding," Art. VI, cl. 2. This means that when federal and state law conflict, federal law prevails and state law is preempted.

[34] The legislative powers granted to Congress are sizable, but they are not unlimited. The Constitution confers on Congress not plenary legislative power but only certain enumerated powers. Therefore, all other legislative power is reserved for the States, as the Tenth Amendment confirms. And conspicuously absent from the list of powers given to Congress is the power to issue direct orders to the governments of the States. The anticommandeering doctrine simply represents the recognition of this limit on congressional authority.

[35] Although the anticommandeering principle is simple and basic, it did not emerge in our cases until relatively recently, when Congress attempted in a few isolated instances to extend its authority in unprecedented ways. The pioneering case was New York v. United States, 505 U.S. 144, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992), which concerned a federal law that required a State, under certain circumstances, either to "take title" to low-level radioactive waste or to "regulat[e] according to the instructions of Congress." Id., at 175, 112 S.Ct. 2408. In enacting this provision, Congress issued orders to either the legislative or executive branch of state government (depending on the branch authorized by state law to take the actions demanded). Either way, the Court held, the provision was unconstitutional because "the Constitution does not empower Congress to subject state governments to this type of instruction." Id., at 176, 112 S.Ct. 2408.

[36] Justice O'Connor's opinion for the Court traced this rule to the basic structure of government established under the Constitution. The Constitution, she noted, "confers upon Congress the power to regulate individuals, not States." Id., at 166, 112 S.Ct. 2408. In this respect, the Constitution represented a sharp break from the Articles of Confederation. "Under the Articles of Confederation, Congress lacked the authority in most respects to govern the people directly." Id., at 163, 112 S.Ct. 2408. Instead, Congress was limited to acting "`only upon the States.'" Id., at 162, 112 S.Ct. 2408 (quoting Lane County v. Oregon, 7 Wall. 71, 76, 19 L.Ed. 101 (1869)). Alexander Hamilton, among others, saw this as "`[t]he great and radical vice in ... the existing Confederation.'" 505 U.S., at 163, 112 S.Ct. 2408 (quoting The Federalist No. 15, at 108). The Constitutional Convention considered plans that would have preserved this basic structure, but it rejected them in favor of a plan under which "Congress would exercise its legislative authority directly over individuals rather than over States." 505 U.S., at 165, 112 S.Ct. 2408.

[37] As to what this structure means with regard to Congress's authority to control state legislatures, New York was clear and emphatic. The opinion recalled that "no Member of the Court ha[d] ever suggested" that even "a particularly strong federal interest" "would enable Congress to command a state government to enact state regulation." Id., at 178, 112 S.Ct. 2408 (emphasis in original). "We have *1477 always understood that even where Congress has the authority under the Constitution to pass laws requiring or prohibiting certain acts, it lacks the power directly to compel the States to require or prohibit those acts." Id., at 166, 112 S.Ct. 2408. "Congress may not simply `commandee[r] the legislative processes of the States by directly compelling them to enact and enforce a federal regulatory program.'" Id., at 161, 112 S.Ct. 2408 (quoting Hodel v. Virginia Surface Mining & Reclamation Assn., Inc., 452 U.S. 264, 288, 101 S.Ct. 2352, 69 L.Ed.2d 1 (1981)). "Where a federal interest is sufficiently strong to cause Congress to legislate, it must do so directly; it may not conscript state governments as its agents." 505 U.S., at 178, 112 S.Ct. 2408.

[38] Five years after New York, the Court applied the same principles to a federal statute requiring state and local law enforcement officers to perform background checks and related tasks in connection with applications for handgun licenses. Printz, 521 U.S. 898, 117 S.Ct. 2365, 138 L.Ed.2d 914. Holding this provision unconstitutional, the Court put the point succinctly: "The Federal Government" may not "command the States' officers, or those of their political subdivisions, to administer or enforce a federal regulatory program." Id., at 935, 117 S.Ct. 2365. This rule applies, Printz held, not only to state officers with policymaking responsibility but also to those assigned more mundane tasks. Id., at 929-930, 117 S.Ct. 2365.

B
[39] Our opinions in New York and Printz explained why adherence to the anticommandeering principle is important. Without attempting a complete survey, we mention several reasons that are significant here.

[40] First, the rule serves as "one of the Constitution's structural protections of liberty." Printz, supra, at 921, 117 S.Ct. 2365. "The Constitution does not protect the sovereignty of States for the benefit of the States or state governments as abstract political entities." New York, supra, at 181, 112 S.Ct. 2408. "To the contrary, the Constitution divides authority between federal and state governments for the protection of individuals." Ibid. "`[A] healthy balance of power between the States and the Federal Government [reduces] the risk of tyranny and abuse from either front.'" Id., at 181-182, 112 S.Ct. 2408 (quoting Gregory, 501 U.S., at 458, 111 S.Ct. 2395).

[41] Second, the anticommandeering rule promotes political accountability. When Congress itself regulates, the responsibility for the benefits and burdens of the regulation is apparent. Voters who like or dislike the effects of the regulation know who to credit or blame. By contrast, if a State imposes regulations only because it has been commanded to do so by Congress, responsibility is blurred. See New York, supra, at 168-169, 112 S.Ct. 2408; Printz, supra, at 929-930, 117 S.Ct. 2365.
Third, the anticommandeering principle prevents Congress from shifting the costs of regulation to the States. If Congress enacts a law and requires enforcement by the Executive Branch, it must appropriate the funds needed to administer the program. It is pressured to weigh the expected benefits of the program against its costs. But if Congress can compel the States to enact and enforce its program, Congress need not engage in any such analysis. See, e.g., E. Young, Two Cheers for Process Federalism, 46 Vill. L. Rev. 1349, 1360-1361 (2001).*1478

IV
A
[42] The PASPA provision at issue here — prohibiting state authorization of sports gambling — violates the anticommandeering rule. That provision unequivocally dictates what a state legislature may and may not do. And this is true under either our interpretation or that advocated by respondents and the United States. In either event, state legislatures are put under the direct control of Congress. It is as if federal officers were installed in state legislative chambers and were armed with the authority to stop legislators from voting on any offending proposals. A more direct affront to state sovereignty is not easy to imagine.

[43] Neither respondents nor the United States contends that Congress can compel a State to enact legislation, but they say that prohibiting a State from enacting new laws is another matter. See Brief for Respondents 19; Brief for United States 12. Noting that the laws challenged in New York and Printz "told states what they must do instead of what they must not do," respondents contend that commandeering occurs "only when Congress goes beyond precluding state action and affirmatively commands it." Brief for Respondents 19 (emphasis deleted).

[44] This distinction is empty. It was a matter of happenstance that the laws challenged in New York and Printz commanded "affirmative" action as opposed to imposing a prohibition. The basic principle — that Congress cannot issue direct orders to state legislatures — applies in either event.

[45] Here is an illustration. PASPA includes an exemption for States that permitted sports betting at the time of enactment, § 3704, but suppose Congress did not adopt such an exemption. Suppose Congress ordered States with legalized sports betting to take the affirmative step of criminalizing that activity and ordered the remaining States to retain their laws prohibiting sports betting. There is no good reason why the former would intrude more deeply on state sovereignty than the latter.

B
[46] Respondents and the United States claim that prior decisions of this Court show that PASPA's anti-authorization provision is constitutional, but they misread those cases. In none of them did we uphold the constitutionality of a federal statute that commanded state legislatures to enact or refrain from enacting state law.

[47] In South Carolina v. Baker, 485 U.S. 505, 108 S.Ct. 1355, 99 L.Ed.2d 592 (1988), the federal law simply altered the federal tax treatment of private investments. Specifically, it removed the federal tax exemption for interest earned on state and local bonds unless they were issued in registered rather than bearer form. This law did not order the States to enact or maintain any existing laws. Rather, it simply had the indirect effect of pressuring States to increase the rate paid on their bearer bonds in order to make them competitive with other bonds paying taxable interest.

[48] In any event, even if we assume that removal of the tax exemption was tantamount to an outright prohibition of the issuance of bearer bonds, see id., at 511, 108 S.Ct. 1355, the law would simply treat state bonds the same as private bonds. The anticommandeering doctrine does not apply when Congress evenhandedly regulates an activity in which both States and private actors engage.

[49] That principle formed the basis for the Court's decision in Reno v. Condon, 528 U.S. 141, 120 S.Ct. 666, 145 L.Ed.2d 587 *1479 (2000), which concerned a federal law restricting the disclosure and dissemination of personal information provided in applications for driver's licenses. The law applied equally to state and private actors. It did not regulate the States' sovereign authority to "regulate their own citizens." Id., at 151, 120 S.Ct. 666.

[50] In Hodel, 452 U.S., at 289, 101 S.Ct. 2352, the federal law, which involved what has been called "cooperative federalism," by no means commandeered the state legislative process. Congress enacted a statute that comprehensively regulated surface coal mining and offered States the choice of "either implement[ing]" the federal program "or else yield[ing] to a federally administered regulatory program." Ibid. Thus, the federal law allowed but did not require the States to implement a federal program. "States [were] not compelled to enforce the [federal] standards, to expend any state funds, or to participate in the federal regulatory program in any manner whatsoever." Id., at 288, 101 S.Ct. 2352. If a State did not "wish" to bear the burden of regulation, the "full regulatory burden [would] be borne by the Federal Government." Ibid.

[51] Finally, in FERC v. Mississippi, 456 U.S. 742, 102 S.Ct. 2126, 72 L.Ed.2d 532 (1982), the federal law in question issued no command to a state legislature. Enacted to restrain the consumption of oil and natural gas, the federal law directed state utility regulatory commissions to consider, but not necessarily to adopt, federal "`rate design' and regulatory standards." Id., at 746, 102 S.Ct. 2126. The Court held that this modest requirement did not infringe the States' sovereign powers, but the Court warned that it had "never ... sanctioned explicitly a federal command to the States to promulgate and enforce laws and regulations." Id., at 761-762, 102 S.Ct. 2126. FERC was decided well before our decisions in New York and Printz, and PASPA, unlike the law in FERC, does far more than require States to consider Congress's preference that the legalization of sports gambling be halted. See Printz, 521 U.S., at 929, 117 S.Ct. 2365 (distinguishing FERC).

[52] In sum, none of the prior decisions on which respondents and the United States rely involved federal laws that commandeered the state legislative process. None concerned laws that directed the States either to enact or to refrain from enacting a regulation of the conduct of activities occurring within their borders. Therefore, none of these precedents supports the constitutionality of the PASPA provision at issue here.

V
[53] Respondents and the United States defend the anti-authorization prohibition on the ground that it constitutes a valid preemption provision, but it is no such thing. Preemption is based on the Supremacy Clause, and that Clause is not an independent grant of legislative power to Congress. Instead, it simply provides "a rule of decision." Armstrong v. Exceptional Child Center, Inc., 575 U.S. ___, ___, 135 S.Ct. 1378, 1383, 191 L.Ed.2d 471 (2015). It specifies that federal law is supreme in case of a conflict with state law. Therefore, in order for the PASPA provision to preempt state law, it must satisfy two requirements. First, it must represent the exercise of a power conferred on Congress by the Constitution; pointing to the Supremacy Clause will not do. Second, since the Constitution "confers upon Congress the power to regulate individuals, not States," New York, 505 U.S., at 166, 112 S.Ct. 2408, the PASPA provision at issue must be best read as one that regulates private actors. *1480 Our cases have identified three different types of preemption — "conflict," "express," and "field," see English v. General Elec. Co., 496 U.S. 72, 78-79, 110 S.Ct. 2270, 110 L.Ed.2d 65 (1990) — but all of them work in the same way: Congress enacts a law that imposes restrictions or confers rights on private actors; a state law confers rights or imposes restrictions that conflict with the federal law; and therefore the federal law takes precedence and the state law is preempted.

[54] This mechanism is shown most clearly in cases involving "conflict preemption." A recent example is Mutual Pharmaceutical Co. v. Bartlett, 570 U.S. 472, 133 S.Ct. 2466, 186 L.Ed.2d 607 (2013). In that case, a federal law enacted under the Commerce Clause regulated manufacturers of generic drugs, prohibiting them from altering either the composition or labeling approved by the Food and Drug Administration. A State's tort law, however, effectively required a manufacturer to supplement the warnings included in the FDA-approved label. Id., at 480-486, 133 S.Ct. 2466. We held that the state law was preempted because it imposed a duty that was inconsistent — i.e., in conflict — with federal law. Id., at 493, 133 S.Ct. 2466.

[55] "Express preemption" operates in essentially the same way, but this is often obscured by the language used by Congress in framing preemption provisions. The provision at issue in Morales v. Trans World Airlines, Inc., 504 U.S. 374, 112 S.Ct. 2031, 119 L.Ed.2d 157 (1992), is illustrative. The Airline Deregulation Act of 1978 lifted prior federal regulations of airlines, and "[t]o ensure that the States would not undo federal deregulation with regulation of their own," id., at 378, 112 S.Ct. 2031, the Act provided that "no State or political subdivision thereof ... shall enact or enforce any law, rule, regulation, standard, or other provision having the force and effect of law relating to rates, routes, or services of any [covered] air carrier." 49 U.S.C. App. § 1305(a)(1) (1988 ed.).

[56] This language might appear to operate directly on the States, but it is a mistake to be confused by the way in which a preemption provision is phrased. As we recently explained, "we do not require Congress to employ a particular linguistic formulation when preempting state law." Coventry Health Care of Mo., Inc. v. Nevils, 581 U.S. ___, ___-___, 137 S.Ct. 1190, 1199, 197 L.Ed.2d 572 (2017). And if we look beyond the phrasing employed in the Airline Deregulation Act's preemption provision, it is clear that this provision operates just like any other federal law with preemptive effect. It confers on private entities (i.e., covered carriers) a federal right to engage in certain conduct subject only to certain (federal) constraints.

[57] "Field preemption" operates in the same way. Field preemption occurs when federal law occupies a "field" of regulation "so comprehensively that it has left no room for supplementary state legislation." R.J. Reynolds Tobacco Co. v. Durham County, 479 U.S. 130, 140, 107 S.Ct. 499, 93 L.Ed.2d 449 (1986). In describing field preemption, we have sometimes used the same sort of shorthand employed by Congress in express preemption provisions. See, e.g., Oneok, Inc. v. Learjet, Inc., 575 U.S. ___, ___, 135 S.Ct. 1591, 1595, 191 L.Ed.2d 511 (2015) ("Congress has forbidden the State to take action in the field that the federal statute pre-empts"). But in substance, field preemption does not involve congressional commands to the States. Instead, like all other forms of preemption, it concerns a clash between a constitutional exercise of Congress's legislative power and conflicting state law. See Crosby v. National *1481 Foreign Trade Council, 530 U.S. 363, 372, n. 6, 120 S.Ct. 2288, 147 L.Ed.2d 352 (2000).

[58] The Court's decision in Arizona v. United States, 567 U.S. 387, 132 S.Ct. 2492, 183 L.Ed.2d 351 (2012), shows how this works.

[59]  Noting that federal statutes "provide a full set of standards governing alien registration," we concluded that these laws "reflect[ ] a congressional decision to foreclose any state regulation in the area, even if it is parallel to federal standards." Id., at 401, 132 S.Ct. 2492. What this means is that the federal registration provisions not only impose federal registration obligations on aliens but also confer a federal right to be free from any other registration requirements.

[60] In sum, regardless of the language sometimes used by Congress and this Court, every form of preemption is based on a federal law that regulates the conduct of private actors, not the States.

[61] Once this is understood, it is clear that the PASPA provision prohibiting state authorization of sports gambling is not a preemption provision because there is no way in which this provision can be understood as a regulation of private actors. It certainly does not confer any federal rights on private actors interested in conducting sports gambling operations. (It does not give them a federal right to engage in sports gambling.) Nor does it impose any federal restrictions on private actors. If a private citizen or company started a sports gambling operation, either with or without state authorization, § 3702(1) would not be violated and would not provide any ground for a civil action by the Attorney General or any other party. Thus, there is simply no way to understand the provision prohibiting state authorization as anything other than a direct command to the States. And that is exactly what the anticommandeering rule does not allow.

