delivered the opinion of the Court.

The constitutionality of Georgia’s congressional redistriet-ing plan is at issue here. In Shaw v. Reno, 509 U. S. 630 (1993), we held that a plaintiff states a claim under the Equal Protection Clause by alleging that a state redistricting plan, on its face, has no rational explanation save as an effort to separate voters on the basis of race. The question we now decide is whether Georgia’s new Eleventh District gives rise to a valid equal protection claim under the principles an*904nounced in Shaw, and, if so, whether it can be sustained nonetheless as narrowly tailored to serve a compelling governmental interest.

I

A

The Equal Protection Clause of the Fourteenth Amendment provides that no State shall “deny to any person within its jurisdiction the equal protection of the laws.” U. S. Const., Amdt. 14, § 1. Its central mandate is racial neutrality in governmental decisionmaking. See, e. g., Loving v. Virginia, 388 U. S. 1, 11 (1967); McLaughlin v. Florida, 379 U. S. 184, 191-192 (1964); see also Brown v. Board of Education, 347 U. S. 483 (1954). Though application of this imperative raises difficult questions, the basic principle is straightforward: “Racial and ethnic distinctions of any sort are inherently suspect and thus call for the most exacting judicial examination. . . . This perception of racial and ethnic distinctions is rooted in our Nation’s constitutional and demographic history.” Regents of Univ. of Cal. v. Bakke, 438 U. S. 265, 291 (1978) (opinion of Powell, J.). This rule obtains with equal force regardless of “the race of those burdened or benefited by a particular classification.” Richmond v. J. A. Croson Co., 488 U. S. 469, 494 (1989) (plurality opinion) (citations omitted); id., at 520 (Scalia, J., concurring in judgment) (“I agree... with Justice O’Connor’s conclusion that strict scrutiny must be applied to all governmental classification by race”); see also Adarand Constructors, Inc. v. Peña, ante, at 224; Bakke, supra, at 289-291 (opinion of Powell, J.). Laws classifying citizens on the basis of race cannot be upheld unless they are narrowly tailored to achieving a compelling state interest. See, e. g., Adarand, ante, at 227; Croson, supra, at 494 (plurality opinion); Wygant v. Jackson Bd. of Ed., 476 U. S. 267, 274, 280, and n. 6 (1986) (plurality opinion).

*905In Shaw v. Reno, supra, we recognized that these equal protection principles govern a State’s drawing of congressional districts, though, as our cautious approach there discloses, application of these principles to electoral districting is a most delicate task. Our analysis began from the premise that “[l]aws that explicitly distinguish between individuals on racial grounds fall within the core of [the Equal Protection Clause’s] prohibition.” Id., at 642. This prohibition extends not just to explicit racial classifications, but also to laws neutral on their face but “‘unexplainable on grounds other than race.’ ” Id., at 644 (quoting Arlington Heights v. Metropolitan Housing Development Corp., 429 U. S. 252, 266 (1977)). Applying this basic equal protection analysis in the voting rights context, we held that “redistricting legislation that is so bizarre on its face that it is ‘unexplainable on grounds other than race,’ . . . demands the same close scrutiny that we give other state laws that classify citizens by race.” 509 U. S., at 644 (quoting Arlington Heights, supra, at 266).

This litigation requires us to apply the principles articulated in Shaw to the most recent congressional redistricting plan enacted by the State of Georgia.

B

In 1965, the Attorney General designated Georgia a covered jurisdiction under § 4(b) of the Voting Rights Act (Act), 79 Stat. 438, as amended, 42 U. S. C. § 1973b(b). 30 Fed.Reg. 9897 (1965); see 28 CFR pt. 51, App.; see also City of Rome v. United States, 446 U. S. 156, 161 (1980). In consequence, § 5 of the Act requires Georgia to obtain either administrative preclearance by the Attorney General or approval by the United States District Court for the District of Columbia of any change in a “standard, practice, or procedure with respect to voting” made after November 1, 1964. 42 U. S. C. § 1973c. The preclearance mechanism applies to *906congressional redistricting plans, see, e. g., Beer v. United States, 425 U. S. 130, 133 (1976), and requires that the proposed change “not have the purpose and will not have the effect of denying or abridging the right to vote on account of race or color.” 42 U. S. C. § 1973c. “[T]he purpose of § 5 has always been to insure that no voting-procedure changes would be made that would lead to a retrogression in the position of racial minorities with respect to their effective exercise of the electoral franchise.” Beer, supra, at 141.

Between 1980 and 1990, one of Georgia’s 10 congressional districts was a majority-black district, that is, a majority of the district’s voters were black. The 1990 Decennial Census indicated that Georgia’s population of 6,478,216 persons, 27% of whom are black, entitled it to an additional eleventh congressional seat, App. 9, prompting Georgia’s General Assembly to redraw the State’s congressional districts. Both the House and the Senate adopted redistricting guidelines which, among other things, required single-member districts of equal population, contiguous geography, nondilution of minority voting strength, fidelity to precinct lines where possible, and compliance with §§ 2 and 5 of the Act, 42 U. S. C. §§ 1973, 1973c. See App. 11-12. Only after these requirements were met did the guidelines permit drafters to consider other ends, such as maintaining the integrity of political subdivisions, preserving the core of existing districts, and avoiding contests between incumbents. Id., at 12.

A special session opened in August 1991, and the General Assembly submitted a congressional redistricting plan to the Attorney General for preclearance on October 1, 1991. The legislature’s plan contained two majority-minority districts, the Fifth and Eleventh, and an additional district, the Second, in which blacks comprised just over 35% of the voting age population. Despite the plan’s increase in the number of majority-black districts from one to two and the absence of any evidence of an intent to discriminate against minority voters, 864 F. Supp. 1354, 1363, and n. 7 (SD Ga. 1994), the *907Department of Justice refused preclearance on January 21, 1992. App. 99-107. The Department’s objection letter noted a concern that Georgia had created only two majority-minority districts, and that the proposed plan did not “recognize” certain minority populations by placing them in a majority-black district. Id., at 105, 105-106.

The General Assembly returned to the drawing board. A new plan was enacted and submitted for preclearance. This second attempt assigned the black population in Central Georgia’s Baldwin County to the Eleventh District and increased the black populations in the Eleventh, Fifth, and Second Districts. The Justice Department refused preclearance again, relying on alternative plans proposing three majority-minority districts. Id., at 120-126. One of the alternative schemes relied on by the Department was the so-called “max-black” plan, 864 F. Supp., at 1360, 1362-1363, drafted by the American Civil Liberties Union (ACLU) for the General Assembly’s black caucus. The key to the ACLU’s plan was the “Macon/Savannah trade.” The dense black population in the Macon region would be transferred from the Eleventh District to the Second, converting the Second into a majority-black district, and the Eleventh District’s loss in black population would be offset by extending the Eleventh to include the black populations in Savannah. Id., at 1365-1366. Pointing to the General Assembly’s refusal to enact the Macon/Savannah swap into law, the Justice Department concluded that Georgia had “failed to explain adequately” its failure to create a third majority-minority district. App. 125. The State did not seek a declaratory judgment from the District Court for the District of Columbia. 864 F. Supp., at 1366, n. 11.

Twice spurned, the General Assembly set out to create three majority-minority districts to gain preclearance. Id., at 1366. Using the ACLU’s “max-black” plan as its benchmark, id., at 1366-1367, the General Assembly enacted a plan that

*908“bore all the signs of [the Justice Department’s] involvement: The black population of Meriwether County was gouged out of the Third District and attached to the Second District by the narrowest of land bridges; Ef-fingham and Chatham Counties were split to make way for the Savannah extension, which itself split, the City of Savannah; and the plan as a whole split 26 counties, 23 more than the existing congressional districts.” Id., at 1367.

See Appendix A, infra, following p. 928. The new plan also enacted the Macon/Savannah swap necessary to create a third majority-black district. The Eleventh District lost the black population of Macon, but picked up Savannah, thereby connecting the black neighborhoods of metropolitan Atlanta and the poor black populace of coastal Chatham County, though 260 miles apart in distance and worlds apart in culture. In short, the social, political, and economic makeup of the Eleventh District tells a tale of disparity, not community. See 864 F. Supp., at 1376-1377, 1389-1390; Plaintiff’s Exh. No. 85, pp. 10-27 (report of Timothy G. O’Rourke, Ph.D.). As the appendices to this opinion attest,

“[t]he populations of the Eleventh are centered around four discrete, widely spaced urban centers that have absolutely nothing to do with each other, and stretch the district hundreds of miles across rural counties and narrow swamp corridors.” 864 F. Supp., at 1389 (footnote omitted).

“The dense population centers of the approved Eleventh District were all majority-black, all at the periphery of the district, and in the case of Atlanta, Augusta and Savannah, all tied to a sparsely populated rural core by even less populated land bridges. Extending from Atlanta to the Atlantic, the Eleventh covered 6,784.2 square miles, splitting eight counties and five municipalities along the way.” Id., at 1367 (footnote omitted).

