dissenting.
We have only today held that the city of Mobile does not violate the Constitution by maintaining an at-large system of electing city officials unless voters can prove that system is a product of purposeful discrimination. City of Mobile v. Bolden, ante, p. 55. This result is reached even though the black residents of Mobile have demonstrated that racial “bloc” voting has prevented them from electing a black representative to the city government. The Court correctly concluded that a city has no obligation under the Constitution *207to structure its representative system in a manner that maximizes the black community’s ability to elect a black representative. Yet in the instant case, the city of Rome is prevented from instituting precisely the type of structural changes which the Court says Mobile may maintain consistently with the Civil War Amendments, so long as their purpose be legitimate, because Congress has prohibited these changes under the Voting Rights Act as an exercise of its “enforcement” power conferred by those Amendments.
It is not necessary to hold that Congress is limited to merely providing a forum in which aggrieved plaintiffs may assert rights under the Civil War Amendments in order to disagree with the Court’s decision permitting Congress to straitjacket the city of Rome in this manner. Under § 5 of the Fourteenth Amendment and § 2 of the Fifteenth Amendment, Congress is granted only the power to “enforce” by “appropriate” legislation the limitations on state action embodied in those Amendments. While the presumption of constitutionality is due to any act of a coordinate branch of the Federal Government or of one of the States, it is this Court which is ultimately responsible for deciding challenges to the exercise of power by those entities. Marbury v. Madison, 1 Cranch 137 (1803); United States v. Nixon, 418 U. S. 683 (1974). Today’s decision is nothing less than a total abdication of that authority, rather than an exercise of the deference due to a coordinate branch of the government.
I
The facts of this case readily demonstrate the fallacy underlying the Court’s determination that congressional prohibition of Rome’s conduct can be characterized as enforcement of the Fourteenth or Fifteenth Amendment.1 The *208three-judge District Court entered extensive findings of fact— facts which are conspicuously absent from the Court's opinion. The lower court found that Rome has not employed any discriminatory barriers to black voter registration in the past 17 years. Nor has the city employed any other barriers to black voting or black candidacy. Indeed, the court found that white elected officials have encouraged blacks to run for elective posts in Rome, and are “responsive to the needs and interests of the black community.” The city has not discriminated against blacks in the provision of services and has made efforts to upgrade black neighborhoods.
It was also established that although a black has never been elected to political office in Rome, a black was appointed to fill a vacancy in an elective post. White candidates vigorously pursue the support of black voters. Several commissioners testified that they spent proportionately more time campaigning in the black community because they “needed that vote to win.” The court concluded that “blacks often hold the balance of power in Rome elections.”
Despite this political climate, the Attorney General refused to approve a number of city annexations and various changes in the electoral process. The city sought to require majority vote for election to the City Commission and Board of Education ; to create numbered posts and staggered terms for those elections; and to establish a ward residency requirement for Board of Education elections. In addition, during the years *209between 1964 and 1973, the city effected 60 annexations. Appellees concede that none of the annexations were sought for discriminatory purposes. All of the electoral changes and 13 of the annexations were opposed by the Attorney General on the grounds that their adoption would lessen the likelihood that blacks would be successful in electing a black city official, assuming racial-bloc voting on the part of both whites and blacks. Each of the changes was considered to be an impermissible “vote-dilution” device.
Rome sought judicial relief and the District Court found that the city had met its burden of proving that these electoral changes and annexations were not enacted with the purpose of discriminating against blacks. The changes were nevertheless prohibited because of their perceived disparate effect.2
II
The Court holds today that the city of Rome can constitutionally be compelled to seek congressional approval for most of its governmental changes even though it has not engaged in any discrimination against blacks for at least 17 years. Moreover, the Court also holds that federal approval can be constitutionally denied even after the city has proved that the changes are not purposefully discriminatory. While I agree with Mr. Justice Powell’s conclusion that requiring localities to submit to preclearance is a significant intrusion on local autonomy, it is an even greater intrusion on that autonomy to deny preclearance sought.