[62] In so holding, we recognize that a closely related provision of PASPA, § 3702(2), does restrict private conduct, but that is not the provision challenged by petitioners. In Part VI-B-2, infra, we consider whether § 3702(2) is severable from the provision directly at issue in these cases.

VI
[63] Having concluded that § 3702(1) violates the anticommandeering doctrine, we consider two additional questions: first, whether the decision below should be affirmed on an alternative ground and, second, whether our decision regarding the anti-authorization provision dooms the remainder of PASPA.

A
[64] Respondents and the United States argue that, even if we disagree with the Third Circuit's decision regarding the constitutionality of the anti-authorization provision, we should nevertheless affirm based on PASPA's prohibition of state "licens[ing]" of sports gambling. Brief for Respondents 43, n. 10; Brief for United States 34-35. Although New Jersey's 2014 Act does not expressly provide for the licensing of sports gambling operations, respondents and the United States contend that the law effectively achieves that result because the only entities that it authorizes to engage in that activity, i.e., casinos and racetracks, are already required to be licensed. Ibid.

[65] We need not decide whether the 2014 Act violates PASPA's prohibition of state "licens[ing]" because that provision suffers from the same defect as the prohibition of state authorization. It issues a *1482 direct order to the state legislature.[29] Just as Congress lacks the power to order a state legislature not to enact a law authorizing sports gambling, it may not order a state legislature to refrain from enacting a law licensing sports gambling.[30]

B
[66] We therefore turn to the question whether, as petitioners maintain, our decision regarding PASPA's prohibition of the authorization and licensing of sports gambling operations dooms the remainder of the Act. In order for other PASPA provisions to fall, it must be "evident that [Congress] would not have enacted those provisions which are within its power, independently of [those] which [are] not." Alaska Airlines, Inc. v. Brock, 480 U.S. 678, 684, 107 S.Ct. 1476, 94 L.Ed.2d 661 (1987) (internal quotation marks omitted). In conducting that inquiry, we ask whether the law remains "fully operative" without the invalid provisions, Free Enterprise Fund v. Public Company Accounting Oversight Bd., 561 U.S. 477, 509, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (internal quotation marks omitted), but "we cannot rewrite a statute and give it an effect altogether different from that sought by the measure viewed as a whole," Railroad Retirement Bd. v. Alton R. Co., 295 U.S. 330, 362, 55 S.Ct. 758, 79 L.Ed. 1468 (1935). We will consider each of the provisions at issue separately.

1
[67] Under 28 U.S.C. § 3702(1), States are prohibited from "operat[ing]," "sponsor[ing]," or "promot[ing]" sports gambling schemes. If the provisions prohibiting state authorization and licensing are stricken but the prohibition on state "operat[ion]" is left standing, the result would be a scheme sharply different from what Congress contemplated when PASPA was enacted. At that time, Congress knew that New Jersey was considering the legalization of sports gambling in the privately owned Atlantic City casinos and that other States were thinking about the institution of state-run sports lotteries. PASPA addressed both of these potential developments. It gave New Jersey one year to legalize sports gambling in Atlantic City but otherwise banned the authorization of sports gambling in casinos, and it likewise prohibited the spread of state-run lotteries. If Congress had known that States would be free to authorize sports gambling in privately owned casinos, would it have nevertheless wanted to prevent States from running sports lotteries?

[68] That seems most unlikely. State-run lotteries, which sold tickets costing only a few dollars, were thought more benign than other forms of gambling, and that is why they had been adopted in many States. Casino gambling, on the other hand, was generally regarded as far more dangerous. A gambler at a casino can easily incur heavy losses, and the legalization of privately owned casinos was known to create the threat of infiltration by organized crime, as Nevada's early experience *1483 had notoriously shown.[31] To the Congress that adopted PASPA, legalizing sports gambling in privately owned casinos while prohibiting state-run sports lotteries would have seemed exactly backwards.

[69] Prohibiting the States from engaging in commercial activities that are permitted for private parties would also have been unusual, and it is unclear what might justify such disparate treatment. Respondents suggest that Congress wanted to prevent States from taking steps that the public might interpret as the endorsement of sports gambling, Brief for Respondents 39, but we have never held that the Constitution permits the Federal Government to prevent a state legislature from expressing its views on subjects of public importance. For these reasons, we do not think that the provision barring state operation of sports gambling can be severed.

[70] We reach the same conclusion with respect to the provisions prohibiting state "sponsor[ship]" and "promot[ion]." The line between authorization, licensing, and operation, on the one hand, and sponsorship or promotion, on the other, is too uncertain. It is unlikely that Congress would have wanted to prohibit such an ill-defined category of state conduct.

2
[71] Nor do we think that Congress would have wanted to sever the PASPA provisions that prohibit a private actor from "sponsor[ing]," "operat[ing]," or "promot[ing]" sports gambling schemes "pursuant to" state law. § 3702(2). These provisions were obviously meant to work together with the provisions in § 3702(1) that impose similar restrictions on governmental entities. If Congress had known that the latter provisions would fall, we do not think it would have wanted the former to stand alone.

[72] The present cases illustrate exactly how Congress must have intended § 3702(1) and § 3702(2) to work. If a State attempted to authorize particular private entities to engage in sports gambling, the State could be sued under § 3702(1), and the private entity could be sued at the same time under § 3702(2). The two sets of provisions were meant to be deployed in tandem to stop what PASPA aimed to prevent: state legalization of sports gambling. But if, as we now hold, Congress lacks the authority to prohibit a State from legalizing sports gambling, the prohibition of private conduct under § 3702(2) ceases to implement any coherent federal policy.

[73] Under § 3702(2), private conduct violates federal law only if it is permitted by state law. That strange rule is exactly the opposite of the general federal approach to gambling. Under 18 U.S.C. § 1955, operating a gambling business violates federal law only if that conduct is illegal under state or local law. Similarly, 18 U.S.C. § 1953, which criminalizes the interstate transmission of wagering paraphernalia, and 18 U.S.C. § 1084, which outlaws the interstate transmission of information that assists in the placing of a bet on a sporting event, apply only if the underlying gambling is illegal under state law. See also 18 U.S.C. § 1952 (making it illegal to travel in interstate commerce to further a gambling business that is illegal under applicable state law).

[74] These provisions implement a coherent federal policy: They respect the policy choices of the people of each State on the controversial issue of gambling. By contrast, if § 3702(2) is severed from § 3702(1), it implements a perverse policy that undermines whatever policy is favored by the people of a State. If the people of a State support the legalization of sports *1484 gambling, federal law would make the activity illegal. But if a State outlaws sports gambling, that activity would be lawful under § 3702(2). We do not think that Congress ever contemplated that such a weird result would come to pass.

[75] PASPA's enforcement scheme reinforces this conclusion. PASPA authorizes civil suits by the Attorney General and sports organizations but does not make sports gambling a federal crime or provide civil penalties for violations. This enforcement scheme is suited for challenging state authorization or licensing or a small number of private operations, but the scheme would break down if a State broadly decriminalized sports gambling. It is revealing that the Congressional Budget Office estimated that PASPA would impose "no cost" on the Federal Government, see S. Rep. No. 102-248, p. 10 (1991), a conclusion that would certainly be incorrect if enforcement required a multiplicity of civil suits and applications to hold illegal bookies and other private parties in contempt.[32]

3
[76] The remaining question that we must decide is whether the provisions of PASPA prohibiting the "advertis[ing]" of sports gambling are severable. See §§ 3702(1)-(2). If these provisions were allowed to stand, federal law would forbid the advertising of an activity that is legal under both federal and state law, and that is something that Congress has rarely done. For example, the advertising of cigarettes is heavily regulated but not totally banned. See Federal Cigarette Labeling and Advertising Act, 79 Stat. 282; Family Smoking Prevention and Tobacco Control Act, §§ 201-204, 123 Stat. 1842-1848.

[77] It is true that at one time federal law prohibited the use of the mail or interstate commerce to distribute advertisements of lotteries that were permitted under state law, but that is no longer the case. See United States v. Edge Broadcasting Co., 509 U.S. 418, 421-423, 113 S.Ct. 2696, 125 L.Ed.2d 345 (1993). In 1975, Congress passed a new statute, codified at 18 U.S.C. § 1307, that explicitly exempts print advertisements regarding a lottery lawfully conducted by States, and in Greater New Orleans Broadcasting Assn., Inc. v. United States, 527 U.S. 173, 176, 119 S.Ct. 1923, 144 L.Ed.2d 161 (1999), we held that the First Amendment protects the right of a radio or television station in a State with a lottery to run such advertisements. In light of these developments, we do not think that Congress would want the advertising provisions to stand if the remainder of PASPA must fall.

[78] For these reasons, we hold that no provision of PASPA is severable from the provision directly at issue in these cases.

* * *

[79] The legalization of sports gambling is a controversial subject. Supporters argue that legalization will produce revenue for the States and critically weaken illegal sports betting operations, which are often run by organized crime. Opponents contend that legalizing sports gambling will hook the young on gambling, encourage people of modest means to squander their savings and earnings, and corrupt professional and college sports.

[80] The legalization of sports gambling requires an important policy choice, but the choice is not ours to make. Congress can regulate sports gambling directly, but if it elects not to do so, each State is free to act *1485 on its own. Our job is to interpret the law Congress has enacted and decide whether it is consistent with the Constitution. PASPA is not. PASPA "regulate[s] state governments' regulation" of their citizens, New York, 505 U.S., at 166, 112 S.Ct. 2408. The Constitution gives Congress no such power.

[81] The judgment of the Third Circuit is reversed.

[82] It is so ordered.

Justice THOMAS, concurring.
[1] I join the Court's opinion in its entirety. I write separately, however, to express my growing discomfort with our modern severability precedents.

[2] I agree with the Court that the Professional and Amateur Sports Protection Act (PASPA) exceeds Congress' Article I authority to the extent it prohibits New Jersey from "authoriz[ing]" or "licens[ing]" sports gambling, 28 U.S.C. § 3702(1). Unlike the dissent, I do "doubt" that Congress can prohibit sports gambling that does not cross state lines. Post, at p. 1489 (opinion of GINSBURG, J.); see License Tax Cases, 5 Wall. 462, 470-471, 18 L.Ed. 497 (1867) (holding that Congress has "no power" to regulate "the internal commerce or domestic trade of the States," including the intrastate sale of lottery tickets); United States v. Lopez, 514 U.S. 549, 587-601, 115 S.Ct. 1624, 131 L.Ed.2d 626 (1995) (THOMAS, J., concurring) (documenting why the Commerce Clause does not permit Congress to regulate purely local activities that have a substantial effect on interstate commerce). But even assuming the Commerce Clause allows Congress to prohibit intrastate sports gambling "directly," it "does not authorize Congress to regulate state governments' regulation of interstate commerce." New York v. United States, 505 U.S. 144, 166, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992). The Necessary and Proper Clause does not give Congress this power either, as a law is not "proper" if it "subvert[s] basic principles of federalism and dual sovereignty." Gonzales v. Raich, 545 U.S. 1, 65, 125 S.Ct. 2195, 162 L.Ed.2d 1 (2005) (THOMAS, J., dissenting). Commandeering the States, as PASPA does, subverts those principles. See Printz v. United States, 521 U.S. 898, 923-924, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997).

[3] Because PASPA is at least partially unconstitutional, our precedents instruct us to determine "which portions of the ... statute we must sever and excise." United States v. Booker, 543 U.S. 220, 258, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005) (emphasis deleted). The Court must make this severability determination by asking a counterfactual question: "`Would Congress still have passed' the valid sections `had it known' about the constitutional invalidity of the other portions of the statute?" Id., at 246, 125 S.Ct. 738 (quoting Denver Area Ed. Telecommunications Consortium, Inc. v. FCC, 518 U.S. 727, 767, 116 S.Ct. 2374, 135 L.Ed.2d 888 (1996) (plurality opinion)). I join the Court's opinion because it gives the best answer it can to this question, and no party has asked us to apply a different test. But in a future case, we should take another look at our severability precedents.

[4] Those precedents appear to be in tension with traditional limits on judicial authority. Early American courts did not have a severability doctrine. See Walsh, Partial Unconstitutionality, 85 N.Y.U. L. Rev. 738, 769 (2010) (Walsh). They recognized that the judicial power is, fundamentally, the power to render judgments in individual cases. See id., at 755; Baude, The Judgment Power, 96 Geo. L.J. 1807, 1815 (2008). Judicial review was a byproduct of that process. See generally P. Hamburger, Law and Judicial Duty (2008); Prakash & Yoo, The Origins of Judicial *1486 Review, 70 U. Chi. L. Rev. 887 (2003). As Chief Justice Marshall famously explained, "[i]t is emphatically the province and duty of the judicial department to say what the law is" because "[t]hose who apply the rule to particular cases, must of necessity expound and interpret that rule." Marbury v. Madison, 1 Cranch 137, 177, 2 L.Ed. 60 (1803). If a plaintiff relies on a statute but a defendant argues that the statute conflicts with the Constitution, then courts must resolve that dispute and, if they agree with the defendant, follow the higher law of the Constitution. See id., at 177-178; The Federalist No. 78, p. 467 (C. Rossiter ed. 1961) (A. Hamilton). Thus, when early American courts determined that a statute was unconstitutional, they would simply decline to enforce it in the case before them. See Walsh 755-766. "[T]here was no `next step' in which courts inquired into whether the legislature would have preferred no law at all to the constitutional remainder." Id., at 777.

[5] Despite this historical practice, the Court's modern cases treat the severability doctrine as a "remedy" for constitutional violations and ask which provisions of the statute must be "excised." See, e.g., Ayotte v. Planned Parenthood of Northern New Eng., 546 U.S. 320, 329, 126 S.Ct. 961, 163 L.Ed.2d 812 (2006); Booker, supra, at 245, 125 S.Ct. 738; Alaska Airlines, Inc. v. Brock, 480 U.S. 678, 686, 107 S.Ct. 1476, 94 L.Ed.2d 661 (1987). This language cannot be taken literally. Invalidating a statute is not a "remedy," like an injunction, a declaration, or damages. See Harrison, Severability, Remedies, and Constitutional Adjudication, 83 Geo. Wash. L. Rev. 56, 82-88 (2014) (Harrison). Remedies "operate with respect to specific parties," not "on legal rules in the abstract." Id., at 85; see also Massachusetts v. Mellon, 262 U.S. 447, 488, 43 S.Ct. 597, 67 L.Ed. 1078 (1923) (explaining that the power "to review and annul acts of Congress" is "little more than the negative power to disregard an unconstitutional enactment" and that "the court enjoins ... not the execution of the statute, but the acts of the official"). And courts do not have the power to "excise" or "strike down" statutes. See 39 Op. Atty. Gen. 22, 22-23 (1937) ("The decisions are practically in accord in holding that the courts have no power to repeal or abolish a statute"); Harrison 82 ("[C]ourts do not make [nonseverable] provisions inoperative.... Invalidation by courts is a figure of speech"); Mitchell, The Writ-of-Erasure Fallacy, 104 Va. L. Rev. (forthcoming 2018) (manuscript, at 4) ("The federal courts have no authority to erase a duly enacted law from the statute books"), online at https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3158038 (as last visited May 11, 2018).

[6] Because courts cannot take a blue pencil to statutes, the severability doctrine must be an exercise in statutory interpretation. In other words, the severability doctrine has courts decide how a statute operates once they conclude that part of it cannot be constitutionally enforced. See Fallon, As-Applied and Facial Challenges and Third-Party Standing, 113 Harv. L. Rev. 1321, 1333-1334 (2000); Harrison 88. But even under this view, the severability doctrine is still dubious for at least two reasons.