*909The Almanac of American Politics has this to say about the Eleventh District: “Geographically, it is á monstrosity, stretching from Atlanta to Savannah. Its core is the plantation country in the center of the state, lightly populated, but heavily black. It links by narrow corridors the black neighborhoods in Augusta, Savannah and southern DeKalb County.” M. Barone & G. Ujifusa, Almanac of American Politics 356 (1994). Georgia’s plan included three majority-black districts, though, and received Justice Department preclearance on April 2, 1992. Plaintiff’s Exh. No. 6; see 864 F. Supp., at 1367.

Elections were held under the new congressional redistricting plan on November 4, 1992, and black candidates were elected to Congress from all three majority-black districts. Id., at 1369. On January 13, 1994, appellees, five white voters from the Eleventh District, filed this action against various state officials (Miller Appellants) in the United States District Court for the Southern District of Georgia. Id., at 1369, 1370. As residents of the challenged Eleventh District, all appellees had standing. See United States v. Hays, ante, at 744-745. Their suit alleged that Georgia’s Eleventh District was a racial gerrymander and so a violation of the Equal Protection Clause as interpreted in Shaw v. Reno. A three-judge court was convened pursuant to 28 U. S. C. § 2284, and the United States and a number of Georgia residents intervened in support of the defendant-state officials.

A majority of the District Court panel agreed that the Eleventh District was invalid under Shaw, with one judge dissenting. 864 F. Supp. 1354 (1994). After sharp criticism of the Justice Department for its use of partisan advocates in its dealings with state officials and for its close cooperation with the ACLU’s vigorous advocacy of minority district maximization, the majority turned to a careful interpretation of our opinion in Shaw. It read Shaw to require strict scrutiny whenever race is the “overriding, predominant force” in the redistrieting process. 864 F. Supp., at *9101372 (emphasis deleted). Citing much evidence of the legislature’s purpose and intent in creating the final plan, as well as the irregular shape of the district (in particular several appendages drawn for the obvious purpose of putting black populations into the district), the court found that race was the overriding and predominant force in the districting determination. Id., at 1378. The court proceeded to apply strict scrutiny. Though rejecting proportional representation as a compelling interest, it was willing to assume that compliance with the Act would be a compelling interest. Id., at 1381-1382. As to the latter, however, the court found that the Act did not require three majority-black districts, and that Georgia’s plan for that reason was not narrowly tailored to the goal of complying with the Act. Id., at 1392-1393.

Appellants filed notices of appeal and requested a stay of the District Court’s judgment, which we granted pending the filing and disposition of the appeals in this litigation, Miller v. Johnson, 512 U. S. 1283 (1994). We later noted probable jurisdiction. 513 U. S. 1071 (1995); see 28 U. S. C. § 1253.

II

A

Finding that the “evidence of the General Assembly’s intent to racially gerrymander the Eleventh District is overwhelming, and practically stipulated by the parties involved,” the District Court held that race was the predominant, overriding factor in drawing the Eleventh District. 864 F. Supp., at 1374; see id., at 1374-1378. Appellants do not take issue with the court’s factual finding of this racial motivation. Rather, they contend that evidence of a legislature’s deliberate classification of voters on the basis of race cannot alone suffice to state a claim under Shaw. They argue that, regardless of the legislature’s purposes, a plaintiff must demonstrate that a district’s shape is so bizarre that it is unexplainable other than on the basis of race, and that *911appellees failed to make that showing here. Appellants’ conception of the constitutional violation misapprehends our holding in Shaw and the equal protection precedent upon which Shaw relied.

Shaw recognized a claim “analytically distinct” from a vote dilution claim. 509 U. S., at 652; see id., at 649-650. Whereas a vote dilution claim alleges that the State has enacted a particular voting scheme as a purposeful device “to minimize or cancel out the voting potential of racial or ethnic minorities,” Mobile v. Bolden, 446 U. S. 55, 66 (1980) (citing cases), an action disadvantaging voters of a particular race, the essence of the equal protection claim recognized in Shaw is that the State has used race as a basis for separating voters into districts. Just as the State may not, absent extraordinary justification, segregate citizens on the basis of race in its public parks, New Orleans City Park Improvement Assn. v. Detiege, 358 U. S. 54 (1958) (per curiam), buses, Gayle v. Browder, 352 U. S. 903 (1956) (per curiam), golf courses, Holmes v. Atlanta, 350 U. S. 879 (1955) (per curiam), beaches, Mayor of Baltimore v. Dawson, 350 U. S. 877 (1955) (per curiam), and schools, Brown v. Board of Education, 347 U. S. 483 (1954), so did we recognize in Shaw that it may not separate its citizens into different voting districts on the basis of race. The idea is a simple one: “At the heart of the Constitution’s guarantee of equal protection lies the simple command that the Government must treat citizens ‘as individuals, not “as simply components of a racial, religious, sexual or national class.” ’ ” Metro Broadcasting, Inc. v. FCC, 497 U. S. 547, 602 (1990) (O’Connor, J., dissenting) (quoting Arizona Governing Comm. for Tax Deferred Annuity and Deferred Compensation Plans v. Norris, 463 U. S. 1073, 1083 (1983)); cf. Northeastern Fla. Chapter, Associated Gen. Contractors of America v. Jacksonville, 508 U. S. 656, 666 (1993) (“ ‘injury in fact’ ” was “denial of equal treatment ..., not the ultimate inability to obtain the benefit”). When the State assigns voters on the basis of race, it engages in *912the offensive and demeaning assumption that voters of a particular race, because of their race, “think alike, share the same political interests, and will prefer the same candidates at the polls.” Shaw, supra, at 647; see Metro Broadcasting, supra, at 636 (Kennedy, J., dissenting). Race-based assignments “embody stereotypes that treat individuals as the product of their race, evaluating their thoughts and efforts— their very worth as citizens — according to a criterion barred to the Government by history and the Constitution.” Metro Broadcasting, supra, at 604 (O’Connor, J., dissenting) (citation omitted); see Powers v. Ohio, 499 U. S. 400, 410 (1991) (“Race cannot be a proxy for determining juror bias or competence”); Palmore v. Sidoti, 466 U. S. 429, 432 (1984) (“Classifying persons according to their race is more likely to reflect racial prejudice than legitimate public concerns; the race, not the person, dictates the category”). They also cause society serious harm. As we concluded in Shaw:

“Racial classifications with respect to voting carry particular dangers. Racial gerrymandering, even for remedial purposes, may balkanize us into competing racial factions; it threatens to carry us further from the goal of a political system in which race no longer matters— a goal that the Fourteenth and Fifteenth Amendments embody, and to which the Nation continues to aspire. It is for these reasons that race-based districting by our state legislatures demands close judicial scrutiny.” Shaw, supra, at 667.

Our observation in Shaw of the consequences of racial stereotyping was not meant to suggest that a district must be bizarre on its face before there is a constitutional violation. Nor was our conclusion in Shaw that in certain instances a district’s appearance (or, to be more precise, its appearance in combination with certain demographic evidence) can give rise to an equal protection claim, 509 U. S., at 649, a holding that bizarreness was a threshold showing, as appellants be*913lieve it to be. Our circumspect approach and narrow holding in Shaw did not erect an artificial rule barring accepted equal protection analysis in other redistricting cases. Shape is relevant not because bizarreness is a necessary ele-. ment of the constitutional wrong or a threshold requirement of proof, but because it may be persuasive circumstantial evidence that race for its own sake, and not other districting principles, was the legislature’s dominant and controlling rationale in drawing its district lines. The logical implication, as courts applying Shaw have recognized, is that parties may rely on evidence other than bizarreness to establish race-based districting. See Shaw v. Hunt, 861 F. Supp. 408, 431 (EDNC 1994); Hays v. Louisiana, 839 F. Supp. 1188, 1195 (WD La. 1993), vacated, 512 U. S. 1230 (1994); but see DeWitt v. Wilson, 856 F. Supp. 1409, 1413 (ED Cal. 1994).

Our reasoning in Shaw compels this conclusion. We recognized in Shaw that, outside the districting context, statutes are subject to strict scrutiny under the Equal Protection Clause not just when they contain express racial classifications, but also when, though race neutral on their face, they are motivated by a racial purpose or object. 509 U. S., at 644. In the rare case, where the effect of government action is a pattern “‘unexplainable on grounds other than race,’ ”• ibid, (quoting Arlington Heights, 429 U. S., at 266), “[t]he evidentiary inquiry is ... relatively easy,” Arlington Heights, supra, at 266 (footnote omitted). As early as Yick Wo v. Hopkins, 118 U. S. 356 (1886), the Court recognized that a laundry permit ordinance was administered in a deliberate way to exclude all Chinese from the laundry business; and in Gomillion v. Lightfoot, 364 U. S. 339 (1960), the Court concluded that the redrawing of Tuskegee, Alabama’s municipal boundaries left no doubt that the plan was designed to exclude blacks. Even in those cases, however, it was the presumed racial purpose of state action, not its stark manifestation, that was the constitutional violation. Patterns of discrimination as conspicuous as these are rare, and *914are not a necessary predicate to a violation of the Equal Protection Clause. Cf. Arlington Heights, supra, at 266, n. 14. In the absence of a pattern as stark as those in Yick Wo or Gomillion, “impact alone is not determinative, and the Court must look to other evidence” of race-based decisionmaking. Arlington Heights, supra, at 266 (footnotes omitted).