The facts of this case signal the necessity for this Court to carefully scrutinize the alleged source of congressional power to intrude so deeply in the governmental structure of the municipal corporations created by some of the 50 States. Section 2 of the Fifteenth Amendment and § 5 of the Four*210teenth provide that Congress shall have the power to “enforce” § 1 “by appropriate legislation.” Congressional power to prohibit the electoral changes proposed by Rome is dependent upon the scope and nature of that power. There are three theories of congressional enforcement power relevant to this case. First, it is clear that if the proposed changes would violate the Constitution, Congress could certainly prohibit their implementation. It has never been seriously maintained, however, that Congress can do no more than the judiciary to enforce the Amendments' commands. Thus, if the electoral changes in issue do not violate the Constitution, as judicially interpreted, it must be determined whether Congress could nevertheless appropriately prohibit these changes under the other two theories of congressional power. Under the second theory, Congress can act remedially to enforce the judicially established substantive prohibitions of the Amendments. If not properly remedial, the exercise of this power could be sustained only if this Court accepts the premise of the third theory that Congress has the authority under its enforcement powers to determine, without more, that electoral changes with a disparate impact on race violate the Constitution, in which case Congress by a legislative Act could effectively amend the Constitution.
I think it is apparent that neither of the first two theories for sustaining the exercise of congressional power supports this application of the Voting Rights Act. After our decision in City of Mobile there is little doubt that Rome has not engaged in constitutionally prohibited conduct.3 I also do not *211believe that prohibition of these changes can genuinely be characterized as a remedial exercise of congressional enforcement powers. Thus, the result of the Court’s holding is that Congress effectively has the power to determine for itself that this conduct violates the Constitution. This result violates previously well-established distinctions between the Judicial Branch and the Legislative or Executive Branches of the Federal Government. See United States v. Nixon, 418 U. S. 683 (1974); Marbury v. Madison, 1 Cranch 137 (1803).
A
If the enforcement power is construed as a “remedial” grant of authority, it is this Court’s duty to ensure that a challenged congressional Act does no more than “enforce” the limitations on state power established in the Fourteenth and Fifteenth Amendments. Marbury v. Madison. The Court has not resolved the question of whether it is an appropriate exercise of remedial power for Congress to prohibit local governments from instituting structural changes in their government, which although not racially motivated, will have the effect of decreasing the ability of a black voting bloc to elect a black candidate.
This Court has found, as a matter of statutory interpretation, that Congress intended to prohibit governmental changes on the basis of no more than disparate impact under the Voting Rights Act. These cases, however, have never directly presented the constitutional questions implicated by the lower court finding in this case that the city has engaged in no purposeful discrimination in enacting these changes, or otherwise, for almost two decades. See Beer v. United States, 425 U. S. 130 (1976); City of Richmond v. United States, 422 U. S. 358 (1975); Perkins v. Matthews, 400 U. S. 379 (1971); Fairley v. Patterson, decided together with Allen v. State Board of Elections, 393 U. S. 544 (1969). In none of these cases was the Court squarely presented with a constitutional challenge to congressional power to prohibit state electoral *212practices after the locality has disproved the existence of any purposeful discrimination.4
The cases in which this Court has actually examined the constitutional questions relating to Congress’ exercise of its powers to enforce the Fourteenth and Fifteenth Amendments also did not purport to resolve this issue.5 But the principles which can be distilled from those precedents require the conclusion that the limitations on state power at issue cannot be sustained as a remedial exercise of power. *213While the Fourteenth and Fifteenth Amendments prohibit only purposeful discrimination, the decisions of this Court have recognized that in some circumstances, congressional prohibition of state or local action which is not purposefully discriminatory may nevertheless be appropriate remedial legislation under the Civil War Amendments. See Oregon v. Mitchell, 400 U. S. 112 (1970); Gaston County v. United States, 395 U. S. 285 (1969).
Those circumstances, however, are not without judicial limits. These decisions indicate that congressional prohibition of some conduct which may not itself violate the Constitution is “appropriate” legislation “to enforce” the Civil War Amendments if that prohibition is necessary to remedy prior constitutional violations by the governmental unit, or if necessary to effectively prevent purposeful discrimination by a governmental unit. In both circumstances, Congress would still be legislating in response to the incidence of state action violative of the Civil War Amendments. These precedents are carefully formulated around a historic tenet of the law that in order to invoke a remedy, there must be a wrong— and under a remedial construction of congressional power to enforce the Fourteenth and Fifteenth Amendments, that wrong must amount to a constitutional violation. Only when the wrong is identified can the appropriateness of the remedy be measured.
