VIETH v. JUBELIRER 1769 - Brennan Center

1769VIETH v. JUBELIRERCite as 124 S.Ct. 1769 (2004)

541 U.S. 267

standing alone. They call attention to un-tidy details, and rightly understood legisla-tion can be untidy: statutes can be un-systematic, redundant, and fuzzy aboutdrawing lines. As a purely textual matter,both the majority’s reading and mine havestrengths and weaknesses. The point isthat the tie breakers cut in favor of sus-taining the South Coast Fleet Rules. Myreading adheres more closely to the legis-lative history of § 209(a). It takes properaccount of the fact that the Fleet Ruleswith this commercial availability conditiondo not require manufacturers, even indi-rectly, to produce a new kind of engine.And, most importantly, my reading ad-heres to the well-established presumptionagainst preemption.

,

541 U.S. 267, 158 L.Ed.2d 546

Richard VIETH, Norma Jean Vieth,and Susan Furey, Appellants,

v.

Robert C. JUBELIRER, President ofthe Pennsylvania Senate, et al.

No. 02–1580.

Argued Dec. 10, 2003.

Decided April 28, 2004.

Background: Voters brought actionagainst Commonwealth of Pennsylvania,challenging constitutionality of congres-sional redistricting plan. Following dis-missal of other claims, 188 F.Supp.2d 532,three-judge panel of the United StatesDistrict Court for the Middle District ofPennsylvania, 195 F.Supp.2d 672, deter-

mined that plan did not satisfy one person-one vote requirement. Following enact-ment of new plan and renewal of constitu-tional challenge, a three-judge panel of theDistrict Court, 241 F.Supp.2d 478, ap-proved the plan, and voters appealed.

Holding: The Supreme Court, JusticeScalia, held that political gerrymanderingclaims were nonjusticiable.

Affirmed.

Justice Kennedy filed opinion concurringin the judgment.

Justice Stevens filed dissenting opinion.

Justice Souter filed dissenting opinion inwhich Justice Ginsburg joined.

Justice Breyer filed dissenting opinion.

Federal Courts O13.20

Claim of political gerrymandering incongressional redistricting plan was non-justiciable, because there were no judicial-ly discernible and manageable standardsfor adjudicating such a claim. (Per JusticeScalia with the Chief Justice and two Jus-tices joining, and one Justice joining inresult).

Syllabus *

After Pennsylvania’s General Assem-bly adopted a congressional redistrictingplan, plaintiffs-appellants sued to enjointhe plan’s implementation, alleging, interalia, that it constituted a political gerry-mander in violation of Article I and theFourteenth Amendment’s Equal Protec-tion Clause. The three-judge DistrictCourt dismissed the gerrymanderingclaim, and the plaintiffs appealed.

Held: The judgment is affirmed.

241 F.Supp.2d 478, affirmed.

* The syllabus constitutes no part of the opinionof the Court but has been prepared by theReporter of Decisions for the convenience of

the reader. See United States v. Detroit Tim-ber & Lumber Co., 200 U.S. 321, 337, 26 S.Ct.282, 50 L.Ed. 499.

1770 124 SUPREME COURT REPORTER 541 U.S. 267

Justice SCALIA, joined by THECHIEF JUSTICE, Justice O’CONNOR,and Justice THOMAS, concluded that po-litical gerrymandering claims are nonjusti-ciable because no judicially discernible andmanageable standards for adjudicatingsuch claims exist. They would thereforeoverrule Davis v. Bandemer, 478 U.S. 109,106 S.Ct. 2797, 92 L.Ed.2d 85, in whichthis Court held that political gerrymander-ing claims are justiciable, but could notagree upon a standard for assessing politi-cal gerrymandering claims. Pp. 1774–1792.

(a) Political gerrymanders existed incolonial times and continued through theframing. The Framers provided a reme-dy for the problem: The Constitutiongives state legislatures the initial power todraw federal election districts, but author-izes Congress to ‘‘make or alter’’ thosedistricts. U.S. Const., Art. I, § 4. In Ban-demer, the Court held that the Equal Pro-tection Clause also grants judges the pow-er—and duty—to control that practice.Pp. 1774–1776.

(b) Neither Art. I, § 2 or § 4, nor theEqual Protection Clause, provides a judi-cially enforceable limit on the political con-siderations that the States and Congressmay take into account when districting.Pp. 1776–1784.

