Chief Justice Roberts wrote the opinion of the Court:
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, 309 (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. See Voting Rights Act of 1965, §4(a), 79Stat. 438.
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.” Northwest Austin Municipal Util. Dist. No. One v. Holder, 557 U. S. 193–204 (2009). 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, Re-ported Voting and Registration, by Sex, Race and His-panic Origin, for States (Nov. 2012) (Table 4b).
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.
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.”
“The first century of congressional enforcement of the Amendment, however, can only be regarded as a failure.” Id., at 197. 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. Katzenbach, 383 U. S., at 310. 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. Id., at 313–314.
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.” 79Stat. 437. 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.” 42 U. S. C. §1973(a). Both the Federal Government and individuals have sued to enforce §2, see, e.g., Johnson v. De Grandy, 512 U. S. 997 (1994) , and injunctive relief is available in appropriate cases to block voting laws from going into effect, see 42 U. S. C. §1973j(d). Section 2 is permanent, applies nationwide, and is not at issue in this case.
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. §4(b), 79Stat. 438. Such tests or devices included literacy and knowledge tests, good moral character requirements, the need for vouchers from registered voters, and the like. §4(c), id., at 438–439. A 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.” §4(a), id., at 438. 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. See 28 CFR pt. 51, App. (2012).
In those jurisdictions, §4 of the Act banned all such tests or devices. §4(a), 79Stat. 438. 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. Id., at 439. 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.” Ibid.
Sections 4 and 5 were intended to be temporary; they were set to expire after five years. See §4(a), id., at 438; Northwest Austin, supra, at 199. 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.” 383 U. S., at 308.
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. Voting Rights Act Amendments of 1970, §§3–4, 84Stat. 315. That swept in several counties in California, New Hampshire, and New York. See 28 CFR pt. 51, App. Congress also extended the ban in §4(a) on tests and devices nationwide. §6, 84Stat. 315.
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. Voting Rights Act Amendments of 1975, §§101, 202, 89Stat. 400, 401. 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. §203, id., at 401–402. 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. See 28 CFR pt. 51, App. 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. §§203, 206, 89Stat. 401, 402. Finally, Congress made the nationwide ban on tests and devices permanent. §102, id., at 400.
In 1982, Congress reauthorized the Act for 25 years, but did not alter its coverage formula. See Voting Rights Act Amendments, 96Stat. 131. 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. §2, id., at 131–133.
We upheld each of these reauthorizations against constitutional challenge. See Georgia v. United States, 411 U. S. 526 (1973) ; City of Rome v. United States, 446 U. S. 156 (1980) ; Lopez v. Monterey County, 525 U. S. 266 (1999) .
In 2006, Congress again reauthorized the Voting Rights Act for 25 years, again without change to its coverage formula. Fannie Lou Hamer, Rosa Parks, and Coretta Scott King Voting Rights Act Reauthorization and Amendments Act, 120Stat. 577. Congress also amended §5 to prohibit more conduct than before. §5, id., at 580–581; see Reno v. Bossier Parish School Bd., 528 U. S. 320, 341(2000) (Bossier II); Georgia v. Ashcroft, 539 U. S. 461, 479 (2003). 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.” 42 U. S. C. §§1973c(b)–(d).
Shortly after this reauthorization, a Texas utility district brought suit, seeking to bail out from the Act’s cover-age and, in the alternative, challenging the Act’s constitutionality. See Northwest Austin, 557 U. S., at 200–201. 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. ”Northwest Austin Municipal Util. Dist. No. One v. Mukasey, 573 F. Supp. 2d 221, 232 (DC 2008). The District Court also rejected the constitutional challenge. Id., at 283.
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.’ ” Northwest Austin, supra, at 205 (quoting Escambia County v. McMillan, 466 U. S. 48, 51 (1984) (per curiam)). 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. Northwest Austin, 557 U. S., at 207. In doing so we expressed serious doubts about the Act’s continued constitutionality.
We explained that §5 “imposes substantial federalism costs” and “differentiates between the States, despite our historic tradition that all the States enjoy equal sovereignty.” Id., at 202, 203 (internal quotation marks omitted). 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.” Id., at 202. Finally, we questioned whether the problems that §5 meant to address were still “concentrated in the jurisdictions singled out for preclearance.” Id., at 203.
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.