The Court today identifies the constitutional wrong which was the object of this congressional exercise of power as purposeful discrimination by local governments in structuring their political processes in an effort to reduce black voting strength. The Court goes on to hold that the prohibitions imposed in this case represent an “appropriate” means of preventing such constitutional violations. The Court does not rest this conclusión on any finding that this prohibition is necessary to remedy any prior discrimination by the locality. Rather, the Court reasons that prohibition of changes dis*214criminatory in effect prevent the incidence of changes which are discriminatory in purpose:
“Congress could rationally have concluded that, because electoral changes by jurisdictions with a demonstrable history of intentional racial discrimination in voting create the risk of purposeful discrimination, it was proper to prohibit changes that have a discriminatory impact." Ante, at 177.
What the Court explicitly ignores is that in this case the city has proved that these changes are not discriminatory in purpose. Neither reason nor precedent supports the conclusion that here it is “appropriate” for Congress to attempt to prevent purposeful discrimination by prohibiting conduct which a locality proves is not purposeful discrimination.
Congress had before it evidence that various governments were énacting electoral changes and annexing territory to prevent the participation of blacks in local government by measures other than outright denial of the franchise.6 Congress could of course remedy and prevent such purposeful discrimination on the part of local governments. See Gomillion v. Lightfoot, 364 U. S. 339, 347 (1960). And given the difficulties of proving that an electoral change or annexation has been undertaken for the purpose of discriminating against blacks, Congress could properly conclude that as a remedial matter it was necessary to place the burden of proving lack of discriminatory purpose on the localities. See South Carolina v. Katzenbach, 383 U. S. 301 (1966). But all of this does not support the conclusion that Congress is acting remedially when it continues the presumption of purposeful discrimination even after the locality has disproved that presumption. Absent other circumstances, it would be a topsy-turvy judicial system which held that electoral changes *215which have been affirmatively proved to be permissible under the Constitution nonetheless violate the Constitution.
The precedent on which the Court relies simply does not support its remedial characterization. Neither Oregon v. Mitchell, 400 U. S. 112 (1970), nor South Carolina v. Katzenbach, supra, legitimizes the use of an irrebuttable presumption that “vote-diluting” changes are motivated by a discriminatory animus. The principal electoral practice in issue in those cases was the use of literacy tests. Yet, the Court simply fails to make any inquiry as to whether the particular electoral practices in issue here are encompassed by the “preventive” remedial rationale invoked in South Carolina and Oregon. The rationale does support congressional prohibition of some electoral practices, but simply has no logical application to the “vote-dilution” devices in issue.
In Oregon, the Court sustained a nationwide prohibition of literacy tests, thereby extending the more limited suspension approved in South Carolina. By upholding this congressional measure, the Court established that under some circumstances, a congressional remedy may be constitutionally overinclusive by prohibiting some state action which might not be purposefully discriminatory. That possibility does not justify the overinclusiveness countenanced by the Court in this case, however. Oregon by no means held that Congress could simply use discriminatory effect as a proxy for discriminatory purpose, as the Court seems to imply. Instead, the Court opinions identified the factors which rendered this prohibition properly remedial. The Court found the nationwide ban to be an appropriate means of effectively preventing purposeful discrimination in the application of the literacy tests as well as an appropriate means of remedying prior constitutional violations by state and local governments in the administration of education to minorities.
The presumption that the literacy tests were either being used to purposefully discriminate, or that the disparate effects of those tests were attributable to discrimination in state-*216administered education was not very wide of the mark. Various opinions of the Court noted that at the time that Congress enacted the ban, few States were utilizing literacy tests, 400 U. S., at 147 (opinion of Douglas, J.), and the voter registration statistics available within those States suggested that a disparate effect was prevalent. Id., at 132-133 (opinion of Black, J.). Even if not adopted with a discriminatory purpose, the tests could readily be applied in a discriminatory fashion. Thus a demonstration by the State that it sought to reinstate the tests for legitimate purposes did not eliminate the substantial risk of discrimination in application. Only a ban could effectively prevent the occurrence of purposeful discrimination.