[7] First, the severability doctrine does not follow basic principles of statutory interpretation. Instead of requiring courts to determine what a statute means, the severability doctrine requires courts to make "a nebulous inquiry into hypothetical congressional intent." Booker, supra, at 320, n. 7, 125 S.Ct. 738 (THOMAS, J., dissenting in part). It requires judges to determine what Congress would have intended had it known that part of its statute was *1487 unconstitutional.[*] But it seems unlikely that the enacting Congress had any intent on this question; Congress typically does not pass statutes with the expectation that some part will later be deemed unconstitutional. See Walsh 740-741; Stern, Separability and Separability Clauses in the Supreme Court, 51 Harv. L. Rev. 76, 98 (1937) (Stern). Without any actual evidence of intent, the severability doctrine invites courts to rely on their own views about what the best statute would be. See Walsh 752-753; Stern 112-113. More fundamentally, even if courts could discern Congress' hypothetical intentions, intentions do not count unless they are enshrined in a text that makes it through the constitutional processes of bicameralism and presentment. See Wyeth v. Levine, 555 U.S. 555, 586-588, 129 S.Ct. 1187, 173 L.Ed.2d 51 (2009) (THOMAS, J., concurring in judgment). Because we have "`a Government of laws, not of men,'" we are governed by "legislated text," not "legislators' intentions" — and especially not legislators' hypothetical intentions. Zuni Public School Dist. No. 89 v. Department of Education, 550 U.S. 81, 119, 127 S.Ct. 1534, 167 L.Ed.2d 449 (2007) (Scalia, J., dissenting). Yet hypothetical intent is exactly what the severability doctrine turns on, at least when Congress has not expressed its fallback position in the text.

[8] Second, the severability doctrine often requires courts to weigh in on statutory provisions that no party has standing to challenge, bringing courts dangerously close to issuing advisory opinions. See Stern 77; Lea, Situational Severability, 103 Va. L. Rev. 735, 788-803 (2017) (Lea). If one provision of a statute is deemed unconstitutional, the severability doctrine places every other provision at risk of being declared nonseverable and thus inoperative; our precedents do not ask whether the plaintiff has standing to challenge those other provisions. See National Federation of Independent Business v. Sebelius, 567 U.S. 519, 696-697, 132 S.Ct. 2566, 183 L.Ed.2d 450 (2012) (joint dissent) (citing, as an example, Williams v. Standard Oil Co. of La., 278 U.S. 235, 242-244, 49 S.Ct. 115, 73 L.Ed. 287 (1929)). True, the plaintiff had standing to challenge the unconstitutional part of the statute. But the severability doctrine comes into play only after the court has resolved that issue — typically the only live controversy between the parties. In every other context, a plaintiff must demonstrate standing for each part of the statute that he wants to challenge. See Lea 789, 751, and nn. 79-80 (citing, as examples, Davis v. Federal Election Comm'n, 554 U.S. 724, 733-734, 128 S.Ct. 2759, 171 L.Ed.2d 737 (2008); DaimlerChrysler Corp. v. Cuno, 547 U.S. 332, 346, 350-353, 126 S.Ct. 1854, 164 L.Ed.2d 589 (2006)). The severability doctrine is thus an unexplained exception to the normal rules of standing, as well as the separation-of-powers principles that those rules protect. See Steel Co. v. Citizens for Better Environment, 523 U.S. 83, 101, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998).

[9] In sum, our modern severability precedents are in tension with longstanding limits on the judicial power. And, though no party in this case has asked us to reconsider these precedents, at some point, it behooves us to do so.
*1488 Justice BREYER, concurring in part and dissenting in part.

[10] I agree with Justice GINSBURG that 28 U.S.C. § 3702(2) is severable from the challenged portion of § 3702(1). The challenged part of subsection (1) prohibits a State from "author[izing]" or "licens[ing]" sports gambling schemes; subsection (2) prohibits individuals from "sponsor[ing], operat[ing], advertis[ing], or promot[ing]" sports gambling schemes "pursuant to the law ... of a governmental entity." The first says that a State cannot authorize sports gambling schemes under state law; the second says that (just in case a State finds a way to do so) sports gambling schemes that a State authorizes are unlawful under federal law regardless. As Justice GINSBURG makes clear, the latter section can live comfortably on its own without the first.

[11] Why would Congress enact both these provisions? The obvious answer is that Congress wanted to "keep sports gambling from spreading." S. Rep. No. 102-248, pp. 4-6 (1991). It feared that widespread sports gambling would "threate[n] to change the nature of sporting events from wholesome entertainment for all ages to devices for gambling." Id., at 4. And it may have preferred that state authorities enforce state law forbidding sports gambling than require federal authorities to bring civil suits to enforce federal law forbidding about the same thing. Alternatively, Congress might have seen subsection (2) as a backup, called into play if subsection (1)'s requirements, directed to the States, turned out to be unconstitutional — which, of course, is just what has happened. Neither of these objectives is unreasonable.

[12] So read, the two subsections both forbid sports gambling but § 3702(2) applies federal policy directly to individuals while the challenged part of § 3702(1) forces the States to prohibit sports gambling schemes (thereby shifting the burden of enforcing federal regulatory policy from the Federal Government to state governments). Section 3702(2), addressed to individuals, standing alone seeks to achieve Congress' objective of halting the spread of sports gambling schemes by "regulat[ing] interstate commerce directly." New York v. United States, 505 U.S. 144, 166, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992). But the challenged part of subsection (1) seeks the same end indirectly by "regulat[ing] state governments' regulation of interstate commerce." Ibid. And it does so by addressing the States (not individuals) directly and telling state legislatures what laws they must (or cannot) enact. Under our precedent, the first provision (directly and unconditionally telling States what laws they must enact) is unconstitutional, but the second (directly telling individuals what they cannot do) is not. See ibid.

[13] As so interpreted, the statutes would make New Jersey's victory here mostly Pyrrhic. But that is because the only problem with the challenged part of § 3702(1) lies in its means, not its end. Congress has the constitutional power to prohibit sports gambling schemes, and no party here argues that there is any constitutional defect in § 3702(2)'s alternative means of doing so.

[14] I consequently join Justice GINSBURG's dissenting opinion in part, and all but Part VI-B of the Court's opinion.

Justice GINSBURG, with whom Justice SOTOMAYOR joins, and with whom Justice BREYER joins in part, dissenting.
[1] The petition for certiorari filed by the Governor of New Jersey invited the Court to consider a sole question: "Does a federal statute that prohibits modification or repeal of state-law prohibitions on private *1489 conduct impermissibly commandeer the regulatory power of States in contravention of New York v. United States, 505 U.S. 144, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992)?" Pet. for Cert. in No. 16-476, p. i.

[2] Assuming, arguendo, a "yes" answer to that question, there would be no cause to deploy a wrecking ball destroying the Professional and Amateur Sports Protection Act (PASPA) in its entirety, as the Court does today. Leaving out the alleged infirmity, i.e., "commandeering" state regulatory action by prohibiting the States from "authoriz[ing]" and "licens[ing]" sports-gambling schemes, 28 U.S.C. § 3702(1), two federal edicts should remain intact. First, PASPA bans States themselves (or their agencies) from "sponsor[ing], operat[ing], advertis[ing], [or] promot[ing]" sports-gambling schemes. Ibid. Second, PASPA stops private parties from "sponsor[ing], operat[ing], advertis[ing], or promot[ing]" sports-gambling schemes if state law authorizes them to do so. § 3702(2).[1] Nothing in these § 3702(1) and § 3702(2) prohibitions commands States to do anything other than desist from conduct federal law proscribes.[2] Nor is there any doubt that Congress has power to regulate gambling on a nationwide basis, authority Congress exercised in PASPA. See Gonzales v. Raich, 545 U.S. 1, 17, 125 S.Ct. 2195, 162 L.Ed.2d 1 (2005) ("Our case law firmly establishes Congress' power to regulate purely local activities that are part of an economic `class of activities' that have a substantial effect on interstate commerce.").

[3] Surely, the accountability concern that gave birth to the anticommandeering doctrine is not implicated in any federal proscription other than the bans on States' authorizing and licensing sports-gambling schemes. The concern triggering the doctrine arises only "where the Federal Government compels States to regulate" or to enforce federal law, thereby creating the appearance that state officials are responsible for policies Congress forced them to enact. New York v. United States, 505 U.S. 144, 168, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992). If States themselves and private parties may not operate sports-gambling schemes, responsibility for the proscriptions is hardly blurred. It cannot be maintained credibly that state officials have anything to do with the restraints. Unmistakably, the foreclosure of sportsgambling schemes, whether state run or privately operated, is chargeable to congressional, not state, legislative action.

[4] When a statute reveals a constitutional flaw, the Court ordinarily engages in a salvage rather than a demolition operation: It "limit[s] the solution [to] severing any problematic portions while leaving the remainder intact." Free Enterprise Fund v. Public Company Accounting Oversight Bd., 561 U.S. 477, 508, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010) (internal quotation marks omitted). The relevant question is whether the Legislature would have wanted unproblematic aspects of the legislation to survive or would want them to fall along with the infirmity.[3] As the Court stated in *1490 New York, "[u]nless it is evident that the Legislature would not have enacted those provisions which are within its power, ... the invalid part may be dropped if what is left is fully operative as a law." 505 U.S., at 186, 112 S.Ct. 2408 (internal quotation marks omitted). Here, it is scarcely arguable that Congress "would have preferred no statute at all," Executive Benefits Ins. Agency v. Arkison, 573 U.S. ___, ___, 134 S.Ct. 2165, 2173, 189 L.Ed.2d 83 (2014), over one that simply stops States and private parties alike from operating sports-gambling schemes.

[5] The Court wields an ax to cut down § 3702 instead of using a scalpel to trim the statute. It does so apparently in the mistaken assumption that private sports-gambling schemes would become lawful in the wake of its decision. In particular, the Court holds that the prohibition on state "operat[ion]" of sports-gambling schemes cannot survive, because it does not believe Congress would have "wanted to prevent States from running sports lotteries" "had [it] known that States would be free to authorize sports gambling in privately owned casinos." Ante, at 1482. In so reasoning, the Court shutters § 3702(2), under which private parties are prohibited from operating sports-gambling schemes precisely when state law authorizes them to do so.[4]

[6] This plain error pervasively infects the Court's severability analysis. The Court strikes Congress' ban on state "sponsor[ship]" and "promot[ion]" of sports-gambling schemes because it has (mistakenly) struck Congress' prohibition on state "operat[ion]" of such schemes. See ante, at Pp. 1482-1483. It strikes Congress' prohibitions on private "sponsor[ship]," "operat[ion]," and "promot[ion]" of sportsgambling schemes because it has (mistakenly) struck those same prohibitions on the States. See ante, at Pp. 1482-1483. And it strikes Congress' prohibition on "advertis[ing]" sports-gambling schemes because it has struck everything else. See ante, at Pp. 1483-1484.
* * *

[7] In PASPA, shorn of the prohibition on modifying or repealing state law, Congress permissibly exercised its authority to regulate commerce by instructing States and private parties to refrain from operating sports-gambling schemes. On no rational ground can it be concluded that Congress would have preferred no statute at all if it could not prohibit States from authorizing or licensing such schemes. Deleting the alleged "commandeering" directions would free the statute to accomplish just what Congress legitimately sought to achieve: stopping sports-gambling regimes while making it clear that the stoppage is attributable to federal, not state, action. I therefore dissent from the Court's determination to destroy PASPA rather than salvage the statute.

Majority Footnotes
[1] See Nat. Gambling Impact Study Comm'n, Final Report, p. 2-1 (1999) (Final Report); S. Durham & K. Hashimoto, The History of Gambling in America 34-35 (2010).
[2] See Atlantic City Racing Assn. v. Attorney General, 98 N.J. 535, 539-541, 489 A.2d 165, 167-168 (1985).
[3] See Note, The Casino Act: Gambling's Past and the Casino Act's Future, 10 Rutgers-Camden L.J. 279, 287 (1979) (The Casino Act).
[4] Id., at 288; see also N.J. Const. Art. 4, § 7, ¶ 2(A); Bingo Licensing Law, N.J. Stat. Ann. § 5:8-24 et seq. (West 2012).
[5] See State Lottery Law, N.J. Stat. Ann. § 5:9-1 et seq.; The Casino Act, at 288; N.J. Const. Art. 4, § 7, ¶ 2(C); Final Report, at 2-1.
[6] Id., at 2-1.
[7] T. White, The Making of the President 1964, p. 275 (1965).
[8] See D. Clary, Gangsters to Governors 152-153 (2017) (Clary).
[9] See The Casino Act, at 289.
[10] See ibid.; N.J. Const. Art. 4, § 7, ¶ 2(D).
[11] Clary 146.
[12] Id., at 146, 158.
[13] Id., at 208-210.
[14] Casinos now operate in New York, Pennsylvania, Delaware, and Maryland. See American Gaming Assn., 2016 State of the States, p. 8, online at https://www.americangaming.org/sites/default/files/2016%20 State%20of%20the%20States_FINAL.pdf (all Internet materials as last visited May 4, 2018).
[15] See, e.g., Brief for American Gaming Assn. as Amicus Curiae 1-2.
[16] See, e.g., Final Report, at 3-10; B. Bradley, The Professional and Amateur Sports Protection Act — Policy Concerns Behind Senate Bill 474, 2 Seton Hall J. Sport L. 5, 7 (1992); Brief for Stop Predatory Gambling et al. as Amici Curiae 22-23.
[17] For example, in 1919, professional gamblers are said to have paid members of the Chicago White Sox to throw the World Series, an episode that was thought to have threatened baseball's status as the Nation's pastime. See E. Asinof, Eight Men Out: The Black Sox and the 1919 World Series 5, 198-199 (1963). And in the early 1950s, the Nation was shocked when several college basketball players were convicted for shaving points. S. Cohen, The Game They Played 183-238 (1977). This scandal is said to have nearly killed college basketball. See generally C. Rosen, Scandals of '51: How the Gamblers Almost Killed College Basketball (1978).
[18] See Professional and Amateur Sports Protection, S. Rep. No. 102-248, p. 8 (1991); Hearing before the Subcommittee on Patents, Copyrights and Trademarks of the Senate Committee on the Judiciary, 102d Cong., 1st Sess., 21, 39, 46-47, 59-60, 227 (1991) (S. Hrg. 102-499) (statements by representatives of major sports leagues opposing sports gambling).
[19] S. Rep. No. 102-248, at 5.
[20] S. Hrg. 102-499, at 10-14.
[21] App. to Pet. for Cert. in No. 16-476, p. 225a.
[22] The statute applies to any "governmental entity," which is defined as "a State, a political subdivision of a State, or an entity or organization ... that has governmental authority within the territorial boundaries of the United States." 28 U.S.C. § 3701(2).
[23] PASPA does not define the term "scheme." The United States has not offered a definition of the term but suggests that it encompasses only those forms of gambling having some unspecified degree of organization or structure. See Brief for United States as Amicus Curiae 28-29. For convenience, we will use the term "sports gambling" to refer to whatever forms of sports gambling fall within PASPA's reach.
[24] The Congressional Budget Office estimated that PASPA would not require the appropriation of any federal funds. S. Rep. No. 102-248, at 10.
[25] Ibid.
[26] Ibid.; 138 Cong. Rec. 12973.
[27] Although this provision did not specifically mention New Jersey or Atlantic City, its requirements — permitting legalization only "in a municipality" with an uninterrupted 10-year history of legal casino gaming — did not fit anyplace else.
[28] See, e.g., A. McCullum, Vermont's legal recreational marijuana law: What you should know, USA Today Network (Jan. 23, 2018), online at https://www.usatoday.com/story/news/nation-now/2018/01/23/vermont-legal-marijuana-law-what-know/XXXXXXXXXX/ ("Vermont ... bec[ame] the first [State] in the country to authorize the recreational use of [marijuana] by an act of a state legislature." (emphasis added)).
[29] Even if the prohibition of state licensing were not itself unconstitutional, we do not think it could be severed from the invalid provision forbidding state authorization. The provision of PASPA giving New Jersey the option of legalizing sports gambling within one year of enactment applied only to casinos operated "pursuant to a comprehensive system of State regulation." § 3704(a)(3)(B). This shows that Congress preferred tightly regulated sports gambling over total deregulation.
[30] The dissent apparently disagrees with our holding that the provisions forbidding state authorization and licensing violate the anti-commandeering principle, but it provides no explanation for its position.
[31] See Clary 84-102.
[32] Of course, one need not rely on the Senate Report for the commonsense proposition that leaving § 3702(2) in place could wildly change the fiscal calculus, "giv[ing] it an effect altogether different from that sought by the measure viewed as a whole." Railroad Retirement Bd. v. Alton R. Co., 295 U.S. 330, 362, 55 S.Ct. 758, 79 L.Ed. 1468 (1935).
[*] The first court to engage in this counterfactual exploration of legislative intent was the Massachusetts Supreme Judicial Court in Warren v. Mayor and Aldermen of Charlestown, 68 Mass. 84, 99 (1854). This Court adopted the Warren formulation in the late 19th century, see Allen v. Louisiana, 103 U.S. 80, 84, 26 L.Ed. 318 (1881), an era when statutory interpretation privileged Congress' unexpressed "intent" over the enacted text, see, e.g., Church of Holy Trinity v. United States, 143 U.S. 457, 472, 12 S.Ct. 511, 36 L.Ed. 226 (1892); United States v. Moore, 95 U.S. 760, 763, 24 L.Ed. 588 (1878).