(1) Among the tests for determiningthe existence of a ‘‘nonjusticiable’’ or ‘‘po-litical’’ question is a lack of judicially dis-coverable and manageable standards forresolving the question. Baker v. Carr, 369U.S. 186, 217, 82 S.Ct. 691, 7 L.Ed.2d 663.Because the Bandemer Court was ‘‘notpersuaded’’ that there are no such stan-dards for deciding political gerrymander-ing cases, 478 U.S., at 123, 106 S.Ct. 2797,such cases were justiciable. However, thesix-Justice majority in Bandemer could notdiscern what the standards might be. Forthe past 18 years, the lower courts have

simply applied the Bandemer S 268plurality’sstandard, almost invariably producing thesame result as would have obtained hadthe question been nonjusticiable: Judicialintervention has been refused. Eighteenyears of judicial effort with virtually noth-ing to show for it justifies revisiting wheth-er the standard promised by Bandemerexists. Pp. 1776–1778.

(2) The Bandemer plurality’s stan-dard—that a political gerrymanderingclaim can succeed only where the plaintiffsshow ‘‘both intentional discriminationagainst an identifiable political group andan actual discriminatory effect on thatgroup,’’ 478 U.S., at 127, 106 S.Ct. 2797—has proved unmanageable in application.Because that standard was misguidedwhen proposed, has not been improved insubsequent application, and is not evendefended by the appellants in this Court, itshould not be affirmed as a constitutionalrequirement. Pp. 1778–1780.

(3) Appellants’ proposed two-prongedstandard based on Art. I, § 2, and theEqual Protection Clause is neither discer-nible nor manageable. Appellants aremistaken when they contend that theirintent prong (‘‘predominant intent’’) is nodifferent from that which this Court hasapplied in racial gerrymandering cases.In those cases, the predominant intent testis applied to the challenged district inwhich the plaintiffs voted, see, e.g., Millerv. Johnson, 515 U.S. 900, 115 S.Ct. 2475,132 L.Ed.2d 762, whereas here appellantsassert that their test is satisfied only whenpartisan advantage was the predominantmotivation behind the entire statewideplan. Vague as a predominant motivationtest might be when used to evaluate singledistricts, it all but evaporates when appliedstatewide. For this and other reasons, theracial gerrymandering cases provide nocomfort. The effects prong of appellants’proposal requires (1) that the plaintiffs


541 U.S. 269

show that the rival party’s voters are sys-tematically ‘‘packed’’ or ‘‘cracked’’; and (2)that the court be persuaded from the total-ity of the circumstances that the map canthwart the plaintiffs’ ability to translate amajority of votes into a majority of seats.This standard is not discernible becausethe Constitution provides no right to pro-portional representation. Even were thestandard discernible, it is not judiciallymanageable. There is no effective way toascertain a party’s majority status, and, inany event, majority status in statewideraces does not establish majority status forparticular district contests. Moreover,even if a majority party could be identi-fied, it would be impossible to ensure thatit won a majority of seats unless theStates’ traditional election structures wereradically revised. Pp. 1780–1784.

(4) For many of the same reasons,Justice Powell’s Bandemer standard—a to-tality-of-the-circumstances analysis thatevaluates districts with an eye to ascer-taining whether the particular gerryman-der is not ‘‘fair’’—must also be rejected.‘‘Fairness’’ is not a judicially manageablestandard. Some criterion more solid andmore demonstrably S 269met than that isnecessary to enable state legislatures todiscern the limits of their districting dis-cretion, to meaningfully constrain thecourts’ discretion, and to win public accep-tance for the courts’ intrusion into a pro-cess that is the very foundation of demo-cratic decisionmaking. Pp. 1784.

(c) Writing separately in dissent, Jus-tices STEVENS, SOUTER, and BREYEReach propose a different standard for adju-dicating political gerrymandering claims.These proposed standards each have theirown deficiencies, but additionally fail forreasons identified with respect to the stan-dards proposed by appellants and thoseproposed in Bandemer. Justice KENNE-DY concurs in the judgment, recognizingthat there are no existing manageable

standards for measuring whether a politi-cal gerrymander burdens the representa-tional rights of a party’s voters. Pp. 1784–1792.

(d) Stare decisis does not require thatBandemer be allowed to stand. Stare de-cisis claims are at their weakest with re-spect to a decision interpreting the Consti-tution, particularly where there has beenno reliance on that decision. P. 1792.