The nationwide ban was also found necessary to effectively remedy past constitutional violations. Without the nationwide ban, a voter who was illiterate due to state discrimination in education could be denied the right to vote on the basis of his illiteracy when he moved into a jurisdiction retaining a literacy test for nondiscriminatory purposes. Id., at 283-284. Finally, Mr. Justice Stewart found that a uniform prohibition had definite advantages for enforcement and federal relations: it reduced tensions with particular regions, and it relieved the Federal Government from the administrative burden implicated by selective state enforcement.
Presumptive prohibition of vote-diluting procedures is not similarly an “appropriate” means of exacting state compliance with the Civil War Amendments. First, these prohibitions are quite unlike the literacy ban, where the disparate effects were traceable to the discrimination of governmental bodies in education even if their present desire to use the tests was legitimate. See Gaston County v. United States, 395 U. S. 285 (1969). Any disparate impact associated with the nondiscriminatory electoral changes in issue here results from bloc voting — private rather than governmental discrimi*217nation. It is clear therefore that these prohibitions do not implicate congressional power to devise an effective remedy for prior constitutional violations by local governments. Nor does the Court invoke this aspect of congressional remedial powers.
It is also clear that while most States still utilizing literacy tests may have been doing so to discriminate, a similar generalization could not be made about all government structures which have some disparate impact on black voting strength. At the time Congress passed the Act, one study demonstrated that 60% of all cities nationwide had at-large elections for city officials, for example. This form of government was adopted by many cities throughout this century as a reform measure designed to overcome wide-scale corruption in the ward system of government. See Jewell, Local Systems of Representation: Political Consequences and Judicial Choices, 36 Geo. Wash. L. Rev. 790, 799 (1967). Obviously, annexations similarly cannot be presumed to be devoid of legitimate uses. Yet both of these practices are regularly prohibited by the Act in most covered cities.
Nor does the prohibition of all practices with a disparate impact enhance congressional prevention of purposeful discrimination. The changes in issue are not, like literacy tests, though fair on their face, subject to discriminatory application by local authorities. See Yick Wo v. Hopkins, 118 U. S. 356 (1886). They are either discriminatory from the outset or not.
Finally, the advantages supporting the imposition of a nationwide ban are simply not implicated in this case. No added administrative burdens are in issue since Congress has provided the mechanism for preclearance suits in any event, and the burden of proof for this issue is on the locality. And it is certain that the only constitutional wrong implicated— purposeful dilution — can be effectively remedied by prohibiting it where it occurs. For all these reasons, I do not think *218that the present case is controlled by the result in Oregon. By prohibiting all electoral changes with a disparate impact, Congress has attempted to prevent disparate impacts — not purposeful discrimination.
Congress unquestionably has the power to prohibit and remedy state action which intentionally deprives citizens of Fourteenth and Fifteenth. Amendment rights. But unless these powers are to be wholly uncanalized, it cannot be appropriate remedial legislation for Congress to prohibit Rome from structuring its government in the manner as its population sees fit absent a finding or unrebutted presumption that Rome has been, or is, intentionally discriminating against its black citizens. Rome has simply committed no constitutional violations, as this Court has defined them.
More is at stake than sophistry at its worst in the Court’s conclusion that requiring the local government to structure its political system in a manner that most effectively enhances black political strength serves to remedy or prevent constitutional wrongs on the part of the local government. The need to prevent this disparate impact is premised on the assumption that white candidates will not represent black interests, and that States should devise a system encouraging blacks to vote in a bloc for black candidates. The findings in this case alone demonstrate the tenuous nature of these assumptions. The court below expressly found that white officials have ably represented the interests of the black community. Even blacks who testified admitted no dissatisfaction, but expressed only a preference to be represented by officials of their own race. The enforcement provisions of the Civil War Amendments were not premised on the notion that Congress could empower a later generation of blacks to “get even” for wrongs inflicted on their forebears. What is now at stake in the city of Rome is the preference of the black community to be represented by a black. This Court has never elevated such a notion, by no means confined to blacks, to the status of a constitutional right. See Whitcomb v. Chavis, *219403 U. S. 124 (1971). This Court concluded in Whitcomb that
“[t]he mere fact that one interest group or another concerned with the outcome of . . . elections has found itself outvoted and without legislative seats of its own provides no basis for invoking constitutional remedies where, as here, there is no indication that this segment of the population is being denied access to the political system.” Id., at 154-155.