Shaw applied these same principles to redistricting. “In some exceptional cases, a reapportionment plan may be so highly irregular that, on its face, it rationally cannot be understood as anything other than an effort to ‘segregate] . . . voters’ on the basis of race.” Shaw, supra, at 646-647 (quoting Gomillion, supra, at 341). In other cases, where the district is not so bizarre on its face that it discloses a racial design, the proof will be more “difficul[t].” 509 U. S., at 646. Although it was not necessary in Shaw to consider further the proof required in these more difficult cases, the logical import of our reasoning is that evidence other than a district’s bizarre shape can be used to support the claim.

Appellants and some of their amici argue that the Equal Protection Clause’s general proscription on race-based deci-sionmaking does not obtain in the districting context because redistricting by definition involves racial considerations. Underlying their argument are the very stereotypical assumptions the Equal Protection Clause forbids. It is true that redistricting in most cases will implicate a political calculus in which various interests compete for recognition, but it does not follow from this that individuals of the same race share a single political interest. The view that they do is “based on the demeaning notion that members of the defined racial groups ascribe to certain ‘minority views’ that must be different from those of other citizens,” Metro Broadcasting, 497 U. S., at 636 (Kennedy, J., dissenting), the precise use of race as a proxy the Constitution prohibits. Nor can the argument that districting cases are excepted from standard equal protection precepts be resuscitated by United Jewish Organizations of Williamsburgh, Inc. v. Carey, 430 U. S. *915144 (1977), where the Court addressed a claim that New York violated the Constitution by splitting a Hasidic Jewish community in order to include additional majority-minority districts. As we explained in Shaw, a majority of the Justices in UJO construed the complaint as stating a vote dilution claim, so their analysis does not apply to a claim that the State has separated voters on the basis of race. 509 U. S., at 652. To the extent any of the opinions in that “highly fractured decision,” id., at 651, can be interpreted as suggesting that a State’s assignment of voters on the basis of race would be subject to anything but our strictest scrutiny, those views ought not be deemed controlling.

In sum, we make clear that parties alleging that a State has assigned voters on the basis of race are neither confined in their proof to evidence regarding the district’s geometry and makeup nor required to make a threshold showing of bizarreness. Today’s litigation requires us further to consider the requirements of the proof necessary to sustain this equal protection challenge.

B

Federal-court review of districting legislation represents a serious intrusion on the most vital of local functions. It is well settled that “reapportionment is primarily the duty and responsibility of the State.” Chapman v. Meier, 420 U. S. 1, 27 (1975); see, e. g., Voinovich v. Quilter, 507 U. S. 146, 156-157 (1993); Growe v. Emison, 507 U. S. 25, 34 (1993). Electoral districting is a most difficult subject for legislatures, and so the States must have discretion to exercise the political judgment necessary to balance competing interests. Although race-based decisionmaking is inherently suspect, e. g., Adarand, ante, at 218 (citing Bakke, 438 U. S., at 291 (opinion of Powell, J.)), until a claimant makes a showing sufficient to support that allegation the good faith of a state legislature must be presumed, see id., at 318-319 (opinion of Powell, J.). The courts, in assessing the sufficiency of a challenge to a districting plan, must be sensitive to the com*916plex interplay of forces that enter a legislature’s redistricting calculus. Redistricting legislatures will, for example, almost always be aware of racial demographics; but it does not follow that race predominates in the redistricting process. Shaw, supra, at 646; see Personnel Administrator of Mass. v. Feeney, 442 U. S. 256, 279 (1979) (“‘[Discriminatory’ purpose’ . . . implies more than intent as volition or intent as awareness of consequences. It implies that the decision-maker ... selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects”) (footnotes and citation omitted). The distinction between being aware of racial considerations and being motivated by them may be difficult to make. This eviden-tiary difficulty, together with the sensitive nature of redistricting and the presumption of good faith that must be accorded legislative enactments, requires courts to exercise extraordinary caution in adjudicating claims that a State has drawn district lines on the basis of race. The plaintiff’s burden is to show, either through circumstantial evidence of a district’s shape and demographics or more direct evidence going to legislative purpose, that race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without a particular district. To make this showing, a plaintiff must prove that the legislature subordinated traditional race-neutral districting principles, including but not limited to compactness, contiguity, and respect for political subdivisions or communities defined by actual shared interests, to racial considerations. Where these or other race-neutral considerations are the basis for redistricting legislation, and are not subordinated to race, a State can “defeat a claim that a district has been gerrymandered on racial lines.” Shaw, supra, at 647. These principles inform the plaintiff’s burden of proof at trial. Of course, courts must also recognize these principles, and the intrusive potential of judicial intervention into the legislative realm, when assessing under the Federal Rules of Civil *917Procedure the adequacy of a plaintiff’s showing at the various stages of litigation and determining whether to permit discovery or trial to proceed. See, e. g., Fed. Rules Civ. Proc. 12(b) and (e), 26(b)(2), 56; see also Celotex Corp. v. Catrett, 477 U. S. 317, 327 (1986).

In our view, the District Court applied the correct analysis, and its finding that race was the predominant factor motivating the drawing of the Eleventh District was not clearly erroneous. The court found it was “exceedingly obvious” from the shape of the Eleventh District, together with the relevant racial demographics, that the drawing of narrow land bridges to incorporate within the district outlying appendages containing nearly 80% of the district’s total black population was a deliberate attempt to bring black populations into the district. 864 F. Supp., at 1375; see id., at 1374-1376. Although by comparison with other districts the geometric shape of the Eleventh District may not seem bizarre on its face, when its shape is considered in conjunction with its racial and population densities, the story of racial gerrymandering seen by the District Court becomes much clearer. See Appendix B, infra, following p. 928; see also App. 133. Although this evidence is quite compelling, we need not determine whether it was, standing alone, sufficient to establish a Shaw claim that the Eleventh District is unexplainable other than by race. The District Court had before it considerable additional evidence showing that the General Assembly was motivated by a predominant, overriding desire to assign black populations to the Eleventh District and thereby permit the creation of a third majority-black district in the Second. 864 F. Supp., at 1372, 1378.

The court found that “it became obvious,” both from the Justice Department’s objection letters and the three pre-clearance rounds in general, “that [the Justice Department] would accept nothing less than abject surrender to its maximization agenda.” Id., at 1366, n. 11; see id., at 1360-1367; see also Arlington Heights, 429 U. S., at 267 (“historical *918background of the decision is one evidentiary source”). It further found that the General Assembly acquiesced and as a consequence was driven by its overriding desire to comply with the Department’s maximization demands. The court supported its conclusion not just with the testimony of Linda Meggers, the operator of “Herschel,” Georgia’s reapportionment computer, and “probably the most knowledgeable person available on the subject of Georgian redistricting,” 864 F. Supp., at 1361, 1363, n. 6, 1366, but also with the State’s own concessions. The State admitted that it “‘would not have added those portions of Effingham and Chatham Counties that are now in the [far southeastern extension of the] present Eleventh Congressional District but for the need to include additional black population in that district to offset the loss of black population caused by the shift of predominantly black portions of Bibb County in the Second Congressional District which occurred in response to the Department of Justice’s March 20th, 1992, objection letter.’ ” Id., at 1377. It conceded further that “[t]o the extent that precincts in the Eleventh Congressional District are split, a substantial reason for their being split was the objective of increasing the black population of that district.” Ibid. And in its brief to this Court, the State concedes that “[i]t is undisputed that Georgia’s eleventh is the product of a desire by the General Assembly to create a majority black district.” Brief for Miller Appellants 30. Hence the trial court had little difficulty concluding that the Justice Department “spent months demanding purely race-based revisions to Georgia’s redistricting plans, and that Georgia spent months attempting to comply.” 864 F. Supp., at 1377. On this record, we fail to see how the District Court could have reached any conclusion other than that race was the predominant factor in drawing Georgia’s Eleventh District; and in any event we conclude the court’s finding is not clearly erroneous. Cf. Wright v. Rockefeller, 376 U. S. 52, 56-57 (1964) (evidence presented “conflicting inferences” and therefore “failed to *919prove that the New York Legislature was either motivated by racial considerations or in fact drew the districts on racial lines”).