Justice KENNEDY, while agreeingthat appellants’ complaint must be dis-missed, concluded that all possibility ofjudicial relief should not be foreclosed incases such as this because a limited andprecise rationale may yet be found to cor-rect an established constitutional violation.Courts confront two obstacles when pre-sented with a claim of injury from partisangerrymandering. First is the lack of com-prehensive and neutral principles fordrawing electoral boundaries. No sub-stantive definition of fairness in districtingcommands general assent. Second is theabsence of rules to limit and confine judi-cial intervention. That courts can grantrelief in districting cases involving racedoes not answer the need for fairness prin-ciples, since those cases involve sortingpermissible districting classifications fromimpermissible ones. Politics is a differentmatter. Gaffney v. Cummings, 412 U.S.735, 93 S.Ct. 2321, 37 L.Ed.2d 298. Adetermination that a gerrymander violatesthe law must rest on something more thanthe conclusion that political classificationswere applied. It must rest instead on aconclusion that the classifications, thoughgenerally permissible, were applied in aninvidious manner or in a way unrelated toany legitimate legislative objective. Theobject of districting is to establish ‘‘fairand effective representation for all citi-zens.’’ Reynolds v. Sims, 377 U.S. 533, 84S.Ct. 1362, 12 L.Ed.2d 506. It might seemthat courts could determine, by the exer-


cise of their judgment, whether politicalclassifications are related to this object orinstead burden representational rights.The lack, however, of any agreed uponmodel of fair and effective representationmakes the analysis difficult. With noagreed upon substantive principles of fairdistricting, there is no basis on which todefine clear, manageSable,270 and politicallyneutral standards for measuring the bur-den a given partisan classification imposeson representational rights. Suitable stan-dards for measuring this burden are criti-cal to our intervention. In this case, theplurality convincingly demonstrates thatthe standards proposed in Davis v. Ban-demer, 478 U.S. 109, 106 S.Ct. 2797, 92L.Ed.2d 85, by the parties here, and by thedissents are either unmanageable or incon-sistent with precedent, or both. Thereare, then, weighty arguments for holdingcases like these to be nonjusticiable.However, they are not so compelling thatthey require the Court now to bar allfuture partisan gerrymandering claims.Baker v. Carr, 369 U.S. 186, 82 S.Ct. 691, 7L.Ed.2d 663, makes clear that the moreabstract standards that guide analysis ofall Fourteenth Amendment claims sufficeto ensure justiciability of claims like these.That a workable standard for measuring agerrymander’s burden on representationalrights has not yet emerged does not meanthat none will emerge in the future. TheCourt should adjudicate only what is in thecase before it. In this case, absent astandard by which to measure the burdenappellants claim has been imposed on theirrepresentational rights, appellants’ evi-dence at best demonstrates only that thelegislature adopted political classifications.That describes no constitutional flaw underthe governing Fourteenth Amendmentstandard. Gaffney, supra, at 752, 93 S.Ct.2321. While the equal protection standardcontinues to govern such cases, the FirstAmendment may prove to offer a sounder

and more prudential basis for judicial in-tervention in political gerrymanderingcases. First Amendment analysis does notdwell on whether a generally permissibleclassification has been used for an imper-missible purpose, but concentrates onwhether the legislation burdens the repre-sentational rights of the complaining par-ty’s voters for reasons of ideology, beliefs,or political association. That analysis al-lows a pragmatic or functional assessmentthat accords some latitude to the States.See, e.g., Eu v. San Francisco CountyDemocratic Central Comm., 489 U.S. 214,109 S.Ct. 1013, 103 L.Ed.2d 271. Pp.1792–1798.

SCALIA, J., announced the judgmentof the Court and delivered an opinion, inwhich REHNQUIST, C.J., andO’CONNOR and THOMAS, JJ., joined.KENNEDY, J., filed an opinionconcurring in the judgment, post, p. 1792.STEVENS, J., filed a dissenting opinion,post, p. 1799. SOUTER, J., filed adissenting opinion, in which GINSBURG,J., joined, post, p. 1815. BREYER, J.,filed a dissenting opinion, post, p. 1822.

Robert B. Hoffman, Reed, Smith LLP,Harrisburg, PA, Paul M. Smith, Counsel ofRecord, Thomas J. Perrelli, Bruce V. Spi-va, Sam Hirsch, Daniel Mach, Jenner &Block, LLC, Washington, DC, for appel-lants.

John P. Krill, Jr., Counsel of Record,Linda J. Shorey, Julia M. Glencer, Kirk-patrick & Lockhart LLP, Harrisburg,Pennsylvania, for Appellees Jubelirer &Perzel.