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. See App. 87a–92a. 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. 811 F. Supp. 2d 424, 508 (2011). The court found that the evidence before Congress in 2006 was sufficient to justify reauthorizing §5 and continuing the §4(b) coverage formula.
The Court of Appeals for the D. C. Circuit affirmed. In assessing §5, the D. C. Circuit considered six primary categories of evidence: Attorney General objections to voting changes, Attorney General requests for more information regarding voting changes, successful §2 suits in covered jurisdictions, the dispatching of federal observers to monitor elections in covered jurisdictions, §5 preclearance suits involving covered jurisdictions, and the deterrent effect of §5. See 679 F. 3d 848, 862–863 (2012). After extensive analysis of the record, the court accepted Congress’s conclusion that §2 litigation remained inadequate in the covered jurisdictions to protect the rights of minority voters, and that §5 was therefore still necessary. Id.,at 873.
Turning to §4, the D. C. Circuit noted that the evidence for singling out the covered jurisdictions was “less robust” and that the issue presented “a close question.” Id., at 879. But the court looked to data comparing the number of successful §2 suits in the different parts of the country. Coupling that evidence with the deterrent effect of §5, the court concluded that the statute continued “to single out the jurisdictions in which discrimination is concentrated,” and thus held that the coverage formula passed constitutional muster. Id., at 883.
Judge Williams dissented. He found “no positive correlation between inclusion in §4(b)’s coverage formula and low black registration or turnout.” Id., at 891. Rather, to the extent there was any correlation, it actually went the other way: “condemnation under §4(b) is a marker of higher black registration and turnout.” Ibid. (emphasis added). Judge Williams also found that “[c]overed jurisdictions have far more black officeholders as a proportion of the black population than do uncovered ones.” Id., at 892. As to the evidence of successful §2 suits, Judge Williams disaggregated the reported cases by State, and concluded that “[t]he five worst uncovered jurisdictions . . . have worse records than eight of the covered jurisdictions.” Id., at 897. He also noted that two covered jurisdictions—Arizona and Alaska—had not had any successful reported §2 suit brought against them during the entire 24 years covered by the data. Ibid. Judge Williams would have held the coverage formula of §4(b) “irrational” and unconstitutional. Id., at 885.
We granted certiorari. 568 U. S. ___ (2012).
In Northwest Austin, we stated that “the Act imposes current burdens and must be justified by current needs.” 557 U. S., at 203. 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.” Ibid. These basic principles guide our review of the question before us.1
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. See 1 Records of the Federal Convention of 1787, pp. 21, 164–168 (M. Farrand ed. 1911); 2 id., at 27–29, 390–392.
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.” Bond v. United States, 564 U. S. ___, ___ (2011) (slip op., at 9). 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.” Ibid. (internal quotation marks omitted).
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.’” Gregory v. Ashcroft, 501 U. S. 452–462 (1991) (quoting Sugarman v. Dougall, 413 U. S. 634, 647 (1973); some internal quotation marks omitted). 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. Art. I, §4, cl. 1; see also Arizona v. Inter Tribal Council of Ariz., Inc., ante, at 4–6. But States have “broad powers to determine the conditions under which the right of suffrage may be exercised.” Carrington v. Rash, 380 U. S. 89, 91 (1965) (internal quotation marks omitted); see also Arizona, ante, at 13–15. And “[e]ach State has the power to prescribe the qualifications of its officers and the manner in which they shall be chosen.” Boyd v. Nebraska ex rel. Thayer, 143 U. S. 135, 161 (1892). Drawing lines for congressional districts is likewise “primarily the duty and responsibility of the State.” Perry v. Perez, 565 U. S. ___, ___ (2012) (per curiam) (slip op., at 3) (internal quotation marks omitted).
Not only do States retain sovereignty under the Constitution, there is also a “fundamental principle of equal sovereignty” among the States. Northwest Austin, supra, at 203 (citing United States v. Louisiana, 363 U. S. 1, 16 (1960); Lessee of Pollard v. Hagan, 3 How. 212, 223 (1845); and Texas v. White, 7 Wall. 700, 725–726 (1869); emphasis added). Over a hundred years ago, this Court explained that our Nation “was and is a union of States, equal in power, dignity and authority.” Coyle v. Smith,221 U. S. 559, 567 (1911). Indeed, “the constitutional equality of the States is essential to the harmonious operation of the scheme upon which the Republic was organized.” Id., at 580. 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. 383 U. S., at 328–329. 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. 557 U. S., at 203.