In light of its well-supported finding, the District Court was justified in rejecting the various alternative explanations offered for the district. Although a legislature’s compliance with “traditional districting principles such as. compactness, contiguity, and respect for political subdivisions” may well suffice to refute a claim of racial gerrymandering, Shaw, 509 U. S., at 647, appellants cannot make such a refutation where, as here, those factors were subordinated to racial objectives. Georgia’s Attorney General objected to the Justice Department’s demand for three majority-black districts on the ground that to do so the State would have to “violate all reasonable standards of compactness and contiguity.” App. 118. This statement from a state official is powerful evidence that the legislature subordinated traditional dis-tricting principles to race when it ultimately enacted a plan creating three majority-black districts, and justified the District Court’s finding that “every [objective districting] factor that could realistically be subordinated to racial tinkering in fact suffered that fate.” 864 F. Supp., at 1384; see id., at 1364, n. 8; id., at 1375 (“While the boundaries of the Eleventh do indeed follow many precinct lines, this is because Ms. Meggers designed the Eleventh District along racial lines, and race data was most accessible to her at the precinct level”).

Nor can the State’s districting legislation be rescued by mere recitation of purported communities of interest. The evidence was compelling “that there are no tangible ‘communities of interest’ spanning the hundreds of miles of the Eleventh District.” Id., at 1389-1390. A comprehensive report demonstrated the fractured political, social, and economic interests within the Eleventh District’s black population. See Plaintiff’s Exh. No. 85, pp. 10-27 (report of Timothy G. O’Rourke, Ph.D.). It is apparent that it was not alleged *920shared interests but rather the object of maximizing the district’s black population and obtaining Justice Department approval that in fact explained the General Assembly’s actions. 864 F. Supp., at 1366, 1378, 1380. A State is free to recognize communities that have a particular racial makeup, provided its action is directed toward some common thread of relevant interests. “[WJhen members of a racial group live together in one community, a reapportionment plan that concentrates members of the group in one district and excludes them from others may reflect wholly legitimate purposes.” Shaw, 509 U. S., at 646. But where the State assumes from a group of voters’ race that they “think alike, share the same political interests, and will prefer the same candidates at the polls,” it engages in racial stereotyping at odds with equal protection mandates. Id., at 647; cf. Powers v. Ohio, 499 U. S. 400, 410 (1991) (“We may not accept as a defense to racial discrimination the very stereotype the law condemns”).

Race was, as the District Court found, the predominant, overriding factor explaining the General Assembly’s decision to attach to the Eleventh District various appendages containing dense majority-black populations. 864 F. Supp., at 1372, 1378. As a result, Georgia’s congressional redistricting plan cannot be upheld unless it satisfies strict scrutiny, our most rigorous and exacting standard of constitutional review.

Ill

To satisfy strict scrutiny, the State must demonstrate that its districting legislation is narrowly tailored to achieve a compelling interest. Shaw, supra, at 653-657; see also Croson, 488 U. S., at 494 (plurality opinion); Wygant, 476 U. S., at 274, 280, and n. 6 (plurality opinion); cf. Adarand, ante, at 227. There is a “significant state interest in eradicating the effects of past racial discrimination.” Shaw, supra, at 656. The State does not argue, however, that it created the Eleventh District to remedy past discrimination, and with good *921reason: There is little doubt that the State’s true interest in designing the Eleventh District was creating a third majority-black district to satisfy the Justice Department’s preclearance demands. 864 F. Supp., at 1378 (“[T]he only interest the General Assembly had in mind when drafting the current congressional plan was satisfying [the Justice Department’s] preclearance requirements”); id., at 1366; compare Wygant, supra, at 277 (plurality opinion) (under strict scrutiny, State must have convincing evidence that remedial action is necessary before implementing affirmative action), with Heller v. Doe, 509 U. S. 312, 320 (1993) (under rational-basis review, legislature need not “‘actually articulate at any time the purpose or rationale supporting its classification’ ”) (quoting Nordlinger v. Hahn, 505 U. S. 1, 15 (1992)). Whether or not in some cases compliance with the Act, standing alone, can provide a compelling interest independent of any interest in remedying past discrimination, it cannot do so here. As we suggested in Shaw, compliance with federal antidiscrimination laws cannot justify race-based dis-tricting where the challenged district was not reasonably necessary under a constitutional reading and application of those laws. See 509 U. S., at 653-655. The congressional plan challenged here was not required by the Act under a correct reading of the statute.

The Justice Department refused to preclear both of Georgia’s first two submitted redistricting plans. The District Court found that the Justice Department had adopted a “black-maximization” policy under §5, and that it was clear from its objection letters that the Department would not grant preclearance until the State made the “Macon/ Savannah trade” and created a third majority-black district. 864 F Supp., at 1366, 1380. It is, therefore, safe to say that the congressional plan enacted in the end was required in order to obtain preclearance. It does not follow, however, that the plan was required by the substantive provisions of the Act.

*922We do not accept the contention that the State has a compelling interest in complying with whatever preclearance mandates the Justice Department issues. When a state governmental entity seeks to justify race-based remedies to cure the effects of past discrimination, we do not accept the government’s mere assertion that the remedial action is required. Rather, we insist on a strong basis in evidence of the harm being remedied. See, e, g., Shaw, supra, at 656; Croson, supra, at 500-501; Wygant, supra, at 276-277 (plurality opinion). “The history of racial classifications in this country suggests that blind judicial deference to legislative or executive pronouncements of necessity has no place in equal protection analysis.” Croson, supra, at 501. Our presumptive skepticism of all racial classifications, see Ada-rand, ante, at 223-224, prohibits us as well from accepting on its face the Justice Department’s conclusion that racial districting is necessary under the Act. Where a State relies on the Department’s determination that race-based districting is necessary to comply with the Act, the judiciary retains an independent obligation in adjudicating consequent equal protection challenges to ensure that the State’s actions are narrowly tailored to achieve a compelling interest. See Shaw, supra, at 654. Were we to accept the Justice Department’s objection itself as a compelling interest adequate to insulate racial districting from constitutional review, we would be surrendering to the Executive Branch our role in enforcing the constitutional limits on race-based official action. We may not do so. See, e. g., United States v. Nixon, 418 U. S. 683, 704 (1974) (judicial power cannot be shared with Executive Branch); Marbury v. Madison, 1 Cranch 137, 177 (1803) (“It is emphatically the province and duty of the judicial department to say what the law is”); cf. Baker v. Carr, 369 U. S. 186, 211 (1962) (Supreme Court is “ultimate interpreter of the Constitution”); Cooper v. Aaron, 358 U. S. 1, 18 (1958) (“permanent and indispensable feature of our constitutional system” is that “the federal *923judiciary is supreme in the exposition of the law of the Constitution”).

For the same reasons, we think it inappropriate for a court engaged in constitutional scrutiny to accord deference to the Justice Department’s interpretation of the Act. Although we have deferred to the Department’s interpretation in certain statutory cases, see, e. g., Presley v. Etowah County Comm’n, 502 U. S. 491, 508-509 (1992), and cases cited therein, we have rejected agency interpretations to which we would otherwise defer where they raise serious constitutional questions. Edward J. DeBartolo Corp. v. Florida Gulf Coast Building & Constr. Trades Council, 485 U. S. 568, 574-575 (1988). When the Justice Department’s interpretation of the Act compels race-based districting, it by definition raises a serious constitutional question, see, e. g., Bakke, 438 U. S., at 291 (opinion of Powell, J.) (“Racial and ethnic distinctions of any sort are inherently supect” under the Equal Protection Clause), and should not receive deference.

Georgia’s drawing of the Eleventh District was not required under the Act because there was no reasonable basis to believe that Georgia’s earlier enacted plans violated §5. Wherever a plan is “ameliorative,” a term we have used to describe plans increasing the number of majority-minority districts, it “cannot violate § 5 unless the new apportionment itself so discriminates on the basis of race or color as to violate the Constitution.” Beer, 425 U. S., at 141. Georgia’s first and second proposed plans increased the number of majority-black districts from 1 out of 10 (10%) to 2 out of 11 (18.18%). These plans were “ameliorative” and could not have violated §5’s nonretrogression principle. Ibid. Acknowledging as much, see Brief for United States 29; 864 F. Supp., at 1384-1385, the United States now relies on the fact that the Justice Department may object to a state proposal either on the ground that it has a prohibited purpose or a prohibited effect, see, e. g., Pleasant Grove v. United *924States, 479 U. S. 462, 469 (1987). The Government justifies its preclearance objections on the ground that the submitted plans violated § 5’s purpose element. The key to the Government’s position, which is plain from its objection letters if not from its briefs to this Court, compare App. 105-106, 124-125 with Brief for United States 31-33, is and always has been that Georgia failed to proffer a nondiscriminatory purpose for its refusal in the first two submissions to take the steps necessary to create a third majority-minority district.