Dissent Footnotes
[1] PASPA was not designed to eliminate any and all sports gambling. The statute targets sports-gambling schemes, i.e., organized markets for sports gambling, whether operated by a State or by a third party under state authorization.
[2] In lieu of a flat ban, PASPA prohibits third parties from operating sports-gambling schemes only if state law permits them to do so. If a state ban is in place, of course, there is no need for a federal proscription.
[3] Notably, in the two decisions marking out and applying the anticommandeering doctrine to invalidate federal law, the Court invalidated only the offending provision, not the entire statute. New York v. United States, 505 U.S. 144, 186-187, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992); Printz v. United States, 521 U.S. 898, 935, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997).
[4] As earlier indicated, see supra, at 1489, direct federal regulation of sports-gambling schemes nationwide, including private-party schemes, falls within Congress' power to regulate activities having a substantial effect on interstate commerce. See Gonzales v. Raich, 545 U.S. 1, 17, 125 S.Ct. 2195, 162 L.Ed.2d 1 (2005). Indeed, according to the Court, direct regulation is precisely what the anticommandeering doctrine requires. Ante, at Pp. 1475-1478.

Justice O'Connor, delivered the opinion of the Court.

[1] The Age Discrimination in Employment Act of 1967, makes it unlawful for an employer, including a State, "to fail or refuse to hire or to discharge any individual or otherwise discriminate against any individual . . . because of such individual's age." In these cases, three sets of plaintiffs filed suit under the Act, seeking money damages for their state employers' alleged discrimination on the basis of age. In these cases, we are asked to consider whether the ADEA contains a clear statement of Congress' intent to abrogate the States' Eleventh Amendment immunity and, if so, whether the ADEA is a proper exercise of Congress' constitutional authority. We conclude that the ADEA does contain a clear statement of Congress' intent to abrogate the States' immunity, but that the abrogation exceeded Congress' authority under § 5 of the Fourteenth Amendment.

[2] To determine whether a federal statute properly subjects States to suits by individuals, we apply a "simple but stringent test: `Congress may abrogate the States' constitutionally secured immunity from suit in federal court only by making its intention unmistakably clear in the language of the statute.' " We agree with petitioners that the ADEA satisfies that test. The ADEA clearly provides for suits by individuals against States.  The ACt authorizes employeees to maintain actions for backpay "against any employer (including a public agency) in any Federal or State court of competent jurisdiction . . . ." Read as a whole, the plain language of these provisions clearly demonstrates Congress' intent to subject the States to suit for money damages at the hands of individual employees.

[3] In Seminole Tribe, we held that Congress lacks power under Article I to abrogate the States' sovereign immunity. Section 5 of the Fourteenth Amendment, however, does grant Congress the authority to abrogate the States' sovereign immunity. In Fitzpatrick v. Bitzer (1976), we recognized that "the Eleventh Amendment, and the principle of state sovereignty which it embodies, are necessarily limited by the enforcement provisions of § 5 of the Fourteenth Amendment."

[4] As we recognized most recently in City of Boerne v. Flores (1997), § 5 is an affirmative grant of power to Congress. "It is for Congress in the first instance to `determin[e] whether and what legislation is needed to secure the guarantees of the Fourteenth Amendment,' and its conclusions are entitled to much deference." Congress' § 5 power is not confined to the enactment of legislation that merely parrots the precise wording of the Fourteenth Amendment. Rather, Congress' power "to enforce" the Amendment includes the authority both to remedy and to deter violation of rights guaranteed thereunder by prohibiting a somewhat broader swath of conduct, including that which is not itself forbidden by the Amendment's text.

[5] Nevertheless, we have also recognized that the same language that serves as the basis for the affirmative grant of congressional power also serves to limit that power. For example, Congress cannot "decree the substance of the Fourteenth Amendment's restrictions on the States. . . . It has been given the power `to enforce,' not the power to determine what constitutes a constitutional violation." The ultimate interpretation and determination of the Fourteenth Amendment's substantive meaning remains the province of the Judicial Branch.

[6] Applying the same "congruence and proportionality" test in these cases, we conclude that the ADEA is not "appropriate legislation" under § 5 of the Fourteenth Amendment. Initially, the substantive requirements the ADEA imposes on state and local governments are disproportionate to any unconstitutional conduct that conceivably could be targeted by the Act. We have considered claims of unconstitutional age discrimination under the Equal Protection Clause three times. See Gregory v. Ashcroft (1991); Vance v. Bradley (1979); Massachusetts Bd. of Retirement v. Murgia (1976) (per curiam). In all three cases, we held that the age classifications at issue did not violate the Equal Protection Clause. Age classifications, unlike governmental conduct based on race or gender, cannot be characterized as "so seldom relevant to the achievement of any legitimate state interest that laws grounded in such considerations are deemed to reflect prejudice and antipathy." Older persons, again, unlike those who suffer discrimination on the basis of race or gender, have not been subjected to a “history of purposeful unequal treatment.” Old age also does not define a discrete and insular minority because all persons, if they live out their normal life spans, will experience it. Accordingly, as we recognized in Murgia, Bradley, and Gregory, age is not a suspect classification under the Equal Protection Clause.

[7] States may discriminate on the basis of age without offending the Fourteenth Amendment if the age classification in question is rationally related to a legitimate state interest. The rationality commanded by the Equal Protection Clause does not require States to match age distinctions and the legitimate interests they serve with razorlike precision. As we have explained, when conducting rational basis review "we will not overturn such [government action] unless the varying treatment of different groups or persons is so unrelated to the achievement of any combination of legitimate purposes that we can only conclude that the [government's] actions were irrational." In contrast, when a State discriminates on the basis of race or gender, we require a tighter fit between the discriminatory means and the legitimate ends they serve. Under the Fourteenth Amendment, a State may rely on age as a proxy for other qualities, abilities, or characteristics that are relevant to the State's legitimate interests. The Constitution does not preclude reliance on such generalizations. That age proves to be an inaccurate proxy in any individual case is irrelevant.

[8] Judged against the backdrop of our equal protection jurisprudence, it is clear that the ADEA is "so out of proportion to a supposed remedial or preventive object that it cannot be understood as responsive to, or designed to prevent, unconstitutional behavior." City of Boerne. The Act, through its broad restriction on the use of age as a discriminating factor, prohibits substantially more state employment decisions and practices than would likely be held unconstitutional under the applicable equal protection, rational basis standard.

[9] That the ADEA prohibits very little conduct likely to be held unconstitutional, while significant, does not alone provide the answer to our § 5 inquiry. Difficult and intractable problems often require powerful remedies, and we have never held that § 5 precludes Congress from enacting reasonably prophylactic legislation. Our task is to determine whether the ADEA is in fact just such an appropriate remedy or, instead, merely an attempt to substantively redefine the States' legal obligations with respect to age discrimination. One means by which we have made such a determination in the past is by examining the legislative record containing the reasons for Congress' action. "The appropriateness of remedial measures must be considered in light of the evil presented. Strong measures appropriate to address one harm may be an unwarranted response to another, lesser one."

[10] Our examination of the ADEA's legislative record confirms that Congress' 1974 extension of the Act to the States was an unwarranted response to a p

erhaps inconsequential problem. Congress never identified any pattern of age discrimination by the States, much less any discrimination whatsoever that rose to the level of constitutional violation. The evidence compiled by petitioners to demonstrate such attention by Congress to age discrimination by the States falls well short of the mark. That evidence consists almost entirely of isolated sentences clipped from floor debates and legislative reports.

[11] Finally, the United States' argument that Congress found substantial age discrimination in the private sector, is beside the point. Although we also have doubts whether the findings Congress did make with respect to the private sector could be extrapolated to support a finding of unconstitutional age discrimination in the public sector, it is sufficient for these cases to note that Congress failed to identify a widespread pattern of age discrimination by the States.

[12] A review of the ADEA's legislative record as a whole, then, reveals that Congress had virtually no reason to believe that state and local governments were unconstitutionally discriminating against their employees on the basis of age. Although that lack of support is not determinative of the § 5 inquiry, Congress' failure to uncover any significant pattern of unconstitutional discrimination here confirms that Congress had no reason to believe that broad prophylactic legislation was necessary in this field. In light of the indiscriminate scope of the Act's substantive requirements, and the lack of evidence of widespread and unconstitutional age discrimination by the States, we hold that the ADEA is not a valid exercise of Congress' power under § 5 of the Fourteenth Amendment. The ADEA's purported abrogation of the States' sovereign immunity is accordingly invalid.

[13] Our decision today does not signal the end of the line for employees who find themselves subject to age discrimination at the hands of their state employers. We hold only that, in the ADEA, Congress did not validly abrogate the States' sovereign immunity to suits by private individuals. State employees are protected by state age discrimination statutes, and may recover money damages from their state employers, in almost every State of the Union. Those avenues of relief remain available today, just as they were before this decision.

Justice Stevens, with whom Justice Souter, Justice Ginsburg, and Justice Breyer join, dissenting in part and concurring in part.

[1] Congress' power to regulate the American economy includes the power to regulate both the public and the private sectors of the labor market. Federal rules outlawing discrimination in the workplace, like the regulation of wages and hours or health and safety standards, may be enforced against public as well as private employers. In my opinion, Congress' power to authorize federal remedies against state agencies that violate federal statutory obligations is coextensive with its power to impose those obligations on the States in the first place. Neither the Eleventh Amendment nor the doctrine of sovereign immunity places any limit on that power.

[2] The application of the ancient judge-made doctrine of sovereign immunity in cases like these is supposedly justified as a freestanding limit on congressional authority, a limit necessary to protect States' "dignity and respect" from impairment by the National Government. The Framers did not, however, select the Judicial Branch as the constitutional guardian of those state interests. Rather, the Framers designed important structural safeguards to ensure that when the National Government enacted substantive law (and provided for its enforcement), the normal operation of the legislative process itself would adequately defend state interests from undue infringement. See generally Wechsler, The Political Safeguards of Federalism: The Role of the States in the Composition and Selection of the National Government, 54 Colum. L. Rev. 543 (1954).

[3] Federalism concerns do make it appropriate for Congress to speak clearly when it regulates state action. But when it does so, as it has in these cases, we can safely presume that the burdens the statute imposes on the sovereignty of the several States were taken into account during the deliberative process leading to the enactment of the measure. Those burdens necessarily include the cost of defending against enforcement proceedings and paying whatever penalties might be incurred for violating the statute. In my judgment, the question whether those enforcement proceedings should be conducted exclusively by federal agencies, or may be brought by private parties as well, is a matter of policy for Congress to decide. In either event, once Congress has made its policy choice, the sovereignty concerns of the several States are satisfied, and the federal interest in evenhanded enforcement of federal law, explicitly endorsed in Article VI of the Constitution, does not countenance further limitations. There is not a word in the text of the Constitution supporting the Court's conclusion that the judgemade doctrine of sovereign immunity limits Congress' power to authorize private parties, as well as federal agencies, to enforce federal law against the States. The importance of respecting the Framers' decision to assign the business of lawmaking to the Congress dictates firm resistance to the present majority's repeated substitution of its own views of federalism for those expressed in statutes enacted by the Congress and signed by the President.

[4] The Eleventh Amendment simply does not support the Court's view. As has been stated before, the Amendment only places a textual limitation on the diversity jurisdiction of the federal courts. Here, however, private petitioners did not invoke the federal courts' diversity jurisdiction; they are citizens of the same State as the defendants and they are asserting claims that arise under federal law. Thus, today's decision (relying as it does on Seminole Tribe ) rests entirely on a novel judicial interpretation of the doctrine of sovereign immunity, which the Court treats as though it were a constitutional precept. It is nevertheless clear to me that if Congress has the power to create the federal rights that these petitioners are asserting, it must also have the power to give the federal courts jurisdiction to remedy violations of those rights, even if it is necessary to "abrogate" the Court's "Eleventh Amendment" version of the common-law defense of sovereign immunity to do so. That is the essence of the Court's holding in Pennsylvania v. Union Gas Co. (1989).

[5] I remain convinced that Union Gas was correctly decided and that the decision of five Justices in Seminole Tribe to overrule that case was profoundly misguided. Despite my respect for stare decisis, I am unwilling to accept Seminole Tribe as controlling precedent. The kind of judicial activism manifested in cases like Seminole Tribe, Alden v. Maine, Florida Prepaid Postsecondary Ed. Expense Bd. v. College Savings Bank, represents such a radical departure from the proper role of this Court that it should be opposed whenever the opportunity arises.

Chief Justice Rehnquist delivered the opinion of the Court.

[1] We decide here whether employees of the State of Alabama may recover money damages by reason of the State's failure to comply with the provisions of Title I of the Americans with Disabilities Act of 1990 (ADA or Act). We hold that such suits are barred by the Eleventh Amendment.

[2] The ADA prohibits certain employers, including the States, from "discriminat[ing] against a qualified individual with a disability because of the disability of such individual in regard to job application procedures, the hiring, advancement, or discharge of employees, employee compensation, job training, and other terms, conditions, and privileges of employment." To this end, the Act requires employers to "mak[e] reasonable accommodations to the known physical or mental limitations of an otherwise qualified individual with a disability who is an applicant or employee, unless [the employer] can demonstrate that the accommodation would impose an undue hardship on the operation of the [employer's] business."

[3] Respondent Patricia Garrett, a registered nurse, was employed as the Director of Nursing, OB/Gyn/Neonatal Services, for the University of Alabama in Birmingham Hospital. In 1994, Garrett was diagnosed with breast cancer and subsequently underwent a lumpectomy, radiation treatment, and chemotherapy. Garrett's treatments required her to take substantial leave from work. Upon returning to work in July 1995, Garrett's supervisor informed Garrett that she would have to give up her Director position. Garrett then applied for and received a transfer to another, lower paying position as a nurse manager.

[4] We have recognized, however, that Congress may abrogate the States' Eleventh Amendment immunity when it both unequivocally intends to do so and "act[s] pursuant to a valid grant of constitutional authority." The first of these requirements is not in dispute here. See 42 U. S. C. § 12202 ("A State shall not be immune under the eleventh amendment to the Constitution of the United States from an action in [a] Federal or State court of competent jurisdiction for a violation of this chapter"). The question, then, is whether Congress acted within its constitutional authority by subjecting the States to suits in federal court for money damages under the ADA.