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.” Id., at 202. 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. See 28 CFR §§51.9, 51.37. If a State seeks preclearance from a three-judge court, the process can take years.
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 noncovered 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.” 679 F. 3d, at 884 (Williams, J., dissenting) (case below).
All this explains why, when we first upheld the Act in 1966, we described it as “stringent” and “potent. ”Katzenbach, 383 U. S., at 308, 315, 337. 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.” Id., at 334. We have since noted that the Act “authorizes federal intrusion into sensitive areas of state and local policymaking,” Lopez, 525 U. S., at 282, and represents an “extraordinary departure from the traditional course of relations between the States and the Federal Government,” Presley v. Etowah County Comm’n, 502 U. S. 491–501 (1992). As we reiterated in Northwest Austin, the Act constitutes “extraordinary legislation otherwise unfamiliar to our federal system.” 557 U. S., at 211.
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.” Katzenbach, 383 U. S., at 308. Several States had enacted a variety of requirements and tests “specifically designed to prevent” African-Americans from voting. Id., at 310. Case-by-case litigation had proved inadequate to prevent such racial discrimination in voting, in part because States “merely switched to discriminatory devices not covered by the federal decrees,” “enacted difficult new tests,” or simply “defied and evaded court orders.” Id., at 314. Shortly before enactment of the Voting Rights Act, only 19.4 percent of African-Americans of voting age were registered to vote in Alabama, only 31.8 percent in Louisiana, and only 6.4 percent in Mississippi. Id., at 313. Those figures were roughly 50 percentage points or more below the figures for whites. Ibid.
In short, we concluded that “[u]nder the compulsion of these unique circumstances, Congress responded in a permissibly decisive manner.” Id., at 334, 335. We also noted then and have emphasized since that this extra-ordinary legislation was intended to be temporary, set to expire after five years. Id., at 333; Northwest Austin, supra, at 199.
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.” Katzenbach, 383 U. S., at 328. 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.” Id., at 330. We explained that “[t]ests and devices are relevant to voting discrimination because of their long history as a tool for perpetrating the evil; a low voting rate is pertinent for the obvious reason that widespread disenfranchisement must inevitably affect the number of actual voters.” Ibid. We therefore concluded that “the coverage formula [was] rational in both practice and theory.” Ibid. It accurately reflected those jurisdictions uniquely characterized by voting discrimination “on a pervasive scale,” linking coverage to the devices used to effectuate discrimination and to the resulting disenfranchisement. Id., at 308. The formula ensured that the “stringent remedies [were] aimed at areas where voting discrimination ha[d] been most flagrant.” Id., at 315.
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.” Northwest Austin, 557 U. S., at 202. The tests and devices that blocked access to the ballot have been forbidden nationwide for over 40 years. See §6,84Stat. 315; §102, 89Stat. 400.
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.” §2(b)(1),120Stat. 577. 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.” H. R. Rep. No. 109–478, p. 12 (2006). 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. Id., at 18.
The following chart, compiled from the Senate and House Reports, compares voter registration numbers from 1965 to those from 2004 in the six originally covered States. These are the numbers that were before Congress when it reauthorized the Act in 2006:
See S. Rep. No. 109–295, p. 11 (2006); H. R. Rep. No. 109–478, at 12. The 2004 figures come from the Census Bureau. 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). The preclearance statistics are also illuminating. In the first decade after enactment of §5, the Attorney General objected to 14.2 percent of proposed voting changes. H. R Rep. No. 109–478, at 22. In the last decade before reenactment, the Attorney General objected to a mere 0.16 percent. S. Rep. No. 109–295, at 13.
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. See §2(b)(1), 120Stat. 577. During the “Freedom Summer” of 1964, in Philadelphia, Mississippi, three men were murdered while working in the area to register African-American voters. See United States v. Price, 383 U. S. 787, 790 (1966). On “Bloody Sunday” in 1965, in Selma, Alabama, police beat and used tear gas against hundreds marching in sup-port of African-American enfranchisement. See Northwest Austin, supra, at 220, n. 3 (Thomas, J., concurring in judgment in part and dissenting in part). 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.