The Government’s position is insupportable. “[Ajmeliora-tive changes, even if they fall short of what might be accomplished in terms of increasing minority representation, cannot be found to violate section 5 unless they so discriminate on the basis of race or color as to violate the Constitution.” Days, Section 5 and the Role of the Justice Department, in B. Grofman & C. Davidson, Controversies in Minority Voting 56 (1992). Although it is true we have held that the State has the burden to prove a nondiscriminatory purpose under § 5, e. g., Pleasant Grove, supra, at 469, Georgia’s Attorney General provided a detailed explanation for the State’s initial decision not to enact the max-black plan, see App. 117-119. The District Court accepted this explanation, 864 F. Supp., at 1365, and found an absence of any discriminatory intent, id., at 1363, and n. 7. The State’s policy of adhering to other districting principles instead of creating as many majority-minority districts as possible does not support an inference that the plan “so discriminates on the basis of race or color as to violate the Constitution,” Beer, supra, at 141; see Mobile v. Bolden, 446 U. S. 55 (1980) (plurality opinion), and thus cannot provide any basis under § 5 for the Justice Department’s objection.

Instead of grounding its objections on evidence of a discriminatory purpose, it would appear the Government was driven by its policy of maximizing majority-black districts. Although the Government now disavows having had that *925policy, see Brief for United States 35, and seems to concede its impropriety, see Tr. of Oral Arg. 32-33, the District Court’s well-documented factual finding was that the Department did adopt a maximization policy and followed it in objecting to Georgia’s first two plans.* One of the two Department of Justice line attorneys overseeing the Georgia preclearance process himself disclosed that “‘what we did and what I did specifically was to take a ... map of the State of Georgia shaded for race, shaded by minority concentration, and overlay the districts that were drawn by the State of Georgia and see how well those lines adequately reflected black voting strength.’” 864 F. Supp., at 1362, n. 4. In utilizing §5 to require States to create majority-minority districts wherever possible, the Department of Justice expanded its authority under the statute beyond what Congress intended and we have upheld.

Section 5 was directed at preventing a particular set of invidious practices that had the effect of “undo[ing] or defeating] the rights recently won by non white voters.” *926H. R. Rep. No. 91-397, p. 8 (1969). As we explained in Beer v. United States,

‘“Section 5 was a response to a common practice in some jurisdictions of staying one step ahead of the federal courts by passing new discriminatory voting laws as soon as the old ones had been struck down. That practice had been possible because each new law remained in effect until the Justice Department or private plaintiffs were able to sustain the burden of proving that the new law, too, was discriminatory.... Congress therefore decided, as the Supreme Court held it could, “to shift the advantage of time and inertia from the perpetrators of the evil to its victim,” by “freezing election procedures in the covered areas unless the changes can be shown to be nondiscriminatory.” ’ ” 425 U. S., at 140 (quoting H. R. Rep. No. 94-196, pp. 57-58 (1975) (footnotes omitted)).

Based on this historical understanding, we recognized in Beer that “the purpose of § 5 has always been to insure that no voting-procedure changes would be made that would lead to a retrogression in the position of racial minorities with respect to their effective exercise of the electoral franchise.” 425 U. S., at 141. The Justice Department’s maximization policy seems quite far removed from this purpose. We are especially reluctant to conclude that § 5 justifies that policy given the serious constitutional concerns it raises. In South Carolina v. Katzenbach, 383 U. S. 301 (1966), we ¡upheld §5 as a necessary and constitutional response to some States’ “extraordinary stratagem[s] of contriving new rules of various kinds for the sole purpose of perpetuating voting discrimination in the face of adverse federal court decrees.” Id., at 335 (footnote omitted); see also City of Rome v. United States, 446 U. S., at 173-183. But our belief in Katzenbach that the federalism costs exacted by § 5 preclearance could be justified by those extraordinary circumstances does not *927mean they can be justified in the circumstances of this litigation. And the Justice Department’s implicit command that States engage in presumptively unconstitutional race-based districting brings the Act, once upheld as a proper exercise of Congress’ authority under §2 of the Fifteenth Amendment, Katzenbach, supra, at 327, 337, into tension with the Fourteenth Amendment. As we recalled in Katzenbach itself, Congress’ exercise of its Fifteenth Amendment authority even when otherwise proper still must “ ‘consist with the letter and spirit of the constitution.’ ” 383 U. S., at 326 (quoting McCulloch v. Maryland, 4 Wheat. 316, 421 (1819)). We need not, however, resolve these troubling and difficult constitutional questions today. There is no indication Congress intended such a far-reaching application of § 5, so we reject the Justice Department’s interpretation of the statute and avoid the constitutional problems that interpretation raises. See, e. g., DeBartolo Corp. v. Florida Gulf Coast Trades Council, 485 U. S., at 575.

>

The Act, and its grant of authority to the federal courts to uncover official efforts to abridge minorities’ right to vote, has been of vital importance in eradicating invidious discrimination from the electoral process and enhancing the legitimacy of our political institutions. Only if our political system and our society cleanse themselves of that discrimination will all members of the polity share an equal opportunity to gain public office regardless of race. As a Nation we share both the obligation and the aspiration of working toward this end. The end is neither assured nor well served, however, by carving electorates into racial blocs. “If our society is to continue to progress as a multiracial democracy, it must recognize that the automatic invocation of race stereotypes retards that progress and causes continued hurt and injury.” Edmondson v. Leesville Concrete Co., 500 U. S. 614, 630-631 (1991). It takes a shortsighted and *928unauthorized view of the Voting Rights Act to invoke that statute, which has played a decisive role in redressing some of our worst forms of discrimination, to demand the very racial stereotyping the Fourteenth Amendment forbids.

* * *

The judgment of the District Court is affirmed, and the cases are remanded for further proceedings consistent with this decision.

It is so ordered.

[Appendices A and B, containing a map of Georgia congressional districts and a population density map of the 11th Congressional District of Georgia, follow this page.]

concurring.

I understand the threshold standard the Court adopts— that “the legislature subordinated traditional race-neutral districting principles ... to racial considerations,” ante, at 916 — to be a demanding one. To invoke strict scrutiny, a plaintiff must show that the State has relied on race in substantial disregard of customary and traditional districting practices. Those practices provide a crucial frame of reference and therefore constitute a significant governing principle in cases of this kind. The standard would be no different if a legislature had drawn the boundaries to favor some other ethnic group; certainly the standard does not treat efforts to create majority-minority districts less favorably than similar efforts on behalf of other groups. Indeed, the driving force behind the adoption of the Fourteenth Amendment was the desire to end legal discrimination against blacks.

Application of the Court’s standard does not throw into doubt the vast majority of the Nation’s 435 congressional districts, where presumably the States have drawn the boundaries in accordance with their customary districting principles. That is so even though race may well have been *929considered in the redistricting process. See Shaw v. Reno, 509 U. S. 630, 646 (1993); ante, at 916. But application of the Court’s standard helps achieve Shaw’s basic objective of making extreme instances of gerrymandering subject to meaningful judicial review. I therefore join the Court’s opinion.

*0[[Image here]]

[[Image here]]

dissenting.

Justice Ginsburg has explained why the District Court’s opinion on the merits was erroneous and why this Court’s law-changing decision will breed unproductive litigation. I join her excellent opinion without reservation. I add these comments because I believe the appellees in these cases, like the appellees in United States v. Hays, ante, p. 737, have not suffered any legally cognizable injury.

In Shaw v. Reno, 509 U. S. 630 (1993), the Court crafted a new cause of action with two novel, troubling features. First, the Court misapplied the term “gerrymander,” previously used to describe grotesque line-drawing by a dominant group to maintain or enhance its political power at a minority’s expense, to condemn the efforts of a majority (whites) to share its power with a minority (African-Americans). Second, the Court dispensed with its previous insistence in vote dilution cases on a showing of injury to an identifiable group of voters, but it failed to explain adequately what showing a plaintiff must make to establish standing to litigate the newly minted Shaw claim. Neither in Shaw itself nor in the cases decided today has the Court coherently articulated what injury this cause of action is designed to redress. Because appellees have alleged no legally cognizable injury, they lack standing, and these cases should be dismissed. See Hays, ante, at 750-751 (Stevens, J., concurring in judgment).

Even assuming the validity of Shaw, I cannot see how appellees in these cases could assert the injury the Court attributes to them. Appellees, plaintiffs below, are white *930voters in Georgia’s Eleventh Congressional District. The Court’s conclusion that they have standing to maintain a Shaw claim appears to rest on a theory that their placement in the Eleventh District caused them “‘representational harms.’” Hays, ante, at 744, cited ante, at 909. The Shaw Court explained the concept of “representational harms” as follows: “When a district obviously is created solely to effectuate the perceived common interests of one racial group, elected officials are more likely to believe that their primary obligation is to represent only the members of that group, rather than their constituency as a whole.” Shaw, 509 U. S., at 648. Although the Shaw Court attributed representational harms solely to a message sent by the legislature’s action, those harms can only come about if the message is received — that is, first, if all or most black voters support the same candidate, and, second, if the successful candidate ignores the interests of her white constituents. Appellees’ standing, in other words, ultimately depends on the very premise the Court purports to abhor: that voters of a particular race “ ‘think alike, share the same political interests, and will prefer the same candidates at the polls.’ ” Ante, at 912 (quoting Shaw, 509 U. S., at 647). This generalization, as the Court recognizes, is “offensive and demeaning.” Ante, at 912.