[5] Congress may not, of course, base its abrogation of the States' Eleventh Amendment immunity upon the powers enumerated in Article I. In Fitzpatrick v. Blitzer (1976), however, we held that "the Eleventh Amendment, and the principle of state sovereignty which it embodies, are necessarily limited by the enforcement provisions of § 5 of the Fourteenth Amendment." As a result, we concluded, Congress may subject nonconsenting States to suit in federal court when it does so pursuant to a valid exercise of its § 5 power. See ibid. Our cases have adhered to this proposition. Accordingly, the ADA can apply to the States only to the extent that the statute is appropriate § 5 legislation.

[6] Congress is not limited to mere legislative repetition of this Court's constitutional jurisprudence. "Rather, Congress' power `to enforce' the Amendment includes the authority both to remedy and to deter violation of rights guaranteed thereunder by prohibiting a somewhat broader swath of conduct, including that which is not itself forbidden by the Amendment's text.

[7] City of Boerne also confirmed, however, the long-settled principle that it is the responsibility of this Court, not Congress, to define the substance of constitutional guarantees. Accordingly, § 5 legislation reaching beyond the scope of § 1's actual guarantees must exhibit "congruence and proportionality between the injury to be prevented or remedied and the means adopted to that end."

[8] The first step in applying these now familiar principles is to identify with some precision the scope of the constitutional right at issue. Here, that inquiry requires us to examine the limitations § 1 of the Fourteenth Amendment places upon States' treatment of the disabled. We look to our prior decisions under the Equal Protection Clause dealing with this issue.

[9] In Cleburne v. Cleburne Living Center, Inc. (1985), we considered an equal protection challenge to a city ordinance requiring a special use permit for the operation of a group home for the mentally retarded. We conclud[ed] that such legislation incurs only the minimum "rational-basis" review applicable to general social and economic legislation. In a statement that today seems quite prescient, we explained that "if the large and amorphous class of the mentally retarded were deemed quasi-suspect for the reasons given by the Court of Appeals, it would be difficult to find a principled way to distinguish a variety of other groups who have perhaps immutable disabilities setting them off from others, who cannot themselves mandate the desired legislative responses, and who can claim some degree of prejudice from at least part of the public at large. One need mention in this respect only the aging, the disabled, the mentally ill, and the infirm. We are reluctant to set out on that course, and we decline to do so."

[10] Under rational-basis review, where a group possesses "distinguishing characteristics relevant to interests the State has the authority to implement," a State's decision to act on the basis of those differences does not give rise to a constitutional violation. "Such a classification cannot run afoul of the Equal Protection Clause if there is a rational relationship between the disparity of treatment and some legitimate governmental purpose." Heller v. Doe (1993).

[11] Thus, the result of Cleburne is that States are not required by the Fourteenth Amendment to make special accommodations for the disabled, so long as their actions toward such individuals are rational. They could quite hardheadedly—and perhaps hardheartedly—hold to job-qualification requirements which do not make allowance for the disabled. If special accommodations for the disabled are to be required, they have to come from positive law and not through the Equal Protection Clause.

[12] Once we have determined the metes and bounds of the constitutional right in question, we examine whether Congress identified a history and pattern of unconstitutional employment discrimination by the States against the disabled. Just as § 1 of the Fourteenth Amendment applies only to actions committed "under color of state law," Congress' § 5 authority is appropriately exercised only in response to state transgressions. The legislative record of the ADA, however, simply fails to show that Congress did in fact identify a pattern of irrational state discrimination in employment against the disabled.

[13] Congress made a general finding in the ADA that "historically, society has tended to isolate and segregate individuals with disabilities, and, despite some improvements, such forms of discrimination against individuals with disabilities continue to be a serious and pervasive social problem." The record assembled by Congress includes many instances to support such a finding. But the great majority of these incidents do not deal with the activities of States.

[14] Several of these incidents undoubtedly evidence an unwillingness on the part of state officials to make the sort of accommodations for the disabled required by the ADA. Whether they were irrational under our decision in Cleburne is more debatable, particularly when the incident is described out of context. But even if it were to be determined that each incident upon fuller examination showed unconstitutional action on the part of the State, these incidents taken together fall far short of even suggesting the pattern of unconstitutional discrimination on which § 5 legislation must be based. Congress, in enacting the ADA, found that "some 43,000,000 Americans have one or more physical or mental disabilities." In 1990, the States alone employed more than 4.5 million people. It is telling, we think, that given these large numbers, Congress assembled only such minimal evidence of unconstitutional state discrimination in employment against the disabled.

[15] Justice Breyer maintains that Congress applied Title I of the ADA to the States in response to a host of incidents representing unconstitutional state discrimination in employment against persons with disabilities. A close review of the relevant materials, however, undercuts that conclusion. Justice Breyer's Appendix C consists not of legislative findings, but of unexamined, anecdotal accounts of "adverse, disparate treatment by state officials." Of course, as we have already explained, "adverse, disparate treatment" often does not amount to a constitutional violation where rational-basis scrutiny applies. These accounts, moreover, were submitted not directly to Congress but to the Task Force on the Rights and Empowerment of Americans with Disabilities, which made no findings on the subject of state discrimination in employment. And, had Congress truly understood this information as reflecting a pattern of unconstitutional behavior by the States, one would expect some mention of that conclusion in the Act's legislative findings. There is none.

[16] Even were it possible to squeeze out of these examples a pattern of unconstitutional discrimination by the States, the rights and remedies created by the ADA against the States would raise the same sort of concerns as to congruence and proportionality as were found in City of Boerne. For example, whereas it would be entirely rational (and therefore constitutional) for a state employer to conserve scarce financial resources by hiring employees who are able to use existing facilities, the ADA requires employers to "mak[e] existing facilities used by employees readily accessible to and usable by individuals with disabilities." The ADA does except employers from the "reasonable accommodatio[n]" requirement where the employer "can demonstrate that the accommodation would impose an undue hardship on the operation of the business of such covered entity." However, even with this exception, the accommodation duty far exceeds what is constitutionally required in that it makes unlawful a range of alternative responses that would be reasonable but would fall short of imposing an "undue burden" upon the employer. The Act also makes it the employer's duty to prove that it would suffer such a burden, instead of requiring (as the Constitution does) that the complaining party negate reasonable bases for the employer's decision.

[17] The ADA's constitutional shortcomings are apparent when the Act is compared to Congress' efforts in the Voting Rights Act of 1965 to respond to a serious pattern of constitutional violations. In South Carolina v. Katzenbach (1966), we considered whether the Voting Rights Act was "appropriate" legislation to enforce the Fifteenth Amendment's protection against racial discrimination in voting. Concluding that it was a valid exercise of Congress' enforcement power under § 2 of the Fifteenth Amendment, we noted that "[b]efore enacting the measure, Congress explored with great care the problem of racial discrimination in voting."

[18] In that Act, Congress documented a marked pattern of unconstitutional action by the States. State officials, Congress found, routinely applied voting tests in order to exclude African-American citizens from registering to vote. Congress also determined that litigation had proved ineffective and that there persisted an otherwise inexplicable 50-percentage-point gap in the registration of white and African-American voters in some States. Congress' response was to promulgate in the Voting Rights Act a detailed but limited remedial scheme designed to guarantee meaningful enforcement of the Fifteenth Amendment in those areas of the Nation where abundant evidence of States' systematic denial of those rights was identified.

[19] The contrast between this kind of evidence, and the evidence that Congress considered in the present case, is stark. Congressional enactment of the ADA represents its judgment that there should be a "comprehensive national mandate for the elimination of discrimination against individuals with disabilities." Congress is the final authority as to desirable public policy, but in order to authorize private individuals to recover money damages against the States, there must be a pattern of discrimination by the States which violates the Fourteenth Amendment, and the remedy imposed by Congress must be congruent and proportional to the targeted violation. Those requirements are not met here, and to uphold the Act's application to the States would allow Congress to rewrite the Fourteenth Amendment law laid down by this Court in Cleburne. Section 5 does not so broadly enlarge congressional authority.

Justice Breyer, with whom Justice Stevens, Justice Souter, and Justice Ginsburg join, dissenting.

[1] Section 5 grants Congress the "power to enforce, by appropriate legislation," the Fourteenth Amendment's equal protection guarantee. As the Court recognizes, state discrimination in employment against persons with disabilities might "run afoul of the Equal Protection Clause" where there is no "rational relationship between the disparity of treatment and some legitimate governmental purpose." In my view, Congress reasonably could have concluded that the remedy before us constitutes an "appropriate" way to enforce this basic equal protection requirement. And that is all the Constitution requires.

[2] The Court says that its primary problem with this statutory provision is one of legislative evidence. It says that "Congress assembled only . . . minimal evidence of unconstitutional state discrimination in employment." In fact, Congress compiled a vast legislative record documenting "massive, society-wide discrimination" against persons with disabilities. In addition to the information presented at 13 congressional hearings, and its own prior experience gathered over 40 years during which it contemplated and enacted considerable similar legislation, Congress created a special task force to assess the need for comprehensive legislation. That task force held hearings in every State, attended by more than 30,000 people, including thousands who had experienced discrimination first hand. The task force hearings, Congress' own hearings, and an analysis of "census data, national polls, and other studies" led Congress to conclude that "people with disabilities, as a group, occupy an inferior status in our society, and are severely disadvantaged socially, vocationally, economically, and educationally." As to employment, Congress found that "[t]wo-thirds of all disabled Americans between the age of 16 and 64 [were] not working at all," even though a large majority wanted to, and were able to, work productively. And Congress found that this discrimination flowed in significant part from "stereotypic assumptions" as well as "purposeful unequal treatment."

[3] The powerful evidence of discriminatory treatment throughout society in general, including discrimination by private persons and local governments, implicates state governments as well, for state agencies form part of that same larger society. There is no particular reason to believe that they are immune from the "stereotypic assumptions" and pattern of "purposeful unequal treatment" that Congress found prevalent. The Court claims that it "make[s] no sense" to take into consideration constitutional violations committed by local governments. But the substantive obligation that the Equal Protection Clause creates applies to state and local governmental entities alike.

[4] In any event, there is no need to rest solely upon evidence of discrimination by local governments or general societal discrimination. There are roughly 300 examples of discrimination by state governments themselves in the legislative record. I fail to see how this evidence "fall[s] far short of even suggesting the pattern of unconstitutional discrimination on which § 5 legislation must be based."

[5] The congressionally appointed task force collected numerous specific examples, provided by persons with disabilities themselves, of adverse, disparate treatment by state officials. They reveal, not what the Court describes as "half a dozen" instances of discrimination, but hundreds of instances of adverse treatment at the hands of state officials— instances in which a person with a disability found it impossible to obtain a state job, to retain state employment, to use the public transportation that was readily available to others in order to get to work, or to obtain a public education, which is often a prerequisite to obtaining employment. State-imposed barriers also frequently made it difficult or impossible for people to vote, to enter a public building, to access important government services, such as calling for emergency assistance, and to find a place to live due to a pattern of irrational zoning decisions similar to the discrimination that we held unconstitutional in Cleburne.

[6] Regardless, Congress expressly found substantial unjustified discrimination against persons with disabilities. The evidence in the legislative record bears out Congress' finding that the adverse treatment of persons with disabilities was often arbitrary or invidious in this sense, and thus unjustified. For example, one study that was before Congress revealed that "most . . . governmental agencies in [one State] discriminated in hiring against job applicants for an average period of five years after treatment for cancer," based in part on coworkers' misguided belief that "cancer is contagious." A school inexplicably refused to exempt a deaf teacher, who taught at a school for the deaf, from a "listening skills" requirement. A State refused to hire a blind employee as director of an agency for the blind—even though he was the most qualified applicant. Certain state agencies apparently had general policies against hiring or promoting persons with disabilities. A zoo turned away children with Downs Syndrome "because [the zookeeper] feared they would upset the chimpanzees." There were reports of numerous zoning decisions based upon "negative attitudes" or "fear," such as a zoning board that denied a permit for an obviously pretextual reason after hearing arguments that a facility would house "`deviants' " who needed "room to roam." Congress could have reasonably believed that these examples represented signs of a widespread problem of unconstitutional discrimination.

[7] The Court argues in the alternative that the statute's damages remedy is not "congruent" with and "proportional" to the equal protection problem that Congress found. The Court suggests that the Act's "reasonable accommodation" requirement, and disparate-impact standard, "far excee[d] what is constitutionally required." But we have upheld disparate-impact standards in contexts where they were not "constitutionally required."

[8] And what is wrong with a remedy that, in response to unreasonable employer behavior, requires an employer to make accommodations that are reasonable? Of course, what is "reasonable" in the statutory sense and what is "unreasonable" in the constitutional sense might differ. In other words, the requirement may exceed what is necessary to avoid a constitutional violation. But it is just that power— the power to require more than the minimum—that § 5 grants to Congress, as this Court has repeatedly confirmed.

Chief Justice Rehnquist delivered the opinion of the Court.

[1] The Family and Medical Leave Act of 1993 (FMLA or Act) entitles eligible employees to take up to 12 work weeks of unpaid leave annually for any of several reasons, including the onset of a "serious health condition" in an employee's spouse, child, or parent. The Act creates a private right of action to seek both equitable relief and money damages "against any employer (including a public agency) in any Federal or State court of competent jurisdiction," should that employer "interfere with, restrain, or deny the exercise of "FMLA rights, § 2615(a)(1). We hold that employees of the State of Nevada may recover money damages in the event of the State's failure to comply with the family-care provision of the Act.

[2] Petitioners include the Nevada Department of Human Resources (Department) and two of its officers. Respondent William Hibbs (hereinafter respondent) worked for the Department's Welfare Division. In April and May 1997, he sought leave under the FMLA to care for his ailing wife, who was recovering from a car accident and neck surgery. The Department granted his request for the full 12 weeks of FMLA leave and authorized him to use the leave intermittently as needed between May and December 1997. Respondent did so until August 5, 1997, after which he did not return to work. In October 1997, the Department informed respondent that he had exhausted his FMLA leave, that no further leave would be granted, and that he must report to work by November 12, 1997. Respondent failed to do so and was terminated. Respondent sued petitioners in the United States District Court seeking damages and injunctive and declaratory relief.

[3] For over a century now, we have made clear that the Constitution does not provide for federal jurisdiction over suits against nonconsenting States. Congress may, however, abrogate such immunity in federal court if it makes its intention to abrogate unmistakably clear in the language of the statute and acts pursuant to a valid exercise of its power under § 5 of the Fourteenth Amendment. The clarity of Congress' intent here is not fairly debatable. The Act enables employees to seek damages "against any employer (including a public agency) in any Federal or State court of competent jurisdiction," and Congress has defined "public agency" to include both "the government of a State or political subdivision thereof" and "any agency of ... a State, or a political subdivision of a State.”

[4] This case turns, then, on whether Congress acted within its constitutional authority when it sought to abrogate the States' immunity for purposes of the FMLA's family-leave provision.

[5] The FMLA aims to protect the right to be free from gender-based discrimination in the workplace. We have held that statutory classifications that distinguish between males and females are subject to heightened scrutiny. For a gender-based classification to withstand such scrutiny, it must "serv[e] important governmental objectives," and "the discriminatory means employed [must be] substantially related to the achievement of those objectives." The State's justification for such a classification "must not rely on overbroad generalizations about the different talents, capacities, or preferences of males and females." We now inquire whether Congress had evidence of a pattern of constitutional violations on the part of the States in this area.

[6] The history of the many state laws limiting women's employment opportunities is chronicled in — and, until relatively recently, was sanctioned by — this Court's own opinions. The long and extensive history of sex discrimination prompted us to hold that measures that differentiate on the basis of gender warrant heightened scrutiny; here, the persistence of such unconstitutional discrimination by the States justifies Congress' passage of prophylactic § 5 legislation.