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. See 42 U. S. C. §1973b(a)(8). 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. See Bossier II, 528 U. S., at 324, 335–336. In 2006, Congress amended §5to prohibit laws that could have favored such groups but did not do so because of a discriminatory purpose, see 42 U. S. C. §1973c(c), 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,” Bossier II, supra, at 336 (citation and internal quotation marks omitted). In addition, Congress expanded §5 to prohibit any voting law “that has the purpose of or will have the effect of diminishing the ability of any citizens of the United States,” on account of race, color, or language minority status, “to elect their preferred candidates of choice.” §1973c(b). In light of those two amendments, the bar that covered jurisdictions must clear has been raised even as the conditions justifying that requirement have dramatically improved.
We have also previously highlighted the concern that “the preclearance requirements in one State [might]be unconstitutional in another.” Northwest Austin, 557 U. S., at 203; see Georgia v. Ashcroft, 539 U. S., at 491 (Kennedy, J., concurring) (“considerations of race that would doom a redistricting plan under the Fourteenth Amendment or §2 [of the Voting Rights Act] seem to be what save it under §5”). Nothing has happened since to alleviate this troubling concern about the current application of §5.
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.
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.
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.
By 2009, however, we concluded that the “coverage formula raise[d] serious constitutional questions. ” Northwest Austin, 557 U. S., at 204. 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.” Id., at 203. The coverage formula met that test in 1965, but no longer does so.
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. §6, 84Stat. 315; §102, 89Stat. 400. And voter registration and turnout numbers in the covered States have risen dramatically in the years since. H. R. Rep. No. 109–478, at 12. Racial disparity in those numbers was compelling evidence justifying the preclearance remedy and the coverage formula. See, e.g., Katzenbach, supra, at 313, 329–330. There is no longer such a disparity.
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.
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. Brief for Federal Respondent 48–49. Under that reasoning, there need not be any logical relationship be-tween 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.
The Government suggests that Katzenbach sanctioned such an approach, but the analysis in Katzenbach was quite different. Katzenbach reasoned that the coverage formula was rational because the “formula . . . was relevant to the problem”: “Tests and devices are relevant to voting discrimination because of their long history as a tool for perpetrating the evil; a low voting rate is pertinent for the obvious reason that widespread disenfranchisement must inevitably affect the number of actual voters.” 383 U. S., at 329, 330.
Here, by contrast, the Government’s reverse-engineering argument does not even attempt to demonstrate the continued relevance of the formula to the problem it targets. And in the context of a decision as significant as this one—subjecting a disfavored subset of States to “extraordinary legislation otherwise unfamiliar to our federal system,” Northwest Austin, supra, at 211—that failure to establish even relevance is fatal.
The Government falls back to the argument that because the formula was relevant in 1965, its continued use is permissible so long as any discrimination remains in the States Congress identified back then—regardless of how that discrimination compares to discrimination in States unburdened by coverage. Brief for Federal Respondent 49–50. This argument does not look to “current political conditions,” Northwest Austin, supra, at 203, but instead relies on a comparison between the States in 1965. That comparison reflected the different histories of the North and South. It was in the South that slavery was upheld by law until uprooted by the Civil War, that the reign of Jim Crow denied African-Americans the most basic freedoms, and that state and local governments worked tirelessly to disenfranchise citizens on the basis of race. The Court invoked that history—rightly so—in sustaining the disparate coverage of the Voting Rights Act in 1966. See Katzenbach, supra, at 308 (“The constitutional propriety of the Voting Rights Act of 1965 must be judged with reference to the historical experience which it reflects.”).
But history did not end in 1965. By the time the Act was reauthorized in 2006, there had been 40 more years of it. In assessing the “current need[ ]” for a preclearance system that treats States differently from one another today, that history cannot be ignored. During that time, largely because of the Voting Rights Act, voting tests were abolished, disparities in voter registration and turnout due to race were erased, and African-Americans attained political office in record numbers. And yet the coverage formula that Congress reauthorized in 2006 ignores these developments, keeping the focus on decades-old data relevant to decades-old problems, rather than current data reflecting current needs.
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. See Rice v. Cayetano, 528 U. S. 495, 512 (2000) (“Consistent with the design of the Constitution, the [Fifteenth] Amendment is cast in fundamental terms, terms transcending the particular controversy which was the immediate impetus for its enactment.”). 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.