D. Michael Fisher, Attorney General,Commonwealth of Pennsylvania, John G.Knorr, III, Chief Deputy Attorney Gener-al, J. Bart DeLone, Counsel of Record,Senior Deputy Attorney General, Harris-burg, PA, for Appellees Cortes and Accur-ti, Office of Attorney General.


541 U.S. 273

Paul M. Smith, Counsel of Record,Thomas J. Perrelli, Bruce V. Spiva, SamHirsch, Daniel Mach, Jenner & Block,LLC, Washington, DC, Robert B. Hoff-man, Reed, Smith LLP, Harrisburg, PA,for Appellants.

For U.S. Supreme Court briefs, see:

2003 WL 22811823 (Reply.Brief)

2003 WL 22070244 (Appellant.Brief)

2003 WL 22439888 (Appellee.Brief)

2003 WL 22439889 (Appellee.Brief)

Justice SCALIA announced thejudgment of the Court and delivered anopinion, in which THE CHIEF JUSTICE,Justice O’CONNOR, and JusticeTHOMAS join.

S 271Plaintiffs-appellants Richard Vieth,Norma Jean Vieth, and Susan Furey chal-lenge a map drawn by the PennsylvaniaGeneral Assembly establishing districts forthe election of congressional Representa-tives, on the ground that the districtingconstitutes an unconstitutional politicalgerrymander.1 In Davis v. Bandemer, 478U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85(1986), this Court held that political gerry-mandering claims are justiciable, butS 272could not agree upon a standard to ad-judicate them. The present appeal pres-ents the questions whether our decision inBandemer was in error, and, if not, whatthe standard should be.

I

The facts, as alleged by the plaintiffs,are as follows. The population figures de-rived from the 2000 census showed thatPennsylvania was entitled to only 19 Rep-resentatives in Congress, a decrease in 2from the Commonwealth’s previous dele-

gation. Pennsylvania’s General Assemblytook up the task of drawing a new district-ing map. At the time, the RepublicanParty controlled a majority of both stateHouses and held the Governor’s office.Prominent national figures in the Republi-can Party pressured the General Assemblyto adopt a partisan redistricting plan as apunitive measure against Democrats forhaving enacted pro-Democrat redistrictingplans elsewhere. The Republican mem-bers of Pennsylvania’s House and Senateworked together on such a plan. On Janu-ary 3, 2002, the General Assembly passedits plan, which was signed into law byGovernor Schweiker as Act 1.

Plaintiffs, registered Democrats whovote in Pennsylvania, brought suit in theUnited States District Court for the Mid-dle District of Pennsylvania, seeking toenjoin implementation of Act 1 under Rev.Stat. § 1979, 42 U.S.C. § 1983. Defen-dants-appellees were the Commonwealthof Pennsylvania and various executive andlegislative officers responsible for enactingor implementing Act 1. The complaint al-leged, among other things, that the legisla-tion created malapportioned districts, inviolation of the one-person, one-vote re-quirement of Article I, § 2, of the UnitedStates Constitution, and that it constituteda political gerrymander, in violation of Ar-ticle I and the Equal Protection Clause ofthe Fourteenth Amendment. With regardto the latter contention, the complaint al-leged that the districts created by Act 1were ‘‘meandering and irregular,’’ and ‘‘ig-nor[ed] all traditional redistricting criteria,including the preservation of S 273local gov-ernment boundaries, solely for the sake ofpartisan advantage.’’ Juris. Statement136a, ¶ 22, 135a, ¶ 20.

1. The term ‘‘political gerrymander’’ has beendefined as ‘‘[t]he practice of dividing a geo-graphical area into electoral districts, often ofhighly irregular shape, to give one political

party an unfair advantage by diluting the op-position’s voting strength.’’ Black’s Law Dic-tionary 696 (7th ed.1999).


A three-judge panel was convened pur-suant to 28 U.S.C. § 2284. The defen-dants moved to dismiss. The DistrictCourt granted the motion with respect tothe political gerrymandering claim, and (onEleventh Amendment grounds) all claimsagainst the Commonwealth; but it de-clined to dismiss the apportionment claimas to other defendants. See Vieth v.Pennsylvania, 188 F.Supp.2d 532 (M.D.Pa.2002) (Vieth I). On trial of the apportion-ment claim, the District Court ruled infavor of plaintiffs. See Vieth v. Pennsyl-vania, 195 F.Supp.2d 672 (M.D.Pa.2002)(Vieth II). It retained jurisdiction overthe case pending the court’s review andapproval of a remedial redistricting plan.On April 18, 2002, Governor Schweikersigned into law Act No. 2002–34, Pa. Stat.Ann., Tit. 25, § 3595.301 (Purdon Supp.2003) (Act 34), a remedial plan that thePennsylvania General Assembly had enact-ed to cure the apportionment problem ofAct 1.