The Constitution imposes no obligation on local governments to erect institutional safeguards to ensure the election of a black candidate. Nor do I believe that Congress can do so, absent a finding that this obligation would be necessary to remedy constitutional violations on the part of the local government.
It is appropriate to add that even if this Court could find a remedial relationship between the prohibition of all state action with a disparate impact on black voting strength and the incidence of purposeful discrimination, this Court should exercise caution in approving the remedy in issue here absent purposeful dilution. Political theorists can readily differ on the advantages inherent in different governmental structures. As Mr. Justice Harlan noted in his dissent in Fairley v. Patterson, decided together with Allen v. State Board of Elections, 393 U. S. 544 (1969): “[I]t is not clear to me how a court would go about deciding whether an at-large system is to be preferred over a district system. Under one system, Negroes have some influence in the election of all officers; under the other, minority groups have more influence in the selection of fewer officers.” Id., at 586 (emphasis deleted)'.
B
The result reached by the Court today can be sustained only upon the theory that Congress was empowered to determine that structural changes with a disparate impact on a minority group’s ability to elect a candidate of their race *220violates the Fourteenth or Fifteenth Amendment. This construction of the Fourteenth Amendment was rejected in the Civil Rights Cases, 109 U. S. 3 (1883). The Court emphasized that the power conferred was “remedial” only. The Court reasoned that the structure of the Amendment made it clear that it did not “authorize Congress to create a code of municipal law for the regulation of private rights; but to provide modes of redress against the operation of State laws, and the action of State officers . . . , when these are subversive of the fundamental rights specified in the [A]mendment.” Id., at 11. This interpretation is consonant with the legislative history surrounding the enactment of the Amendment.7
This construction has never been refuted by a majority of the Members of this Court. Support for this construction in current years has emerged in South Carolina v. Katzenbach, 383 U. S. 301 (1966), and Oregon v. Mitchell, 400 U. S. 112 (1970).8 See also opinion of Powell, J., ante, at 200-201. In South Carolina v. Katzenbach, the Court observed that Congress could not attack evils not comprehended by the Fifteenth Amendment. 383 U. S., at 326. In Oregon v. Mitchell, five Members of the Court were unwilling to conclude that Congress had the power to determine that estab*221lishing the age limitation for voting at 21 denied equal protection to those between the ages of 18 and 20.
The opinion of Mr. Justice Stewart in that case, joined by Mr. Chief Justice Burger and Mr. Justice Blackmun, reaffirmed that Congress only has the power under the Fourteenth Amendment to “provide the means of eradicating situations that amount to a violation of the Equal Protection Clause” but not to “determine as a matter of substantive constitutional law what situations fall within the ambit of the clause.” Id., at 296. Mr. Justice Harlan, in a separate opinion, reiterated his belief that it is the duty of the Court, and not the Congress, to determine when States have exceeded constitutional limitations imposed upon their powers. Id., at 204-207. Cf. Oregon v. Hass, 420 U. S. 714 (1975); Cooper v. Aaron, 358 U. S. 1, 18 (1958). Mr. Justice Black also was unwilling to accept the broad construction of enforcement powers formulated in the opinion of Mr. Justice Brennan, joined by Justices White and Marshall.9
The Court today fails to heed this prior precedent. To permit congressional power to prohibit the conduct challenged in this case requires state and local governments to cede far more of their powers to the Federal Government than the Civil War Amendments ever envisioned; and it requires the judiciary to cede far more of its power to interpret and enforce the Constitution than ever envisioned. The intrusion is all the more offensive to our constitutional system when it is recognized that the only values fostered are debatable assumptions about political theory which should properly be left to the local democratic process.