In defending the coverage formula, the Government, the intervenors, and the dissent also rely heavily on data from the record that they claim justify disparate coverage. Congress compiled thousands of pages of evidence before reauthorizing the Voting Rights Act. The court below and the parties have debated what that record shows—they have gone back and forth about whether to compare covered to noncovered jurisdictions as blocks, how to disaggregate the data State by State, how to weigh §2 cases as evidence of ongoing discrimination, and whether to consider evidence not before Congress, among other issues. Compare, e.g., 679 F. 3d, at 873–883 (case below), with id., at 889–902 (Williams, J., dissenting). Regardless of how to look at the record, however, no one can fairly say that it shows anything approaching the “pervasive,” “flagrant,” “widespread,” and “rampant” discrimination that faced Congress in 1965, and that clearly distinguished the covered jurisdictions from the rest of the Nation at that time. Katzenbach, supra, at 308, 315, 331; Northwest Austin, 557 U. S., at 201.
But a more fundamental problem remains: Congress did not use the record it compiled to shape a coverage formula grounded in current conditions. It instead reenacted a formula based on 40-year-old facts having no logical relation to the present day. The dissent relies on “second-generation barriers,” which are not impediments to the casting of ballots, but rather electoral arrangements that affect the weight of minority votes. That does not cure the problem. Viewing the preclearance requirements as targeting such efforts simply highlights the irrationality of continued reliance on the §4 coverage formula, which is based on voting tests and access to the ballot, not vote dilution. We cannot pretend that we are reviewing an updated statute, or try our hand at updating the statute ourselves, based on the new record compiled by Congress. Contrary to the dissent’s contention, see post, at 23, we are not ignoring the record; we are simply recognizing that it played no role in shaping the statutory formula before us today.
The dissent also turns to the record to argue that, in light of voting discrimination in Shelby County, the county cannot complain about the provisions that subject it to preclearance. Post, at 23–30. But that is like saying that a driver pulled over pursuant to a policy of stopping all redheads cannot complain about that policy, if it turns out his license has expired. Shelby County’s claim is that the coverage formula here is unconstitutional in all its applications, because of how it selects the jurisdictions subjected to preclearance. The county was selected based on that formula, and may challenge it in court.
The dissent proceeds from a flawed premise. It quotes the famous sentence from McCulloch v. Maryland, 4 Wheat. 316, 421 (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.” Post, at 9 (emphasis in dissent). 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, post, at 9, 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.” Northwest Austin, supra, at 204. The dissent does not explain how those “serious constitutional questions” became untenable in four short years.
The dissent treats the Act as if it were just like any other piece of legislation, but this Court has made clear from the beginning that the Voting Rights Act is far from ordinary. At the risk of repetition, Katzenbach indicated that the Act was “uncommon” and “not otherwise appropriate,” but was justified by “exceptional” and “unique” conditions. 383 U. S., at 334, 335. Multiple decisions since have reaffirmed the Act’s “extraordinary” nature. See, e.g., Northwest Austin, supra, at 211. Yet the dissent goes so far as to suggest instead that the preclearance requirement and disparate treatment of the States should be upheld into the future “unless there [is] no or almost no evidence of unconstitutional action by States.” Post, at 33.
In other ways as well, the dissent analyzes the question presented as if our decision in Northwest Austin never happened. For example, the dissent refuses to con-sider the principle of equal sovereignty, despite Northwest Austin’s emphasis on its significance. Northwest Austin also emphasized the “dramatic” progress since 1965, 557 U. S., at 201, but the dissent describes current levels of discrimination as “flagrant,” “widespread,” and “pervasive,” post, at 7, 17 (internal quotation marks omitted). Despite the fact that Northwest Austin requires an Act’s “disparate geographic coverage” to be “sufficiently related” to its targeted problems, 557 U. S., at 203, the dissent maintains that an Act’s limited coverage actually eases Congress’s burdens, and suggests that a fortuitous relationship should suffice. Although Northwest Austin stated definitively that “current burdens” must be justified by “current needs,” ibid., the dissent argues that the coverage formula can be justified by history, and that the required showing can be weaker on reenactment than when the law was first passed.
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.
Striking down an Act of Congress “is the gravest and most delicate duty that this Court is called on to perform.” Blodgett v. Holden, 275 U. S. 142, 148 (1927) (Holmes, J., concurring). 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.
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.” Presley, 502 U. S., at 500–501. 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.
The judgment of the Court of Appeals is reversed.
It is so ordered.