In particular instances, of course, members of one race may vote by an overwhelming margin for one candidate, and in some cases that candidate will be of the same race. “Racially polarized voting” is one of the circumstances plaintiffs must prove to advance a vote dilution claim. Thornburg v. Gingles, 478 U. S. 30, 56-58 (1986). Such a claim allows voters to allege that gerrymandered district lines have impaired their ability to elect a candidate of their own race. The Court emphasizes, however, that a so-called Shaw claim is “ ‘analytically distinct’ from a vote dilution claim,” ante, at 911 (quoting Shaw, 509 U. S., at 652). Neither in Shaw, nor in Hays, nor in the instant cases has the Court answered the *931question its analytic distinction raises: If the Shaw injury does not flow from an increased probability that white candidates will lose, then how can the increased probability that black candidates will win cause white voters, such as appel-lees, cognizable harm?1

The Court attempts an explanation in these cases by equating the injury it imagines appellees have suffered with the injuries African-Americans suffered under segregation. The heart of appellees’ claim, by the Court’s account, is that “a State’s assignment of voters on the basis of race,” ante, at 915, violates the Equal Protection Clause for the same reason a State may not “segregate citizens on the basis of race in its public parks, New Orleans City Park Improvement Assn. v. Detiege, 358 U. S. 54 (1958) (per curiam), buses, Gayle v. Browder, 352 U. S. 903 (1956) (per curiam), golf courses, Holmes v. Atlanta, 350 U. S. 879 (1955) (per curiam), beaches, Mayor of Baltimore v. Dawson, 350 U. S. 877 (1955) (per curiam), and schools, Brown v. Board of Education, 347 U. S. 483 (1954).” Ante, at 911. This equation, however, fails to elucidate the elusive Shaw injury. Our desegregation cases redressed the exclusion of black citizens from public facilities reserved for whites. In these cases, in contrast, any voter, black or white, may live in the Eleventh District. What appellees contest is the inclusion of too many black voters in the district as drawn. In my view, if appellees allege no vote dilution, that inclusion can cause them no conceivable injury.

The Court’s equation of Shaw claims with our desegregation decisions is inappropriate for another reason. In each of those cases, legal segregation frustrated the public interest in diversity and tolerance by barring African-Americans *932from joining whites in the activities at issue. The districting plan here, in contrast, serves the interest in diversity and tolerance by increasing the likelihood that a meaningful number of black representatives will add their voices to legislative debates. See post, at 947-948 (Ginsburg, J., dissenting). “There is no moral or constitutional equivalence between a policy that is designed to perpetuate a caste system and one that seeks to eradicate racial subordination.” Adarand Constructors, Inc. v. Peña, ante, at 243 (Stevens, J., dissenting); see also Adarand, ante, at 247-248, n. 5 (Stevens, J., dissenting). That racial integration of the sort attempted by Georgia now appears more vulnerable to judicial challenge than some policies alleged to perpetuate racial bias, cf. Allen v. Wright, 468 U. S. 737 (1984), is anomalous, to say the least.

Equally distressing is the Court’s equation of traditional gerrymanders, designed to maintain or enhance a dominant group’s power, with a dominant group’s decision to share its power with a previously underrepresented group. In my view, districting plans violate the Equal Protection Clause when they “serve no purpose other than to favor one segment — whether racial, ethnic, religious, economic, or political — that may occupy a position of strength at a particular point in time, or to disadvantage a politically weak segment of the community.” Karcher v. Daggett, 462 U. S. 725, 748 (1983) (Stevens, J., concurring). In contrast, I do not see how a districting plan that favors a politically weak group can violate equal protection. The Constitution does not mandate any form of proportional representation, but it certainly permits a State to adopt a policy that promotes fair representation of different groups. Indeed, this Court squarely so held in Gaffney v. Cummings, 412 U. S. 735 (1973):

“[N]either we nor the district courts have a constitutional warrant to invalidate a state plan, otherwise *933within tolerable population limits, because it undertakes, not to minimize or eliminate the political strength of any group or party, but to recognize it and, through districting, provide a rough sort of proportional representation in the legislative halls of the State.” Id., at 754.

The Court’s refusal to distinguish an enactment that helps a minority group from enactments that cause it harm is especially unfortunate at the intersection of race and. voting, given that African-Americans and other disadvantaged groups have struggled so long and so hard for inclusion in that most central exercise of our democracy. See post, at 936-938 (Ginsburg, J., dissenting). I have long believed that treating racial groups differently from other identifiable groups of voters, as the Court does today, is itself an invidious racial classification. Racial minorities should receive neither more nor less protection than other groups against gerrymanders.2 A fortiori, racial minorities should not be less eligible than other groups to benefit from districting plans the majority designs to aid them.

I respectfully dissent.

dissenting.

. Legislative districting is highly political business. This Court has generally respected the competence of state legislatures to attend to the task. When race is the issue, however, we have recognized the need for judicial intervention to prevent dilution of minority voting strength. Generations of rank discrimination against African-Americans, as citizens and voters, account for that surveillance.

Two Terms ago, in Shaw v. Reno, 509 U. S. 630 (1993), this Court took up a claim “analytically distinct” from a vote dilution claim. Id., at 652. Shaw authorized judicial intervention in “extremely irregular” apportionments, id., at 642, in which the legislature cast aside traditional districting practices to consider race alone — in the Shaw case, to create a district in North Carolina in which African-Americans would compose a majority of the voters.

Today the Court expands the judicial role, announcing that federal courts are to undertake searching review of any district with contours “predominantly] motivat[ed]” by race: “[S]trict scrutiny” will be triggered not only when traditional districting practices are abandoned, but also when those practices are “subordinated to” — given less weight than — race. See ante, at 916. Applying this new “race-as-predominant-factor” standard, the Court invalidates Georgia’s districting plan even though Georgia’s Eleventh District, the focus of today’s dispute, bears the imprint of familiar districting practices. Because I do not endorse the Court’s new standard and would not upset Georgia’s plan, I dissent.

I

At the outset, it may be useful to note points on which the Court does not divide. First, we agree that federalism and the slim judicial competence to draw district lines weigh *935heavily against judicial intervention in apportionment decisions; as a rule, the task should remain within the domain of state legislatures. See ante, at 915; Reynolds v. Sims, 377 U. S. 533, 586 (1964) (“[Ljegislative reapportionment is primarily a matter for legislative consideration and determination . . . .”). Second, for most of our Nation’s history, the franchise has not been enjoyed equally by black citizens and white voters. To redress past wrongs and to avert any recurrence of exclusion of blacks from political processes, federal courts now respond to Equal Protection Clause and Voting Rights Act complaints of state action that dilutes minority voting strength. See, e. g., Thornburg v. Gingles, 478 U. S. 30 (1986); White v. Regester, 412 U. S. 755 (1973). Third, to meet statutory requirements, state legislatures must sometimes consider race as a factor highly relevant to the drawing of district lines. See Pildes & Niemi, Expressive Harms, “Bizarre Districts,” and Voting Rights: Evaluating Election-District Appearances After Shaw v. Reno, 92 Mich. L. Rev. 483, 496 (1993) (“compliance with the [Voting Rights Act] and Gingles necessarily requires race-conscious districting”). Finally, state legislatures may recognize communities that have a particular racial or ethnic makeup, even in the absence of any compulsion to do so, in order to account for interests common to or shared by the persons grouped together. See Shaw, 509 U. S., at 646 (“[Wjhen members of a racial group live together in one community, a reapportionment plan that concentrates members of the group in one district and excludes them from others may reflect wholly legitimate purposes.”).

Therefore, the fact that the Georgia General Assembly took account of race in drawing district lines — a fact not in dispute — does not render the State’s plan invalid. To offend the Equal Protection Clause, all agree, the legislature had to do more than consider race. How much more, is the issue that divides the Court today.

*936A

“We say once again what has been said on many occasions: reapportionment is primarily the duty and responsibility of the State through its legislature or other body, rather than of a federal court.” Chapman v. Meier, 420 U. S. 1, 27 (1975); see also ante, at 915. The Constitution itself allocates this responsibility to States. U. S. Const., Art. I, § 2; Growe v. Emison, 507 U. S. 25, 34 (1993).

“Districting inevitably has sharp political impact and inevitably political decisions must be made by those charged with the task.” White v. Weiser, 412 U. S. 783, 795-796 (1973). District lines are drawn to accommodate a myriad of factors — geographic, economic, historical, and political — and state legislatures, as arenas of compromise and electoral accountability, áre best positioned to mediate competing claims; courts, with a mandate to adjudicate, are ill equipped for the task.

B

Federal courts have ventured into the political thicket of apportionment when necessary to secure to members of racial minorities equal voting rights — rights denied in many States, including Georgia, until not long ago.