The Voting Rights Act is generally viewed as an exercise of Fifteenth Amendment power. See South Carolina v. Katzenbach, 383 U. S. 301 (1966). Since vote “dilution” devices are in issue in this case, the rights *208at stake are more properly viewed as Fourteenth Amendment rights. See City of Mobile v. Bolden, ante, p. 55. Nevertheless, this Court has upheld the constitutionality of the Act if it is applied to remedy violations of the Fourteenth Amendment. Gaston County v. United States, 395 U. S. 285, 290, n. 5 (1969). Moreover, the nature of the enforcement powers conferred by the Fourteenth and Fifteenth Amendments has always been treated as coextensive. See, e. g., United States v. Guest, 383 U. S. 745, 784 (1966) (opinion of Brennan, J.); James v. Bowman, 190 U. S. 127 (1903). For this reason, it is not necessary to differentiate between the Fourteenth and Fifteenth Amendment powers for the purposes of this opinion.
I share Mr. Justice Powell’s observation that the factual conclusions respecting the discriminatory effect of the annexations are highly questionable. Ante, at 195-196. I rest my dissent, however, on somewhat broader grounds.
At least four Members of the Court in Mobile held that purposeful discrimination would be prerequisite to establishing a constitutional violation in a case alleging vote dilution under the Fourteenth and Fifteenth Amendments. Ante, at 66-68 (opinion of Stewart, J.). While a majority of the Court might adopt this view, see ante, at 94 (opinion of White, J.), the voting procedures adopted by Rome would appear to readily meet the standards of constitutionality established by Mr. Justice SteveNs. See ante, at 90.
In City of Petersburg v. United States, 354 F. Supp. 1021 (DC 1972), summarily aff’d, 410 U. S. 962 (1973), the District Court did find that an annexation scheme could be prohibited solely on the basis of its disparate impact, without a finding of purposeful discrimination on the part of the local government. Petersburg cannot be considered dispositive of the question presented in this case, however. The court did not address any possible constitutional difficulties with its conclusion, and thus it is not clear that these arguments were raised by the parties. An unexplicated summary affirmance by this Court affirms only the judgment, not the reasoning, of the District Court. See Hicks v. Miranda, 422 U. S. 332 (1975).
This issue was also not squarely presented or resolved in United Jewish Organizations v. Carey, 430 U. S. 144 (1977). In UJO, the issue was whether the State could constitutionally take racial criteria into account in drawing its district lines where such redistricting was not strictly necessary to eliminate the effects of past discriminatory districting or apportionment. The Court found that use of these criteria was proper, for differing reasons. In an opinion by MR. Justice White, joined by three other Members of the Court, it was suggested in part that the Voting Rights Act could constitutionally require this. The only question, however, was the constitutionality of state use of racial criteria, vis-a-vis other citizens, and not the constitutionality of congressional Acts which required state governments to use racial criteria against their will. In another part of the opinion, Mr. Justice White reasoned that “the State is [not] powerless to minimize the consequences of racial discrimination by voters when it is regularly practiced at the polls.” Id., at 167. While States may be empowered to voluntarily use racial criteria in order to minimize the effects of racial-bloc voting, that conclusion does not determine the constitutional authority of Congress to require States to use racial criteria in structuring their governments.
See the reference to the legislative history in United Jewish Organizations v. Carey, supra, at 168.
See, e. g., Burt, Miranda And Title II: A Morganatic Marriage, 1969 S. Ct. Rev. 81.
Explicit support can also be derived from Mr. Justice Harlan’s dissenting opinion, joined by Mr. Justice Stewart, in Katzenbach v. Morgan, 384 U. S. 641, 659 (1966). Mr. Justice Harlan clarified the need for the remedial construction of congressional powers. It is also unnecessary, however, to read the majority opinion as establishing the Court’s rejection of the remedial construction of the Civil Rights Cases. While Mr. Justice Brennan^s majority opinion did contain language suggesting a rejection of the “remedial” construction of the enforcement powers, the opinion also advanced a remedial rationale which supports the determination reached by the Court. Compare the rationales forwarded at 384 U. S., at 654 with the statements, id., at 656. It would be particularly inappropriate to construe Katzenbach v. Morgan as a rejection of the remedial interpretation of- congressional powers in view of this Court’s subsequent decision in Oregon v. Mitchell.
Since Mr. Justice Black found that congressional powers were more circumscribed when not acting to counter racial discrimination under the Fourteenth Amendment, he did not have to determine the precise nature of congressional powers when they were exercised in the field of racial relations. His analysis of the nationwide ban on literacy tests, also presented in Oregon v. Mitchell, however, is consistent with a remedial interpretation of those powers.