Plaintiffs moved to impose remedial dis-tricts, arguing that the District Courtshould not consider Act 34 to be a properremedial scheme, both because it was ma-lapportioned, and because it constitutedan unconstitutional political gerrymanderlike its predecessor. The District Courtdenied this motion, concluding that thenew districts were not malapportioned,and rejecting the political gerrymanderingclaim for the reasons previously assignedin Vieth I. Vieth v. Pennsylvania, 241F.Supp.2d 478, 484–485 (M.D.Pa.2003)(Vieth III). The plaintiffs appealed thedismissal of their Act 34 political gerry-mandering claim.2 We noted probable ju-risdiction. 539 U.S. 957, 123 S.Ct. 2652,156 L.Ed.2d 654 (2003).

S 274II

Political gerrymanders are not new tothe American scene. One scholar tracesthem back to the Colony of Pennsylvaniaat the beginning of the 18th century,where several counties conspired to mini-mize the political power of the city ofPhiladelphia by refusing to allow it tomerge or expand into surrounding jurisdic-tions, and denying it additional representa-tives. See E. Griffith, The Rise and De-velopment of the Gerrymander 26–28(1974) (hereinafter Griffith). In 1732, twomembers of His Majesty’s Council and theattorney general and deputy inspector andcomptroller general of affairs of the Prov-ince of North Carolina reported that theGovernor had proceeded to ‘‘divide oldPrecincts established by Law, & to enactnew Ones in Places, whereby his Arts hehas endeavoured to prepossess People in afuture election according to his desire, hisDesigns herein being TTT either to endeav-our by his means to get a Majority of hiscreatures in the Lower House’’ or to dis-rupt the assembly’s proceedings. 3 Colo-nial Records of North Carolina 380–381(W. Saunders ed. 1886); see also Griffith29. The political gerrymander remainedalive and well (though not yet known bythat name) at the time of the framing.There were allegations that Patrick Henryattempted (unsuccessfully) to gerrymanderJames Madison out of the First Congress.See 2 W. Rives, Life and Times of JamesMadison 655, n. 1 (reprint 1970); Letterfrom Thomas Jefferson to William Short,Feb. 9, 1789, reprinted in 5 Works ofThomas Jefferson 451 (P. Ford ed.1904).And in 1812, of course, there occurred the

2. The plaintiffs apparently never amendedtheir complaint to allege that Act 34 was apolitical gerrymander, yet the District Court’sdecision in Vieth III resolved that claim onthe merits. Because subject-matter jurisdic-

tion is not implicated and neither party hasraised the point, we assume that the DistrictCourt deemed the plaintiffs’ original com-plaint to have been constructively amended.


541 U.S. 276

notoriously outrageous political districtingin Massachusetts that gave the gerryman-der its name—an amalgam of the names ofMassachusetts Governor Elbridge Gerryand the creature (‘‘salamander’’) which theoutline of an election district he was credit-ed with forming was thought to resemble.See Webster’s New International Dictio-nary 1052 (2d ed.1945). ‘‘By 1840 the ger-rymander was a recognized force in partypolitics and was generally attempted in alllegislation S 275enacted for the formation ofelection districts. It was generally con-ceded that each party would attempt togain power which was not proportionate toits numerical strength.’’ Griffith 123.

It is significant that the Framers provid-ed a remedy for such practices in theConstitution. Article I, § 4, while leavingin state legislatures the initial power todraw districts for federal elections, permit-ted Congress to ‘‘make or alter’’ thosedistricts if it wished.3 Many objected tothe congressional oversight established bythis provision. In the course of the de-bates in the Constitutional Convention,Charles Pinckney and John Rutledgemoved to strike the relevant language.James Madison responded in defense ofthe provision that Congress must be giventhe power to check partisan manipulationof the election process by the States:

‘‘Whenever the State Legislatures had afavorite measure to carry, they wouldtake care so to mould their regulationsas to favor the candidates they wished tosucceed. Besides, the inequality of theRepresentation in the Legislatures ofparticular States, would produce a likeinequality in their representation in theNatl. Legislature, as it was presumablethat the Counties having the power in

the former case would secure it to them-selves in the latter. What danger couldthere be in giving a controuling power tothe Natl. Legislature?’’ 2 Records ofthe Federal Convention of 1787, pp.240–241 (M. Farrand ed.1911).