[7] As the FMLA's legislative record reflects, a 1990 Bureau of Labor Statistics (BLS) survey stated that 37 percent of surveyed private-sector employees were covered by maternity leave policies, while only 18 percent were covered by paternity leave policies. The corresponding numbers from a similar BLS survey the previous year were 33 percent and 16 percent, respectively. While these data show an increase in the percentage of employees eligible for such leave, they also show a widening of the gender gap during the same period. Thus, stereotype-based beliefs about the allocation of family duties remained firmly rooted, and employers' reliance on them in establishing discriminatory leave policies remained widespread. Congress also heard testimony that "[p]arental leave for fathers ... is rare. Even ... [w]here child-care leave policies do exist, men, both in the public and private sectors, receive notoriously discriminatory treatment in their requests for such leave." Fifteen States provided women up to one year of extended maternity leave, while only four provided men with the same. This and other differential leave policies were not attributable to any differential physical needs of men and women, but rather to the pervasive sex-role stereotype that caring for family members is women's work.

[8] Finally, Congress had evidence that, even where state laws and policies were not facially discriminatory, they were applied in discriminatory ways. It was aware of the "serious problems with the discretionary nature of family leave," because when "the authority to grant leave and to arrange the length of that leave rests with individual supervisors," it leaves "employees open to discretionary and possibly unequal treatment." Testimony supported that conclusion, explaining that "[t]he lack of uniform parental and medical leave policies in the work place has created an environment where [sex] discrimination is rampant."

[9] In sum, the States' record of unconstitutional participation in, and fostering of, gender-based discrimination in the administration of leave benefits is weighty enough to justify the enactment of prophylactic § 5 legislation.

[10] We reached the opposite conclusion in Garrett and Kimel. In those cases, the § 5 legislation under review responded to a purported tendency of state officials to make age-or disability-based distinctions. Under our equal protection case law, discrimination on the basis of such characteristics is not judged under a heightened review standard, and passes muster if there is "a rational basis for doing so at a class-based level, even if it `is probably not true' that those reasons are valid in the majority of cases."  Thus, in order to impugn the constitutionality of state discrimination against the disabled or the elderly, Congress must identify, not just the existence of age-or disability-based state decisions, but a "widespread pattern" of irrational reliance on such criteria. We found no such showing with respect to the ADEA and Title I of the Americans with Disabilities Act of 1990 (ADA). Here, however, Congress directed its attention to state gender discrimination, which triggers a heightened level of scrutiny. Because the standard for demonstrating the constitutionality of a gender-based classification is more difficult to meet than our rational-basis test—it must "serv[e] important governmental objectives" and be "substantially related to the achievement of those objectives,"— it was easier for Congress to show a pattern of state constitutional violations. Congress was similarly successful in South Carolina v. Katzenbach (1966), where we upheld the Voting Rights Act of 1965: Because racial classifications are presumptively invalid, most of the States' acts of race discrimination violated the Fourteenth Amendment.

[11] The impact of the discrimination targeted by the FMLA is significant. Congress determined: "Historically, denial or curtailment of women's employment opportunities has been traceable directly to the pervasive presumption that women are mothers first, and workers second. This prevailing ideology about women's roles has in turn justified discrimination against women when they are mothers or mothers-to-be." Stereotypes about women's domestic roles are reinforced by parallel stereotypes presuming a lack of domestic responsibilities for men. Because employers continued to regard the family as the woman's domain, they often denied men similar accommodations or discouraged them from taking leave. These mutually reinforcing stereotypes created a self-fulfilling cycle of discrimination that forced women to continue to assume the role of primary family caregiver, and fostered employers' stereotypical views about women's commitment to work and their value as employees. Those perceptions, in turn, Congress reasoned, lead to subtle discrimination that may be difficult to detect on a case-by-case basis.

[12] We also find significant the many other limitations that Congress placed on the scope of this measure. The FMLA requires only unpaid leave, and applies only to employees who have worked for the employer for at least one year and provided 1,250 hours of service within the last 12 months. Employees in high-ranking or sensitive positions are simply ineligible for FMLA leave; of particular importance to the States, the FMLA expressly excludes from coverage state elected officials, their staffs, and appointed policymakers. Employees must give advance notice of foreseeable leave, and employers may require certification by a health care provider of the need for leave. In choosing 12 weeks as the appropriate leave floor, Congress chose "a middle ground, a period long enough to serve `the needs of families' but not so long that it would upset `the legitimate interests of employers.'" Moreover, the cause of action under the FMLA is a restricted one: The damages recoverable are strictly defined and measured by actual monetary losses, and the accrual period for backpay is limited by the Act's 2-year statute of limitations (extended to three years only for willful violations).

[13] For the above reasons, we conclude that § 2612(a)(1)(C) is congruent and proportional to its remedial object, and can "be understood as responsive to, or designed to prevent, unconstitutional behavior."

Justice Scalia, dissenting.

[1] I join Justice Kennedy’s dissent, and add one further observation: The constitutional violation that is a prerequisite to "prophylactic" congressional action to "enforce" the Fourteenth Amendment is a violation by the State against which the enforcement action is taken. There is no guilt by association, enabling the sovereignty of one State to be abridged under § 5 of the Fourteenth Amendment because of violations by another State, or by most other States, or even  by 49 other States.

[2] Today's opinion for the Court does not even attempt to demonstrate that each one of the 50 States covered by 29 U.S.C. § 2612(a)(1)(C) was in violation of the Fourteenth Amendment. It treats "the States" as some sort of collective entity which is guilty or innocent as a body. "[T]he States' record of unconstitutional participation in, and fostering of, gender-based discrimination," it concludes, "is weighty enough to justify the enactment of prophylactic § 5 legislation." This will not do. Prophylaxis in the sense of extending the remedy beyond the violation is one thing; prophylaxis in the sense of extending the remedy beyond the violator is something else.

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

[1] The Court is unable to show that States have engaged in a pattern of unlawful conduct which warrants the remedy of opening state treasuries to private suits. The inability to adduce evidence of alleged discrimination, coupled with the inescapable fact that the federal scheme is not a remedy but a benefit program, demonstrates the lack of the requisite link between any problem Congress has identified and the program it mandated.

[2] The relevant question, as the Court seems to acknowledge, is whether, notwithstanding the passage of Title VII and similar state legislation, the States continued to engage in widespread discrimination on the basis of gender in the provision of family leave benefits. If such a pattern were shown, the Eleventh Amendment would not bar Congress from devising a congruent and proportional remedy. The evidence to substantiate this charge must be far more specific, however, than a simple recitation of a general history of employment discrimination against women. When the federal statute seeks to abrogate state sovereign immunity, the Court should be more careful to insist on adherence to the analytic requirements set forth in its own precedents. Persisting overall effects of gender-based discrimination at the workplace must not be ignored; but simply noting the problem is not a substitute for evidence which identifies some real discrimination the family leave rules are designed to prevent. Respondents fail to make the requisite showing. The Act's findings of purpose are devoid of any discussion of the relevant evidence.

[3] As the Court seems to recognize, the evidence considered by Congress concerned discriminatory practices of the private sector, not those of state employers. The statistical information compiled by the Bureau of Labor Statistics (BLS), which are the only factual findings the Court cites, surveyed only private employers. Ante, at 730. While the evidence of discrimination by private entities may be relevant, it does not, by itself, justify the abrogation of States' sovereign immunity.

[4] Even if there were evidence that individual state employers, in the absence of clear statutory guidelines, discriminated in the administration of leave benefits, this circumstance alone would not support a finding of a state-sponsored pattern of discrimination. The evidence could perhaps support the charge of disparate impact, but not a charge that States have engaged in a pattern of intentional discrimination prohibited by the Fourteenth Amendment. The federal-state equivalence upon which the Court places such emphasis is a deficient rationale at an even more fundamental level, however; for the States appear to have been ahead of Congress in providing gender-neutral family leave benefits. Thirty States, the District of Columbia, and Puerto Rico had adopted some form of family-care leave in the years preceding the Act's adoption.

[5] The Court acknowledges that States have adopted family leave programs prior to federal intervention, but argues these policies suffered from serious imperfections. Even if correct, this observation proves, at most, that programs more generous and more effective than those operated by the States were feasible. That the States did not devise the optimal programs is not, however, evidence that the States were perpetuating unconstitutional discrimination. Given that the States assumed a pioneering role in the creation of family leave schemes, it is not surprising these early efforts may have been imperfect. This is altogether different, however, from purposeful discrimination.

[6] The Court's lengthy discussion of the allegedly deficient state policies falls short of meeting this standard. A great majority of these programs exhibit no constitutional defect and, in fact, are authorized by this Court's precedent. The Court points out that seven States adopted leave provisions applicable only to women. Yet it must acknowledge that three of these schemes concerned solely pregnancy disability leave. Our cases make clear that a State does not violate the Equal Protection Clause by granting pregnancy disability leave to women without providing for a grant of parenting leave to men. Geduldig v. Aiello (1974).

[7] Considered in its entirety, the evidence fails to document a pattern of unconstitutional conduct sufficient to justify the abrogation of States' sovereign immunity. The few incidents identified by the Court "fall far short of even suggesting the pattern of unconstitutional discrimination on which § 5 legislation must be based." Juxtaposed to this evidence is the States' record of addressing gender-based discrimination in the provision of leave benefits on their own volition.

[8] Our concern with gender discrimination, which is subjected to heightened scrutiny, as opposed to age-or disability-based distinctions, which are reviewed under rational standard, does not alter this conclusion. The application of heightened scrutiny is designed to ensure gender-based classifications are not based on the entrenched and pervasive stereotypes which inhibit women's progress in the workplace. Ante, at 736. This consideration does not divest respondents of their burden to show that "Congress identified a history and pattern of unconstitutional employment discrimination by the States." The Court seems to reaffirm this requirement. In my submission, however, the Court does not follow it. Given the insufficiency of the evidence that States discriminated in the provision of family leave, the unfortunate fact that stereotypes about women continue to be a serious and pervasive social problem would not alone support the charge that a State has engaged in a practice designed to deny its citizens the equal protection of the laws.

[9] It bears emphasis that, even were the Court to bar unconsented federal suits by private individuals for money damages from a State, individuals whose rights under the Act were violated would not be without recourse. The Act is likely a valid exercise of Congress' power under the Commerce Clause, and so the standards it prescribes will be binding upon the States. The United States may enforce these standards in actions for money damages; and private individuals may bring actions against state officials for injunctive relief under Ex parte Young (1908). What is at issue is only whether the States can be subjected, without consent, to suits brought by private persons seeking to collect moneys from the state treasury. Their immunity cannot be abrogated without documentation of a pattern of unconstitutional acts by the States, and only then by a congruent and proportional remedy. There has been a complete failure by respondents to carry their burden to establish each of these necessary propositions. I would hold that the Act is not a valid abrogation of state sovereign immunity and dissent with respect from the Court's conclusion to the contrary.

Chief Justice ROBERTS delivered the opinion of the Court.

[1] The Voting Rights Act of 1965 employed extraordinary measures to address an extraordinary problem. Section 5 of the Act required States to obtain federal permission before enacting any law related to voting — a drastic departure from basic principles of federalism. And § 4 of the Act applied that requirement only to some States — an equally dramatic departure from the principle that all States enjoy equal sovereignty. This was strong medicine, but Congress determined it was needed to address entrenched racial discrimination in voting, "an insidious and pervasive evil which had been perpetuated in certain parts of our country through unremitting and ingenious defiance of the Constitution." South Carolina v. Katzenbach, 383 U.S. 301 (1966). As we explained in upholding the law, "exceptional conditions can justify legislative measures not otherwise appropriate." Id., at 334. Reflecting the unprecedented nature of these measures, they were scheduled to expire after five years.

[2] Nearly 50 years later, they are still in effect; indeed, they have been made more stringent, and are now scheduled to last until 2031. There is no denying, however, that the conditions that originally justified these measures no longer characterize voting in the covered jurisdictions. By 2009, "the racial gap in voter registration and turnout [was] lower in the States originally covered by § 5 than it [was] nationwide." Since that time, Census Bureau data indicate that African-American voter turnout has come to exceed white voter turnout in five of the six States originally covered by § 5, with a gap in the sixth State of less than one half of one percent. See Dept. of Commerce, Census Bureau, Reported Voting and Registration, by Sex, Race and Hispanic Origin, for States (Nov. 2012) (Table 4b).

[3] At the same time, voting discrimination still exists; no one doubts that. The question is whether the Act's extraordinary measures, including its disparate treatment of the States, continue to satisfy constitutional requirements. As we put it a short time ago, "the Act imposes current burdens and must be justified by current needs." Northwest Austin, 557 U.S., at 203.

I

A

[4] The Fifteenth Amendment was ratified in 1870, in the wake of the Civil War. It provides that "[t]he right of citizens of the United States to vote shall not be denied or abridged by the United States or by any State on account of race, color, or previous condition of servitude," and it gives Congress the "power to enforce this article by appropriate legislation."

[5] "The first century of congressional enforcement of the Amendment, however, can only be regarded as a failure.”  In the 1890s, Alabama, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, and Virginia began to enact literacy tests for voter registration and to employ other methods designed to prevent African-Americans from voting. Congress passed statutes outlawing some of these practices and facilitating litigation against them, but litigation remained slow and expensive, and the States came up with new ways to discriminate as soon as existing ones were struck down. Voter registration of African-Americans barely improved.

[6] Inspired to action by the civil rights movement, Congress responded in 1965 with the Voting Rights Act. Section 2 was enacted to forbid, in all 50 States, any "standard, practice, or procedure ... imposed or applied ... to deny or abridge the right of any citizen of the United States to vote on account of race or color." The current version forbids any "standard, practice, or procedure" that "results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color." Both the Federal Government and individuals have sued to enforce § 2, and injunctive relief is available in appropriate cases to block voting laws from going into effect. Section 2 is permanent, applies nationwide, and is not at issue in this case.

[7] Other sections targeted only some parts of the country. At the time of the Act's passage, these "covered" jurisdictions were those States or political subdivisions that had maintained a test or device as a prerequisite to voting as of November 1, 1964, and had less than 50 percent voter registration or turnout in the 1964 Presidential election. Such tests or devices included literacy and knowledge tests, good moral character requirements, the need for vouchers from registered voters, and the like. Covered jurisdiction could "bail out" of coverage if it had not used a test or device in the preceding five years "for the purpose or with the effect of denying or abridging the right to vote on account of race or color." In 1965, the covered States included Alabama, Georgia, Louisiana, Mississippi, South Carolina, and Virginia. The additional covered subdivisions included 39 counties in North Carolina and one in Arizona.

[8] In those jurisdictions, § 4 of the Act banned all such tests or devices. Section 5 provided that no change in voting procedures could take effect until it was approved by federal authorities in Washington, D.C. — either the Attorney General or a court of three judges. A jurisdiction could obtain such "preclearance" only by proving that the change had neither "the purpose [nor] the effect of denying or abridging the right to vote on account of race or color."

[9] Sections 4 and 5 were intended to be temporary; they were set to expire after five years. In South Carolina v. Katzenbach, we upheld the 1965 Act against constitutional challenge, explaining that it was justified to address "voting discrimination where it persists on a pervasive scale."

[7] In 1970, Congress reauthorized the Act for another five years, and extended the coverage formula in § 4(b) to jurisdictions that had a voting test and less than 50 percent voter registration or turnout as of 1968. That swept in several counties in California, New Hampshire, and New York. Congress also extended the ban in § 4(a) on tests and devices nationwide.