The Fifteenth Amendment, ratified in 1870, declares that the right to vote “shall not be denied ... by any State on account of race.” That declaration, for generations, was often honored in the breach; it was greeted by a near century of “unremitting and ingenious defiance” in several States, including Georgia. South Carolina v. Katzenbach, 383 U. S. 301, 309 (1966). After a brief interlude of black suffrage enforced by federal troops but accompanied by rampant violence against blacks, Georgia held a constitutional convention in 1877. Its purpose, according to the convention’s leader, was to “‘fix it so that the people shall rule and the Negro shall never be heard from.’” McDonald, Binford, & Johnson, Georgia, in Quiet Revolution in the South 68 (C. David*937son & B. Grofman eds. 1994) (quoting Robert Toombs). In pursuit of this objective, Georgia enacted a cumulative poll tax, requiring voters to show they had paid past as well as current poll taxes; one historian described this tax as the “most effective bar to Negro suffrage ever devised.” A. Stone, Studies in the American Race Problem 354-355 (1908).

In 1890, the Georgia General Assembly authorized “white primaries”; keeping blacks out of the Democratic primary effectively excluded them from Georgia’s political life, for victory in the Democratic primary was tantamount to election. McDonald, Binford, & Johnson, supra, at 68-69. Early in this century, Georgia Governor Hoke Smith persuaded the legislature to pass the “Disenfranchisement Act of 1908”; true to its title, this measure added various property, “good character,” and literacy requirements that, as administered, served to keep blacks from voting. Id., at 69; see also Katzenbach, 383 U. S., at 310 (tests of this order were “specifically designed to prevent Negroes from voting”). The result, as one commentator observed 25 years later, was an “ ‘almost absolute exclusion of the Negro voice in state and federal elections.’” McDonald, Binford, & Johnson, supra, at 70 (quoting R. Wardlaw, Negro Suffrage in Georgia, 1867-1930, p. 69 (unpublished 1932)).

Faced with a political situation scarcely open to self-correction — disenfranchised blacks had no electoral influence, hence no muscle to lobby the legislature for change— the Court intervened. It invalidated white primaries, see Smith v. Allwright, 321 U. S. 649 (1944), and other burdens on minority voting. See, e. g., Schnell v. Davis, 336 U. S. 933 (1949) (per curiam) (discriminatory application of voting tests); Lane v. Wilson, 307 U. S. 268 (1939) (procedural hurdles); Guinn v. United States, 238 U. S. 347 (1915) (grandfather clauses).

It was against this backdrop that the Court, construing the Equal Protection Clause, undertook to ensure that ap*938portionment plans do not dilute minority voting strength. See, e. g., Rogers v. Lodge, 458 U. S. 613, 617 (1982); Regester, 412 U. S., at 765; Wright v. Rockefeller, 376 U. S. 52, 57 (1964). By enacting the Voting Rights Act of 1965, Congress heightened federal judicial involvement in apportionment, and also fashioned a role for the Attorney General. Section 2 creates a federal right of action to challenge vote dilution. Section 5 requires States with a history of discrimination to preclear any changes in voting practices with either a federal court (a three-judge United States District Court for the District of Columbia) or the Attorney General.

These Court decisions and congressional directions significantly reduced voting discrimination against minorities. In the 1972 election, Georgia gained its first black Member of Congress since Reconstruction, and the 1981 apportionment created the State’s first majority-minority district.1 This voting district, however, was not gained easily. Georgia created it only after the United States District Court for the District of Columbia refused to preclear a predecessor apportionment plan that included no such district — an omission due in part to the influence of Joe Mack Wilson, then Chairman of the Georgia House Reapportionment Committee. As Wilson put it only 14 years ago, “ T don’t want to draw nigger districts.’” Busbee v. Smith, 549 F. Supp. 494, 501 (DC 1982).

a

>

Before Shaw v. Reno, 509 U. S. 630 (1993), this Court invoked the Equal Protection Clause to justify intervention in the quintessential^ political task of legislative districting in two circumstances: to enforce the one-person-one-vote requirement, see Reynolds v. Sims, 377 U. S. 533 (1964); and *939to prevent dilution of a minority group’s voting strength, see Regester, 412 U. S., at 765; Wright, 376 U. S., at 57.2

In Shaw, the Court recognized a third basis for an equal protection challenge to a State’s apportionment plan. The Court wrote cautiously, emphasizing that judicial intervention is exceptional: Strict judicial scrutiny is in order, the Court declared, if a district is “so extremely irregular on its face that it rationally can be viewed only as an effort to segregate the races for purposes of voting.” 509 U. S., at 642.

“[E]xtrem[e] irregular[ity]” was evident in Shaw, the Court explained, setting out this description of the North Carolina voting district under examination:

“It is approximately 160 miles long and, for much of its length, no wider than the I-85 corridor. It winds in snakelike fashion through tobacco country, financial centers, and manufacturing areas until it gobbles in enough enclaves of black neighborhoods. Northbound and southbound drivers on I-85 sometimes find themselves in separate districts in one county, only to ‘trade’ districts when they enter the next county. Of the 10 counties through which District 12 passes, 5 are cut into 3 different districts; even towns are divided. At one point the district remains contiguous only because it intersects at a single point with two other districts before crossing over them. One state legislator has remarked that ‘ “[i]f you drove down the interstate with both car *940doors open, you’d kill most of the people in the district.” ’ Washington Post, Apr. 20, 1993, p. A4. The district even has inspired poetry: Ask not for whom the line is drawn; it is drawn to avoid thee.’ Grofman, Would Vince Lombardi Have Been Right If He Had Said: 'When It Comes to Redistricting, Race Isn’t Everything, It’s the Only Thing’?, 14 Cardozo L. Rev. 1237, 1261, n. 96 (1993) (internal quotation marks omitted).” Id., at 635-636 (some citations and internal quotation marks omitted).

The problem in Shaw was not the plan architects’ consideration of race as relevant in redistricting. Rather, in the Court’s estimation, it was the virtual exclusion of other factors from the calculus. Traditional districting practices were cast aside, the Court concluded, with race alone steering placement of district lines.

B

The record before us does not show that race similarly overwhelmed traditional districting practices in Georgia. Although the Georgia General Assembly prominently considered race in shaping the Eleventh District, race did not crowd out all other factors, as the Court found it did in North Carolina’s delineation of the Shaw district.

In contrast to the snake-like North Carolina district inspected in Shaw, Georgia’s Eleventh District is hardly “bizarre,” “extremely irregular,” or “irrational on its face.” Id., at 642, 644,658. Instead, the Eleventh District’s design reflects significant consideration of “traditional districting factors (such as keeping political subdivisions intact) and the usual political process of compromise and trades for a variety of nonracial reasons.” 864 F. Supp. 1354, 1397, n. 5 (SD Ga. 1994) (Edmondson, J., dissenting); cf. ante, at 917 (“geometric shape of the Eleventh District may not seem bizarre on its face”). The district covers a core area in central and east-*941era Georgia, and its total land area of 6,780 square miles is about average for the State. Defendant’s Exh. 177, p. 4.3 The border of the Eleventh District runs 1,184 miles, in line with Georgia’s Second District, which has a 1,243-mile border, and the State’s Eighth District, with a border running 1,155 miles. See 864 F. Supp., at 1396 (Edmondson, J., dissenting).4

Nor does the Eleventh District disrespect the boundaries of political subdivisions. Of the 22 counties in the district, 14 are intact and 8 are divided. See Joint Exh. 17. That puts the Eleventh District at about the state average in divided counties. By contrast, of the Sixth District’s five counties, none are intact, ibid., and of the Fourth District’s four counties, just one is intact. Ibid.5 Seventy-one percent of the Eleventh District’s boundaries track the borders of political subdivisions. See 864 F. Supp., at 1396 (Edmond-son, J., dissenting). Of the State’s 11 districts, 5 score worse than the Eleventh District on this criterion, and 5 score bet*942ter. See Defendant’s Exh. 177, p. 4.6 Eighty-three percent of the Eleventh District’s geographic area is composed of intact counties, above average for the State’s congressional districts. 864 F. Supp., at 1396 (Edmondson, J., dissenting).7 And notably, the Eleventh District’s boundaries largely follow precinct lines.8

Evidence at trial similarly shows that considerations other than race went into determining the Eleventh District’s boundaries. For a “political reason” — to accommodate the request of an incumbent State Senator regarding the placement of the precinct in which his son lived — the DeKalb County portion of the Eleventh District was drawn to include a particular (largely white) precinct. 2 Tr. 187, 202. The corridor through Effingham County was substantially narrowed at the request of a (white) State Representative. 2 Tr. 189-190, 212-214. In Chatham County, the district was trimmed to exclude a heavily black community in Garden City because a State Representative wanted to keep the city intact inside the neighboring First District. 2 Tr. 218-219. The Savannah extension was configured by “the narrowest means possible” to avoid splitting the city of Port Wentworth. 4 Tr. 172-174, 175-178, 181-183.

*943Georgia’s Eleventh District, in sum, is not an outlier district shaped without reference to familiar districting techniques. Tellingly, the district that the Court’s decision today unsettles is not among those on a statistically calculated list of the 28 most bizarre districts in the United States, a study prepared in the wake of our decision in Shaw. See Pildes & Niemi, 92 Mich. L. Rev., at 565.