Although the motion of Pinckney and Rut-ledge failed, opposition to the ‘‘make oralter’’ provision of Article I, § 4—and thedefense that it was needed to prevent po-litical gerrymanSdering276—continued to bevoiced in the state ratifying debates. Adelegate to the Massachusetts conventionwarned that state legislatures

‘‘might make an unequal and partial di-vision of the states into districts for theelection of representatives, or theymight even disqualify one third of theelectors. Without these powers in Con-gress, the people can have no remedy;but the 4th section provides a remedy, acontrolling power in a legislature, com-posed of senators and representatives oftwelve states, without the influence ofour commotions and factions, who willhear impartially, and preserve and re-store to the people their equal and sa-cred rights of election.’’ 2 Debates onthe Federal Constitution 27 (J. Elliot 2ded. 1876).

The power bestowed on Congress toregulate elections, and in particular to re-strain the practice of political gerryman-dering, has not lain dormant. In the Ap-portionment Act of 1842, 5 Stat. 491,Congress provided that Representativesmust be elected from single-member dis-tricts ‘‘composed of contiguous territory.’’See Griffith 12 (noting that the law was‘‘an attempt to forbid the practice of thegerrymander’’). Congress again imposedthese requirements in the Apportionment

3. Article I, § 4, provides as follows:‘‘The Times, Places and Manner of holding

Elections for Senators and Representatives,shall be prescribed in each State by the Legis-

lature thereof; but the Congress may at anytime by Law make or alter such Regulations,except as to the Places of chusing Senators.’’


Act of 1862, 12 Stat. 572, and in 1872further required that districts ‘‘contai[n]as nearly as practicable an equal numberof inhabitants,’’ 17 Stat. 28, § 2. In theApportionment Act of 1901, Congress im-posed a compactness requirement. 31Stat. 733. The requirements of contigu-ity, compactness, and equality of popula-tion were repeated in the 1911 apportion-ment legislation, 37 Stat. 13, but were notthereafter continued. Today, only thesingle-member-district-requirement re-mains. See 2 U.S.C. § 2c. Recent histo-ry, however, attests to Congress’s aware-ness of the sort of districting practicesappellants protest, and of its power underArticle I, § 4, to control them. Since1980, no fewer than five bills have beenintroduced to regulate gerrySmandering277

in congressional districting. See H.R.5037, 101st Cong., 2d Sess. (1990); H.R.1711, 101st Cong., 1st Sess. (1989); H.R.3468, 98th Cong., 1st Sess. (1983); H.R.5529, 97th Cong., 2d Sess. (1982); H.R.2349, 97th Cong., 1st Sess. (1981).4

Eighteen years ago, we held that theEqual Protection Clause grants judges thepower—and duty—to control political ger-rymandering, see Davis v. Bandemer, 478U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85(1986). It is to consideration of this prece-dent that we now turn.

III

As Chief Justice Marshall proclaimedtwo centuries ago, ‘‘[i]t is emphatically theprovince and duty of the judicial depart-ment to say what the law is.’’ Marbury v.Madison, 1 Cranch 137, 177, 2 L.Ed. 60(1803). Sometimes, however, the law isthat the judicial department has no busi-

ness entertaining the claim of unlawful-ness—because the question is entrusted toone of the political branches or involves nojudicially enforceable rights. See, e.g.,Nixon v. United States, 506 U.S. 224, 113S.Ct. 732, 122 L.Ed.2d 1 (1993) (challengeto procedures used in Senate impeachmentproceedings); Pacific States Telephone &Telegraph Co. v. Oregon, 223 U.S. 118, 32S.Ct. 224, 56 L.Ed. 377 (1912) (claims aris-ing under the Guaranty Clause of ArticleIV, § 4). Such questions are said to be‘‘nonjusticiable,’’ or ‘‘political questions.’’