[8] In 1975, Congress reauthorized the Act for seven more years, and extended its coverage to jurisdictions that had a voting test and less than 50 percent voter registration or turnout as of 1972.  Congress also amended the definition of "test or device" to include the practice of providing English-only voting materials in places where over five percent of voting-age citizens spoke a single language other than English. As a result of these amendments, the States of Alaska, Arizona, and Texas, as well as several counties in California, Florida, Michigan, New York, North Carolina, and South Dakota, became covered jurisdictions. Congress correspondingly amended sections 2 and 5 to forbid voting discrimination on the basis of membership in a language minority group, in addition to discrimination on the basis of race or color. Finally, Congress made the nationwide ban on tests and devices permanent.

[9] In 1982, Congress reauthorized the Act for 25 years, but did not alter its coverage formula. See Voting Rights Act Amendments. Congress did, however, amend the bailout provisions, allowing political subdivisions of covered jurisdictions to bail out. Among other prerequisites for bailout, jurisdictions and their subdivisions must not have used a forbidden test or device, failed to receive preclearance, or lost a § 2 suit, in the ten years prior to seeking bailout.

[10] We upheld each of these reauthorizations against constitutional challenge.

[11] In 2006, Congress again reauthorized the Voting Rights Act for 25 years, again without change to its coverage formula. Congress also amended § 5 to prohibit more conduct than before. Section 5 now forbids voting changes with "any discriminatory purpose" as well as voting changes that diminish the ability of citizens, on account of race, color, or language minority status, "to elect their preferred candidates of choice."

[12] Shortly after this reauthorization, a Texas utility district brought suit, seeking to bail out from the Act's coverage and, in the alternative, challenging the Act's constitutionality. A three-judge District Court explained that only a State or political subdivision was eligible to seek bailout under the statute, and concluded that the utility district was not a political subdivision, a term that encompassed only "counties, parishes, and voter-registering subunits." The District Court also rejected the constitutional challenge.

[13] We reversed. We explained that "`normally the Court will not decide a constitutional question if there is some other ground upon which to dispose of the case.'" Concluding that "underlying constitutional concerns," among other things, "compel[led] a broader reading of the bailout provision," we construed the statute to allow the utility district to seek bailout. In doing so we expressed serious doubts about the Act's continued constitutionality.

[14] We explained that § 5 "imposes substantial federalism costs" and "differentiates between the States, despite our historic tradition that all the States enjoy equal sovereignty." We also noted that "[t]hings have changed in the South. Voter turnout and registration rates now approach parity. Blatantly discriminatory evasions of federal decrees are rare. And minority candidates hold office at unprecedented levels. Finally, we questioned whether the problems that § 5 meant to address were still "concentrated in the jurisdictions singled out for preclearance."

[15] Eight Members of the Court subscribed to these views, and the remaining Member would have held the Act unconstitutional. Ultimately, however, the Court's construction of the bailout provision left the constitutional issues for another day.

B

[16] Shelby County is located in Alabama, a covered jurisdiction. It has not sought bailout, as the Attorney General has recently objected to voting changes proposed from within the county. Instead, in 2010, the county sued the Attorney General in Federal District Court in Washington, D.C., seeking a declaratory judgment that sections 4(b) and 5 of the Voting Rights Act are facially unconstitutional, as well as a permanent injunction against their enforcement. The District Court ruled against the county and upheld the Act. The court found that the evidence before Congress in 2006 was sufficient to justify reauthorizing § 5 and continuing the § 4(b) coverage formula.

[17] The Court of Appeals for the D.C. Circuit affirmed.

[***]

[18] We granted certiorari.

II

[19] In Northwest Austin, we stated that "the Act imposes current burdens and must be justified by current needs." And we concluded that "a departure from the fundamental principle of equal sovereignty requires a showing that a statute's disparate geographic coverage is sufficiently related to the problem that it targets." These basic principles guide our review of the question before us.

A

[20] The Constitution and laws of the United States are "the supreme Law of the Land." U.S. Const., Art. VI, cl. 2. State legislation may not contravene federal law. The Federal Government does not, however, have a general right to review and veto state enactments before they go into effect. A proposal to grant such authority to "negative" state laws was considered at the Constitutional Convention, but rejected in favor of allowing state laws to take effect, subject to later challenge under the Supremacy Clause.

[21] Outside the strictures of the Supremacy Clause, States retain broad autonomy in structuring their governments and pursuing legislative objectives. Indeed, the Constitution provides that all powers not specifically granted to the Federal Government are reserved to the States or citizens. Amdt. 10. This "allocation of powers in our federal system preserves the integrity, dignity, and residual sovereignty of the States." But the federal balance "is not just an end in itself: Rather, federalism secures to citizens the liberties that derive from the diffusion of sovereign power."

[22] More specifically, "`the Framers of the Constitution intended the States to keep for themselves, as provided in the Tenth Amendment, the power to regulate elections.'" Of course, the Federal Government retains significant control over federal elections. For instance, the Constitution authorizes Congress to establish the time and manner for electing Senators and Representatives. But States have "broad powers to determine the conditions under which the right of suffrage may be exercised." And "[e]ach State has the power to prescribe the qualifications of its officers and the manner in which they shall be chosen." Drawing lines for congressional districts is likewise "primarily the duty and responsibility of the State."

[23] Not only do States retain sovereignty under the Constitution, there is also a "fundamental principle of equal sovereignty" among the States. Over a hundred years ago, this Court explained that our Nation "was and is a union of States, equal in power, dignity and authority." Indeed, "the constitutional equality of the States is essential to the harmonious operation of the scheme upon which the Republic was organized." Coyle concerned the admission of new States, and Katzenbach rejected the notion that the principle operated as a bar on differential treatment outside that context. At the same time, as we made clear in Northwest Austin, the fundamental principle of equal sovereignty remains highly pertinent in assessing subsequent disparate treatment of States.

[24] The Voting Rights Act sharply departs from these basic principles. It suspends "all changes to state election law — however innocuous — until they have been precleared by federal authorities in Washington, D.C." States must beseech the Federal Government for permission to implement laws that they would otherwise have the right to enact and execute on their own, subject of course to any injunction in a § 2 action. The Attorney General has 60 days to object to a preclearance request, longer if he requests more information. If a State seeks preclearance from a three-judge court, the process can take years.

[25] And despite the tradition of equal sovereignty, the Act applies to only nine States (and several additional counties). While one State waits months or years and expends funds to implement a validly enacted law, its neighbor can typically put the same law into effect immediately, through the normal legislative process. Even if a non-covered jurisdiction is sued, there are important differences between those proceedings and preclearance proceedings; the preclearance proceeding "not only switches the burden of proof to the supplicant jurisdiction, but also applies substantive standards quite different from those governing the rest of the nation."

[26] All this explains why, when we first upheld the Act in 1966, we described it as "stringent" and "potent." We recognized that it "may have been an uncommon exercise of congressional power," but concluded that "legislative measures not otherwise appropriate" could be justified by "exceptional conditions." We have since noted that the Act "authorizes federal intrusion into sensitive areas of state and local policymaking," and represents an "extraordinary departure from the traditional course of relations between the States and the Federal Government," As we reiterated in Northwest Austin, the Act constitutes "extraordinary legislation otherwise unfamiliar to our federal system."

B

[27] In 1966, we found these departures from the basic features of our system of government justified. The "blight of racial discrimination in voting" had "infected the electoral process in parts of our country for nearly a century."

[ * * * ]

[28] In short, we concluded that "[u]nder the compulsion of these unique circumstances, Congress responded in a permissibly decisive manner." We also noted then and have emphasized since that this extraordinary legislation was intended to be temporary, set to expire after five years.

[29] At the time, the coverage formula — the means of linking the exercise of the unprecedented authority with the problem that warranted it — made sense. We found that "Congress chose to limit its attention to the geographic areas where immediate action seemed necessary.” The areas where Congress found "evidence of actual voting discrimination" shared two characteristics: "the use of tests and devices for voter registration, and a voting rate in the 1964 presidential election at least 12 points below the national average."

[ * * * ]

C

[30] Nearly 50 years later, things have changed dramatically. Shelby County contends that the preclearance requirement, even without regard to its disparate coverage, is now unconstitutional. Its arguments have a good deal of force. In the covered jurisdictions, "[v]oter turnout and registration rates now approach parity. Blatantly discriminatory evasions of federal decrees are rare. And minority candidates hold office at unprecedented levels." The tests and devices that blocked access to the ballot have been forbidden nationwide for over 40 years.

[31] Those conclusions are not ours alone. Congress said the same when it reauthorized the Act in 2006, writing that "[s]ignificant progress has been made in eliminating first generation barriers experienced by minority voters, including increased numbers of registered minority voters, minority voter turnout, and minority representation in Congress, State legislatures, and local elected offices." The House Report elaborated that "the number of African-Americans who are registered and who turn out to cast ballots has increased significantly over the last 40 years, particularly since 1982," and noted that "[i]n some circumstances, minorities register to vote and cast ballots at levels that surpass those of white voters." That Report also explained that there have been "significant increases in the number of African-Americans serving in elected offices"; more specifically, there has been approximately a 1,000 percent increase since 1965 in the number of African-American elected officials in the six States originally covered by the Voting Rights Act.

[ * * * ]

[32] Census Bureau data from the most recent election indicate that African-American voter turnout exceeded white voter turnout in five of the six States originally covered by § 5, with a gap in the sixth State of less than one half of one percent. See Dept. of Commerce, Census Bureau, Reported Voting and Registration, by Sex, Race and Hispanic Origin, for States (Table 4b).

[ * * * ]

[33] There is no doubt that these improvements are in large part because of the Voting Rights Act. The Act has proved immensely successful at redressing racial discrimination and integrating the voting process. During the "Freedom Summer" of 1964, in Philadelphia, Mississippi, three men were murdered while working in the area to register African-American voters. On "Bloody Sunday" in 1965, in Selma, Alabama, police beat and used tear gas against hundreds marching in support of African-American enfranchisement. Today both of those towns are governed by African-American mayors. Problems remain in these States and others, but there is no denying that, due to the Voting Rights Act, our Nation has made great strides.

[34] Yet the Act has not eased the restrictions in § 5 or narrowed the scope of the coverage formula in § 4(b) along the way. Those extraordinary and unprecedented features were reauthorized — as if nothing had changed. In fact, the Act's unusual remedies have grown even stronger. When Congress reauthorized the Act in 2006, it did so for another 25 years on top of the previous 40 — a far cry from the initial five-year period. Congress also expanded the prohibitions in § 5. We had previously interpreted § 5 to prohibit only those redistricting plans that would have the purpose or effect of worsening the position of minority groups. In 2006, Congress amended § 5 to prohibit laws that could have favored such groups but did not do so because of a discriminatory purpose, even though we had stated that such broadening of § 5 coverage would "exacerbate the substantial federalism costs that the preclearance procedure already exacts, perhaps to the extent of raising concerns about § 5's constitutionality."

[ * * * ]

[35] Respondents do not deny that there have been improvements on the ground, but argue that much of this can be attributed to the deterrent effect of § 5, which dissuades covered jurisdictions from engaging in discrimination that they would resume should § 5 be struck down. Under this theory, however, § 5 would be effectively immune from scrutiny; no matter how "clean" the record of covered jurisdictions, the argument could always be made that it was deterrence that accounted for the good behavior.

[36] The provisions of § 5 apply only to those jurisdictions singled out by § 4. We now consider whether that coverage formula is constitutional in light of current conditions.

III

A

[37] When upholding the constitutionality of the coverage formula in 1966, we concluded that it was "rational in both practice and theory." Katzenbach, 383 U.S., at 330. The formula looked to cause (discriminatory tests) and effect (low voter registration and turnout), and tailored the remedy (preclearance) to those jurisdictions exhibiting both.

[38] By 2009, however, we concluded that the "coverage formula raise[d] serious constitutional questions." As we explained, a statute's "current burdens" must be justified by "current needs," and any "disparate geographic coverage" must be "sufficiently related to the problem that it targets." The coverage formula met that test in 1965, but no longer does so.

[39] Coverage today is based on decades-old data and eradicated practices. The formula captures States by reference to literacy tests and low voter registration and turnout in the 1960s and early 1970s. But such tests have been banned nationwide for over 40 years. And voter registration and turnout numbers in the covered States have risen dramatically in the years since. Racial disparity in those numbers was compelling evidence justifying the preclearance remedy and the coverage formula. There is no longer such a disparity.

[40] In 1965, the States could be divided into two groups: those with a recent history of voting tests and low voter registration and turnout, and those without those characteristics. Congress based its coverage formula on that distinction. Today the Nation is no longer divided along those lines, yet the Voting Rights Act continues to treat it as if it were.

B

[41] The Government's defense of the formula is limited. First, the Government contends that the formula is "reverse-engineered": Congress identified the jurisdictions to be covered and then came up with criteria to describe them. Under that reasoning, there need not be any logical relationship between the criteria in the formula and the reason for coverage; all that is necessary is that the formula happen to capture the jurisdictions Congress wanted to single out.

[ * * * ]

[42] The Fifteenth Amendment commands that the right to vote shall not be denied or abridged on account of race or color, and it gives Congress the power to enforce that command. The Amendment is not designed to punish for the past; its purpose is to ensure a better future. To serve that purpose, Congress — if it is to divide the States — must identify those jurisdictions to be singled out on a basis that makes sense in light of current conditions. It cannot rely simply on the past. We made that clear in Northwest Austin, and we make it clear again today.

[ * * * ]

D

[43] The dissent proceeds from a flawed premise. It quotes the famous sentence from McCulloch v. Maryland, 4 Wheat. 316 (1819), with the following emphasis: "Let the end be legitimate, let it be within the scope of the constitution, and all means which are appropriate, which are plainly adapted to that end, which are not prohibited, but consist with the letter and spirit of the constitution, are constitutional." But this case is about a part of the sentence that the dissent does not emphasize — the part that asks whether a legislative means is "consist[ent] with the letter and spirit of the constitution." The dissent states that "[i]t cannot tenably be maintained" that this is an issue with regard to the Voting Rights Act, but four years ago, in an opinion joined by two of today's dissenters, the Court expressly stated that "[t]he Act's preclearance requirement and its coverage formula raise serious constitutional questions." The dissent does not explain how those "serious constitutional questions" became untenable in four short years.

[***]

[44] There is no valid reason to insulate the coverage formula from review merely because it was previously enacted 40 years ago. If Congress had started from scratch in 2006, it plainly could not have enacted the present coverage formula. It would have been irrational for Congress to distinguish between States in such a fundamental way based on 40-year-old data, when today's statistics tell an entirely different story. And it would have been irrational to base coverage on the use of voting tests 40 years ago, when such tests have been illegal since that time. But that is exactly what Congress has done.

* * *

[45] Striking down an Act of Congress "is the gravest and most delicate duty that this Court is called on to perform." We do not do so lightly. That is why, in 2009, we took care to avoid ruling on the constitutionality of the Voting Rights Act when asked to do so, and instead resolved the case then before us on statutory grounds. But in issuing that decision, we expressed our broader concerns about the constitutionality of the Act. Congress could have updated the coverage formula at that time, but did not do so. Its failure to act leaves us today with no choice but to declare § 4(b) unconstitutional. The formula in that section can no longer be used as a basis for subjecting jurisdictions to preclearance.

[46] Our decision in no way affects the permanent, nationwide ban on racial discrimination in voting found in § 2. We issue no holding on § 5 itself, only on the coverage formula. Congress may draft another formula based on current conditions. Such a formula is an initial prerequisite to a determination that exceptional conditions still exist justifying such an "extraordinary departure from the traditional course of relations between the States and the Federal Government." Our country has changed, and while any racial discrimination in voting is too much, Congress must ensure that the legislation it passes to remedy that problem speaks to current conditions.

[47] The judgment of the Court of Appeals is reversed.

[48] It is so ordered.

Justice THOMAS, concurring.

[1] I join the Court's opinion in full but write separately to explain that I would find § 5 of the Voting Rights Act unconstitutional as well.