C

The Court suggests that it was not Georgia’s Legislature, but the U. S. Department of Justice, that effectively drew the lines, and that Department officers did so with nothing but race in mind. Yet the “Max-Black” plan advanced by the Attorney General was not the plan passed by the Georgia General Assembly.9 See 864 F. Supp., at 1396-1397, n. 5 (Edmondson, J., dissenting) (“The Max-Black plan did influence to some degree the shape of the ultimate Eleventh District .... [But] the actual Eleventh is not identical to the Max-Black plan. The Eleventh, to my eye, is significantly different in shape in many ways. These differences show . . . consideration of other matters beyond race .. . .”).10

And although the Attorney General refused preclearance to the first two plans approved by Georgia’s Legislature, the State was not thereby disarmed; Georgia could have demanded relief from the Department’s objections by instituting a civil action in the United States District Court for the District of Columbia, with ultimate review in this Court. Instead of pursuing that avenue, the State chose to adopt the plan here in controversy — a plan the State forcefully defends *944before us. We should respect Georgia’s choice by taking its position on brief as genuine.

D

Along with attention to size, shape, and political subdivisions, the Court recognizes as an appropriate districting principle, “respect for . . . communities defined by actual shared interests.” Ante, at 916. The Court finds no community here, however, because a report in the record showed “fractured political, social, and economic interests within the Eleventh District’s black population.” Ante, at 919.

But ethnicity itself can tie people together, as volumes of social science literature have documented — even people with divergent economic interests. For this reason, ethnicity is a significant force in political life. As stated in a classic study of ethnicity in one city of immigrants:

“[M]any elements — history, family and feeling, interest, formal organizational life — operate to keep much of New York life channeled within the bounds of the ethnic group. . ..

“... The political realm ... is least willing to consider [ethnicity] a purely private affair. . . .

“[Political life itself emphasizes the ethnic character of the city, with its balanced tickets and its special appeals . . . .” N. Glazer & D. Moynihan, Beyond the Melting Pot 19-20 (1963).

See also, e.g., E. Litt, Beyond Pluralism: Ethnic Politics in America 2 (1970) (“[E]thnic forces play a surprisingly persistent role in our politics.”); Ethnic Group Politics, Preface ix (H. Bailey & E. Katz eds. 1969) (“[EJthnic identifications do exist and . . . one cannot really understand the American political process without giving special attention to racial, religious and national minorities.”).

To accommodate the reality of ethnic bonds, legislatures have long drawn voting districts along ethnic lines. Our *945Nation’s cities are full of districts identified by their ethnic character — Chinese, Irish, Italian, Jewish, Polish, Russian, for example. See, e.g., S. Erie, Rainbow’s End: Irish-Americans and the Dilemmas of Urban Machine Politics, 1840-1985, p. 91 (1988) (describing Jersey City’s “Horseshoe district” as “lumping most of the city’s Irish together”); Coveted Landmarks Add a Twist to Redistricting Task, Los Angeles Times, Sept. 10, 1991, pp. Al, A24 (“In San Francisco in 1961,... an Irish Catholic [State Assembly member] 'wanted his district drawn following [Catholic] parish lines so all the parishes where he went to baptisms, weddings and funerals would be in his district’ . . . .”); Stone, Goode: Bad and Indifferent, Washington Monthly, July-Aug. 1986, pp. 27, 28 (discussing “The Law of Ethnic Loyalty— ... a universal law of politics,” and identifying “predominantly Italian wards of South Philadelphia,” a “Jewish Los Angeles district,” and a “Polish district in Chicago”). The creation of ethnic districts reflecting felt identity is not ordinarily viewed as offensive or demeaning to those included in the delineation.

Ill

To separate permissible and impermissible use of race in legislative apportionment, the Court orders strict scrutiny for districting plans “predominantly motivated” by race. No longer can a State avoid judicial oversight by giving — as in this case — genuine and measurable consideration to traditional districting practices. Instead, a federal case can be mounted whenever plaintiffs plausibly allege that other factors carried less weight than race. This invitation to litigate against the State seems to me neither necessary nor proper.

A

The Court derives its test from diverse opinions on the relevance of race in contexts distinctly unlike apportionment. *946See ante, at 911-912.11 The controlling idea, the Court says, is “ ‘the simple command [at the heart of the Constitution’s guarantee of equal protection] that the Government must treat citizens as individuals, not as simply components of a racial, religious, sexual or national class.’ ” See ante, at 911 (quoting Metro Broadcasting, Inc. v. FCC, 497 U. S. 547, 602 (1990) (O’Connor, J., dissenting)) (some internal quotation marks omitted). But cf. Strauder v. West Virginia, 100 U. S. 303, 307 (1880) (pervading purpose of post-Civil War Amendments was to bar discrimination against once-enslaved race).

*947In adopting districting plans, however, States do not treat people as individuals. Apportionment schemes, by their very nature, assemble people in groups. States do not assign voters to districts based on merit or achievement, standards States might use in hiring employees or engaging contractors. Rather, legislators classify voters in groups— by economic, geographical, political, or social characteristics — and then “reconcile the competing claims of [these] groups.” Davis v. Bandemer, 478 U. S. 109, 147 (1986) (O’Connor, J., concurring in judgment).

That ethnicity defines some of these groups is a political reality. Until now, no constitutional infirmity has been seen in districting Irish or Italian voters together, for example, so long as the delineation does not abandon familiar apportionment practices. See supra, at 944-945. If Chinese-Americans and Russian-Americans may seek and secure group recognition in the delineation of voting districts, then African-Americans should not be dissimilarly treated. Otherwise, in the name of equal protection, we would shut out “the very minority group whose history in the United States gave birth to the Equal Protection Clause.” See Shaw, 509 U. S., at 679 (Stevens, J., dissenting).12

B

Under the Court’s approach, judicial review of the same intensity, i. e., strict scrutiny, is in order once it is determined that an apportionment is predominantly motivated by race. It matters not at all, in this new regime, whether the apportionment dilutes or enhances minority voting strength. As very recently observed, however, “[t]here is no moral or *948constitutional equivalence between a policy that is designed to perpetuate a caste system and one that seeks to eradicate racial subordination.” Adarand Constructors, Inc. v. Peña, ante, at 243 (Stevens, J., dissenting).

Special circumstances justify vigilant judicial inspection to protect minority voters — circumstances that do not apply to majority voters. A history of exclusion from state politics left racial minorities without clout to extract provisions for fair representation in the lawmaking forum. See supra, at 936-938. The equal protection rights of minority voters thus could have remained unrealized absent the Judiciary’s close surveillance. Cf. United States v. Carolene Products Co., 304 U. S. 144, 153, n. 4 (1938) (referring to the “more searching judicial inquiry” that may properly attend classifications adversely affecting “discrete and insular minorities”). The majority, by definition, encounters no such blockage. White voters in Georgia do not lack means to exert strong pressure on their state legislators. The force of their numbers is itself a powerful determiner of what the legislature will do that does not coincide with perceived majority interests.

State legislatures like Georgia’s today operate under federal constraints imposed by the Voting Rights Act — constraints justified by history and designed by Congress to make once-subordinated people free and equal citizens. But these federal constraints do not leave majority voters in need of extraordinary judicial solicitude. The Attorney General, who administers the Voting Rights Act’s preclearance requirements, is herself a political actor. She has a duty to enforce the law Congress passed, and she is no doubt aware of the political cost of venturing too far to the detriment of majority voters. Majority voters, furthermore, can press the State to seek judicial review if the Attorney General refuses to preclear a plan that the voters favor. Finally, the Act is itself a political measure, subject to modification in the political process.

*949C

The Court’s disposition renders redistricting perilous work for state legislatures. Statutory mandates and political realities may require States to consider race when drawing district lines. See supra, at 935. But today’s decision is a counterforce; it opens the way for federal litigation if “traditional ... districting principles” arguably were accorded less weight than race. See ante, at 916. Genuine attention to traditional districting practices and avoidance of bizarre configurations seemed, under Shaw, to provide a safe harbor. See 509 U. S., at 647 (“[Traditional districting principles such as compactness, contiguity, and respect for political subdivisions . . . are objective factors that may serve to defeat a claim that a district has been gerrymandered on racial lines.”). In view of today’s decision, that is no longer the case.

Only after litigation — under either the Voting Rights Act, the Court’s new Miller standard, or both — will States now be assured that plans conscious of race are safe. Federal judges in large numbers may be drawn into the fray. This enlargement of the judicial role is unwarranted. The reapportionment plan that resulted from Georgia’s political process merited this Court’s approbation, not its condemnation. Accordingly, I dissent.

[Appendixes A and B, containing maps of Georgia’s proposed and current Eleventh Districts, and Appendix C, containing a map of the Shaw v. Reno District, follow this page.]

*0[[Image here]]

[[Image here]]