In Baker v. Carr, 369 U.S. 186, 82 S.Ct.691, 7 L.Ed.2d 663 (1962), we set forth sixindependent tests for the existence of apolitical question:

‘‘[1] a textually demonstrable constitu-tional commitment of the issue to a coor-dinate political department; or [2] a lackof judicially discoverable and managea-ble standSards278 for resolving it; or [3]the impossibility of deciding without aninitial policy determination of a kindclearly for nonjudicial discretion; or [4]the impossibility of a court’s undertak-ing independent resolution without ex-pressing lack of the respect due coordi-nate branches of government; or [5] anunusual need for unquestioning adher-ence to a political decision already made;or [6] the potentiality of embarrassmentfrom multifarious pronouncements byvarious departments on one question.’’Id., at 217, 82 S.Ct. 691.

These tests are probably listed in descend-ing order of both importance and certain-ty. The second is at issue here, and thereis no doubt of its validity. ‘‘The judicialPower’’ created by Article III, § 1, of the

4. The States, of course, have taken their ownsteps to prevent abusive districting practices.A number have adopted standards for redis-tricting, and measures designed to insulatethe process from politics. See, e.g., IowaCode § 42.4(5) (2003); N.J. Const., Art. II,

§ 2; Haw.Rev.Stat. § 25–2 (1993); IdahoCode § 72–1506 (1948–1999); Me.Rev.Stat.Ann., Tit. 21–A, §§ 1206, 1206–A (West Supp.2003); Mont.Code Ann. § 5–1–115 (2003);Wash. Rev.Code § 44.05.090 (1994).


541 U.S. 279

Constitution is not whatever judges chooseto do, see Valley Forge Christian Collegev. Americans United for Separation ofChurch and State, Inc., 454 U.S. 464, 487,102 S.Ct. 752, 70 L.Ed.2d 700 (1982); cf.Grupo Mexicano de Desarrollo, S.A. v.Alliance Bond Fund, Inc., 527 U.S. 308,332–333, 119 S.Ct. 1961, 144 L.Ed.2d 319(1999), or even whatever Congress choosesto assign them, see Lujan v. Defenders ofWildlife, 504 U.S. 555, 576–577, 112 S.Ct.2130, 119 L.Ed.2d 351 (1992); Chicago &Southern Air Lines, Inc. v. Waterman S.S.Corp., 333 U.S. 103, 110–114, 68 S.Ct. 431,92 L.Ed. 568 (1948). It is the power to actin the manner traditional for English andAmerican courts. One of the most obviouslimitations imposed by that requirement isthat judicial action must be governed bystandard, by rule. Laws promulgated bythe Legislative Branch can be inconsistent,illogical, and ad hoc; law pronounced bythe courts must be principled, rational, andbased upon reasoned distinctions.

Over the dissent of three Justices, theCourt held in Davis v. Bandemer that,since it was ‘‘not persuaded that there areno judicially discernible and manageablestandards by which political gerrymandercases are to be decided,’’ 478 U.S., at 123,106 S.Ct. 2797, such cases were justiciable.The clumsy shifting of the burden of prooffor the premise (the Court was ‘‘not per-suaded’’ that standards do not exist, ratherthan ‘‘persuaded’’ S 279that they do) was ne-cessitated by the uncomfortable fact thatthe six-Justice majority could not discernwhat the judicially discernable standardsmight be. There was no majority on thatpoint. Four of the Justices finding justici-

ability believed that the standard was onething, see id., at 127, 106 S.Ct. 2797 (plu-rality opinion of White, J., joined by Bren-nan, Marshall, and Blackmun, JJ.); twobelieved it was something else, see id., at161, 106 S.Ct. 2797 (Powell, J., joined bySTEVENS, J., concurring in part and dis-senting in part). The lower courts havelived with that assurance of a standard (ormore precisely, lack of assurance thatthere is no standard), coupled with thatinability to specify a standard, for the past18 years. In that time, they have consid-ered numerous political gerrymanderingclaims; this Court has never revisited theunanswered question of what standardgoverns.

Nor can it be said that the lower courtshave, over 18 years, succeeded in shapingthe standard that this Court was initiallyunable to enunciate. They have simplyapplied the standard set forth in Bandem-er’s four-Justice plurality opinion. Thismight be thought to prove that the four-Justice plurality standard has met the testof time—but for the fact that its applica-tion has almost invariably produced thesame result (except for the incurring ofattorney’s fees) as would have obtained ifthe question were nonjusticiable: Judicialintervention has been refused. As onecommentary has put it, ‘‘[t]hroughout itssubsequent history, Bandemer has servedalmost exclusively as an invitation to litiga-tion without much prospect of redress.’’S. Issacharoff, P. Karlan, & R. Pildes, TheLaw of Democracy 886 (rev.2d ed.2002).The one case in which relief was provided(and merely preliminary relief, at that) didnot involve the drawing of district lines; 5