[* * * ]

Justice GINSBURG, with whom Justice BREYER, Justice SOTOMAYOR, and Justice KAGAN join, dissenting.

[1] In the Court's view, the very success of § 5 of the Voting Rights Act demands its dormancy. Congress was of another mind. Recognizing that large progress has been made, Congress determined, based on a voluminous record, that the scourge of discrimination was not yet extirpated. The question this case presents is who decides whether, as currently operative, § 5 remains justifiable, this Court, or a Congress charged with the obligation to enforce the post-Civil War Amendments "by appropriate legislation." With overwhelming support in both Houses, Congress concluded that, for two prime reasons, § 5 should continue in force, unabated. First, continuance would facilitate completion of the impressive gains thus far made; and second, continuance would guard against backsliding. Those assessments were well within Congress' province to make and should elicit this Court's unstinting approbation.

I

[2] "[V]oting discrimination still exists; no one doubts that." But the Court today terminates the remedy that proved to be best suited to block that discrimination. The Voting Rights Act of 1965 (VRA) has worked to combat voting discrimination where other remedies had been tried and failed. Particularly effective is the VRA's requirement of federal preclearance for all changes to voting laws in the regions of the country with the most aggravated records of rank discrimination against minority voting rights.

[3] A century after the Fourteenth and Fifteenth Amendments guaranteed citizens the right to vote free of discrimination on the basis of race, the "blight of racial discrimination in voting" continued to "infec[t] the electoral process in parts of our country." Early attempts to cope with this vile infection resembled battling the Hydra. Whenever one form of voting discrimination was identified and prohibited, others sprang up in its place. This Court repeatedly encountered the remarkable "variety and persistence" of laws disenfranchising minority citizens. To take just one example, the Court, in 1927, held unconstitutional a Texas law barring black voters from participating in primary elections, the Court struck down a "reenacted" and slightly altered version of the same law, and in 1953, the Court once again confronted an attempt by Texas to "circumven[t]" the Fifteenth Amendment by adopting yet another variant of the all-white primary.

[4] During this era, the Court recognized that discrimination against minority voters was a quintessentially political problem requiring a political solution. As Justice Holmes explained: If "the great mass of the white population intends to keep the blacks from voting," "relief from [that] great political wrong, if done, as alleged, by the people of a State and the State itself, must be given by them or by the legislative and political department of the government of the United States."

[5] Congress learned from experience that laws targeting particular electoral practices or enabling case-by-case litigation were inadequate to the task. In the Civil Rights Acts of 1957, 1960, and 1964, Congress authorized and then expanded the power of "the Attorney General to seek injunctions against public and private interference with the right to vote on racial grounds." Katzenbach, 383 U.S., at 313. But circumstances reduced the ameliorative potential of these legislative Acts:

[6] "Voting suits are unusually onerous to prepare, sometimes requiring as many as 6,000 man-hours spent combing through registration records in preparation for trial. Litigation has been exceedingly slow, in part because of the ample opportunities for delay afforded voting officials and others involved in the proceedings. Even when favorable decisions have finally been obtained, some of the States affected have merely switched to discriminatory devices not covered by the federal decrees or have enacted difficult new tests designed to prolong the existing disparity between white and Negro registration. Alternatively, certain local officials have defied and evaded court orders or have simply closed their registration offices to freeze the voting rolls." Id.

[7] Patently, a new approach was needed.

[8] Answering that need, the Voting Rights Act became one of the most consequential, efficacious, and amply justified exercises of federal legislative power in our Nation's history. Requiring federal preclearance of changes in voting laws in the covered jurisdictions — those States and localities where opposition to the Constitution's commands were most virulent — the VRA provided a fit solution for minority voters as well as for States. Under the preclearance regime established by § 5 of the VRA, covered jurisdictions must submit proposed changes in voting laws or procedures to the Department of Justice (DOJ), which has 60 days to respond to the changes. A change will be approved unless DOJ finds it has "the purpose [or] ... the effect of denying or abridging the right to vote on account of race or color." In the alternative, the covered jurisdiction may seek approval by a three-judge District Court in the District of Columbia.

[9] After a century's failure to fulfill the promise of the Fourteenth and Fifteenth Amendments, passage of the VRA finally led to signal improvement on this front. "The Justice Department estimated that in the five years after [the VRA's] passage, almost as many blacks registered [to vote] in Alabama, Mississippi, Georgia, Louisiana, North Carolina, and South Carolina as in the entire century before 1965." And in assessing the overall effects of the VRA in 2006, Congress found that "[s]ignificant progress has been made in eliminating first generation barriers experienced by minority voters, including increased numbers of registered minority voters, minority voter turnout, and minority representation in Congress, State legislatures, and local elected offices. This progress is the direct result of the Voting Rights Act of 1965. On that matter of cause and effects there can be no genuine doubt.

[10] Although the VRA wrought dramatic changes in the realization of minority voting rights, the Act, to date, surely has not eliminated all vestiges of discrimination against the exercise of the franchise by minority citizens. Jurisdictions covered by the preclearance requirement continued to submit, in large numbers, proposed changes to voting laws that the Attorney General declined to approve, auguring that barriers to minority voting would quickly resurface were the preclearance remedy eliminated. Congress also found that as "registration and voting of minority citizens increas[ed], other measures may be resorted to which would dilute increasing minority voting strength." Efforts to reduce the impact of minority votes, in contrast to direct attempts to block access to the ballot, are aptly described as "second-generation barriers" to minority voting.

[11] Second-generation barriers come in various forms. One of the blockages is racial gerrymandering, the redrawing of legislative districts in an "effort to segregate the races for purposes of voting." Another is adoption of a system of at-large voting in lieu of district-by-district voting in a city with a sizable black minority. By switching to at-large voting, the overall majority could control the election of each city council member, effectively eliminating the potency of the minority's votes.

[ * * * ]

[12] After considering the full legislative record, Congress made the following findings: The VRA has directly caused significant progress in eliminating first-generation barriers to ballot access, leading to a marked increase in minority voter registration and turnout and the number of minority elected officials. But despite this progress, "second generation barriers constructed to prevent minority voters from fully participating in the electoral process" continued to exist, as well as racially polarized voting in the covered jurisdictions, which increased the political vulnerability of racial and language minorities in those jurisdictions. Extensive "[e]vidence of continued discrimination," Congress concluded, "clearly show[ed] the continued need for Federal oversight" in covered jurisdictions. The overall record demonstrated to the federal lawmakers that, "without the continuation of the Voting Rights Act of 1965 protections, racial and language minority citizens will be deprived of the opportunity to exercise their right to vote, or will have their votes diluted, undermining the significant gains made by minorities in the last 40 years."

[13] Based on these findings, Congress reauthorized preclearance for another 25 years, while also undertaking to reconsider the extension after 15 years to ensure that the provision was still necessary and effective. 42 U.S.C. § 1973b(a)(7), (8) (2006 ed., Supp. V). The question before the Court is whether Congress had the authority under the Constitution to act as it did.

II

[14] In answering this question, the Court does not write on a clean slate. It is well established that Congress' judgment regarding exercise of its power to enforce the Fourteenth and Fifteenth Amendments warrants substantial deference. The VRA addresses the combination of race discrimination and the right to vote, which is "preservative of all rights." When confronting the most constitutionally invidious form of discrimination, and the most fundamental right in our democratic system, Congress' power to act is at its height.

[15] The basis for this deference is firmly rooted in both constitutional text and precedent. The Fifteenth Amendment, which targets precisely and only racial discrimination in voting rights, states that, in this domain, "Congress shall have power to enforce this article by appropriate legislation." In choosing this language, the Amendment's framers invoked Chief Justice Marshall's formulation of the scope of Congress' powers under the Necessary and Proper Clause:

[16] "Let the end be legitimate, let it be within the scope of the constitution, and all means which are appropriate, which are plainly adapted to that end, which are not prohibited, but consist with the letter and spirit of the constitution, are constitutional." McCulloch v. Maryland, 4 Wheat. 316, 421 (1819).

[17] It cannot tenably be maintained that the VRA, an Act of Congress adopted to shield the right to vote from racial discrimination, is inconsistent with the letter or spirit of the Fifteenth Amendment, or any provision of the Constitution read in light of the Civil War Amendments. Nowhere in today's opinion, or in Northwest Austin, is there clear recognition of the transformative effect the Fifteenth Amendment aimed to achieve. Notably, "the Founders' first successful amendment told Congress that it could `make no law' over a certain domain"; in contrast, the Civil War Amendments used "language [that] authorized transformative new federal statutes to uproot all vestiges of unfreedom and inequality" and provided "sweeping enforcement powers ... to enact `appropriate' legislation targeting state abuses."

[18] The stated purpose of the Civil War Amendments was to arm Congress with the power and authority to protect all persons within the Nation from violations of their rights by the States. In exercising that power, then, Congress may use "all means which are appropriate, which are plainly adapted" to the constitutional ends declared by these Amendments. So when Congress acts to enforce the right to vote free from racial discrimination, we ask not whether Congress has chosen the means most wise, but whether Congress has rationally selected means appropriate to a legitimate end. "It is not for us to review the congressional resolution of [the need for its chosen remedy]. It is enough that we be able to perceive a basis upon which the Congress might resolve the conflict as it did."

[19] Until today, in considering the constitutionality of the VRA, the Court has accorded Congress the full measure of respect its judgments in this domain should garner. “As against the reserved powers of the States, Congress may use any rational means to effectuate the constitutional prohibition of racial discrimination in voting." Faced with subsequent reauthorizations of the VRA, the Court has reaffirmed this standard. Today's Court does not purport to alter settled precedent establishing that the dispositive question is whether Congress has employed "rational means."

[***]

[20] In summary, the Constitution vests broad power in Congress to protect the right to vote, and in particular to combat racial discrimination in voting. This Court has repeatedly reaffirmed Congress' prerogative to use any rational means in exercise of its power in this area. And both precedent and logic dictate that the rational-means test should be easier to satisfy, and the burden on the statute's challenger should be higher, when what is at issue is the reauthorization of a remedy that the Court has previously affirmed, and that Congress found, from contemporary evidence, to be working to advance the legislature's legitimate objective.

III

[21] The 2006 reauthorization of the Voting Rights Act fully satisfies the standard stated in McCulloch, 4 Wheat., at 421: Congress may choose any means "appropriate" and "plainly adapted to" a legitimate constitutional end. As we shall see, it is implausible to suggest otherwise.

 [ * * * ]

[22] Congress designed the VRA to be a dynamic statute, capable of adjusting to changing conditions. True, many covered jurisdictions have not been able to bail out due to recent acts of noncompliance with the VRA, but that truth reinforces the congressional judgment that these jurisdictions were rightfully subject to preclearance, and ought to remain under that regime.

IV

[23] Congress approached the 2006 reauthorization of the VRA with great care and seriousness. The same cannot be said of the Court's opinion today. The Court makes no genuine attempt to engage with the massive legislative record that Congress assembled. Instead, it relies on increases in voter registration and turnout as if that were the whole story. Without even identifying a standard of review, the Court dismissively brushes off arguments based on "data from the record," and declines to enter the "debat[e about] what [the] record shows." One would expect more from an opinion striking at the heart of the Nation's signal piece of civil-rights legislation.

[24] I note the most disturbing lapses. First, by what right, given its usual restraint, does the Court even address Shelby County's facial challenge to the VRA? Second, the Court veers away from controlling precedent regarding the "equal sovereignty" doctrine without even acknowledging that it is doing so. Third, hardly showing the respect ordinarily paid when Congress acts to implement the Civil War Amendments, and as just stressed, the Court does not even deign to grapple with the legislative record.

[ * * * ]

[25] This Court has consistently rejected constitutional challenges to legislation enacted pursuant to Congress' enforcement powers under the Civil War Amendments upon finding that the legislation was constitutional as applied to the particular set of circumstances before the Court. A similar approach is warranted here.

[ * * * ]

[26] Today's unprecedented extension of the equal sovereignty principle outside its proper domain — the admission of new States — is capable of much mischief. Federal statutes that treat States disparately are hardly novelties. Of gravest concern, Congress relied on our pathmarking Katzenbach decision in each reauthorization of the VRA. It had every reason to believe that the Act's limited geographical scope would weigh in favor of, not against, the Act's constitutionality. Congress could hardly have foreseen that the VRA's limited geographic reach would render the Act constitutionally suspect.

[27] In the Court's conception, it appears, defenders of the VRA could not prevail upon showing what the record overwhelmingly bears out, i.e., that there is a need for continuing the preclearance regime in covered States. In addition, the defenders would have to disprove the existence of a comparable need elsewhere. I am aware of no precedent for imposing such a double burden on defenders of legislation.

[ * * * ]

C

[28] The Court has time and again declined to upset legislation of this genre unless there was no or almost no evidence of unconstitutional action by States. No such claim can be made about the congressional record for the 2006 VRA reauthorization. Given a record replete with examples of denial or abridgment of a paramount federal right, the Court should have left the matter where it belongs: in Congress' bailiwick.

[29] Instead, the Court strikes § 4(b)'s coverage provision because, in its view, the provision is not based on "current conditions." It discounts, however, that one such condition was the preclearance remedy in place in the covered jurisdictions, a remedy Congress designed both to catch discrimination before it causes harm, and to guard against return to old ways. Volumes of evidence supported Congress' determination that the prospect of retrogression was real. Throwing out preclearance when it has worked and is continuing to work to stop discriminatory changes is like throwing away your umbrella in a rainstorm because you are not getting wet.

[30] But, the Court insists, the coverage formula is no good; it is based on "decades-old data and eradicated practices." Even if the legislative record shows, as engaging with it would reveal, that the formula accurately identifies the jurisdictions with the worst conditions of voting discrimination, that is of no moment, as the Court sees it. Congress, the Court decrees, must "star[t] from scratch." I do not see why that should be so.

[***]

[31] Consider once again the components of the record before Congress in 2006. The coverage provision identified a known list of places with an undisputed history of serious problems with racial discrimination in voting. Recent evidence relating to Alabama and its counties was there for all to see. There was extensive evidence that, due to the preclearance mechanism, conditions in the covered jurisdictions had notably improved. And there was evidence that preclearance was still having a substantial real-world effect, having stopped hundreds of discriminatory voting changes in the covered jurisdictions since the last reauthorization. In addition, there was evidence that racial polarization in voting was higher in covered jurisdictions than elsewhere, increasing the vulnerability of minority citizens in those jurisdictions. And countless witnesses, reports, and case studies documented continuing problems with voting discrimination in those jurisdictions. In light of this record, Congress had more than a reasonable basis to conclude that the existing coverage formula was not out of sync with conditions on the ground in covered areas. And certainly Shelby County was no candidate for release through the mechanism Congress provided.

[ * * * ]

[32] The sad irony of today's decision lies in its utter failure to grasp why the VRA has proven effective. The Court appears to believe that the VRA's success in eliminating the specific devices extant in 1965 means that preclearance is no longer needed. With that belief, and the argument derived from it, history repeats itself. The same assumption — that the problem could be solved when particular methods of voting discrimination are identified and eliminated — was indulged and proved wrong repeatedly prior to the VRA's enactment. Unlike prior statutes, which singled out particular tests or devices, the VRA is grounded in Congress' recognition of the "variety and persistence" of measures designed to impair minority voting rights. In truth, the evolution of voting discrimination into more subtle second-generation barriers is powerful evidence that a remedy as effective as preclearance remains vital to protect minority voting rights and prevent backsliding.

[33] Beyond question, the VRA is no ordinary legislation. It is extraordinary because Congress embarked on a mission long delayed and of extraordinary importance: to realize the purpose and promise of the Fifteenth Amendment. For a half century, a concerted effort has been made to end racial discrimination in voting. Thanks to the Voting Rights Act, progress once the subject of a dream has been achieved and continues to be made.

* * *

[34] For the reasons stated, I would affirm the judgment of the Court of Appeals.