5. See Republican Party of North Carolina v.Martin, 980 F.2d 943 (C.A.4 1992) (upholdingdenial of Federal Rule of Civil Procedure12(b)(6) judgment for the defendants); Repub-lican Party of North Carolina v. North Car-olina State Bd. of Elections, 27 F.3d 563(C.A.4 1994) (unpublished opinion) (uphold-

ing, as modified, a preliminary injunction).Martin dealt with North Carolina’s system ofelecting superior court judges statewide, asystem that had resulted in the election ofonly a single Republican judge since 1900.980 F.2d, at 948. Later developments in thecase are described in n. 8, infra.


in all of the cases we are aware of involv-ing that most S 280common form of politicalgerrymandering, relief was denied.6

Moreover, although the case in which reliefwas provided seemingly involved the neplus ultra of partisan manipulation, see n.5, supra, we would be at a loss to explainwhy the Bandemer line should have beendrawn just there, and should not haveembraced several districting plans thatwere upheld despite allegations of extremepartisan discriminaStion,281 bizarrely shapeddistricts, and disproportionate results.See, e.g., Session v. Perry, 298 F.Supp.2d451 (E.D.Tex.2004) (per curiam); O’Learv. Miller, 222 F.Supp.2d 850 (E.D.Mich.),summarily aff’d, 537 U.S. 997, 123 S.Ct.512, 154 L.Ed.2d 391 (2002); Badham v.Eu, 694 F.Supp. 664, 670 (N.D.Cal.1988),summarily aff’d, 488 U.S. 1024, 109 S.Ct.829, 102 L.Ed.2d 962 (1989). To think thatthis lower court jurisprudence has broughtforth ‘‘judicially discernible and managea-ble standards’’ would be fantasy.

Eighteen years of judicial effort withvirtually nothing to show for it justify us in

revisiting the question whether the stan-dard promised by Bandemer exists. Asthe following discussion reveals, no judi-cially discernible and manageable stan-dards for adjudicating political gerryman-dering claims have emerged. Lackingthem, we must conclude that political ger-rymandering claims are nonjusticiable andthat Bandemer was wrongly decided.

A

We begin our review of possible stan-dards with that proposed by JusticeWhite’s plurality opinion in Bandemer be-cause, as the narrowest ground for ourdecision in that case, it has been the stan-dard employed by the lower courts. Theplurality concluded that a political gerry-mandering claim could succeed only whereplaintiffs showed ‘‘both intentional discrim-ination against an identifiable politicalgroup and an actual discriminatory effecton that group.’’ 478 U.S., at 127, 106 S.Ct.2797. As to the intent element, the plural-ity acknowledged that ‘‘[a]s long as redis-

6. For cases in which courts rejected prayersfor relief under Davis v. Bandemer, 478 U.S.109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986),see, e.g., Duckworth v. State AdministrationBd. of Election Laws, 332 F.3d 769 (C.A.42003); Smith v. Boyle, 144 F.3d 1060 (C.A.71998); La Porte County Republican CentralComm. v. Board of Comm’rs of County of LaPorte, 43 F.3d 1126 (C.A.7 1994); Session v.Perry, 298 F.Supp.2d 451 (E.D.Tex.2004) (percuriam); Martinez v. Bush, 234 F.Supp.2d1275 (S.D.Fla.2002) (three-judge panel);O’Lear v. Miller, 222 F.Supp.2d 850(E.D.Mich.), summarily aff’d, 537 U.S. 997,123 S.Ct. 512, 154 L.Ed.2d 391 (2002); Mary-landers for Fair Representation, Inc. v. Schae-fer, 849 F.Supp. 1022 (D.Md.1994) (three-judge panel); Terrazas v. Slagle, 821 F.Supp.1162 (W.D.Tex.1993) (three-judge panel);Pope v. Blue, 809 F.Supp. 392 (W.D.N.C.)(three-judge panel), summarily aff’d, 506 U.S.801, 113 S.Ct. 30, 121 L.Ed.2d 3 (1992); Illi-nois Legislative Redistricting Comm’n v. LaP-aille, 782 F.Supp. 1272 (N.D.Ill.1992); Fundfor Accurate and Informed Representation, Inc.

v. Weprin, 796 F.Supp. 662 (N.D.N.Y.) (three-judge panel), summarily aff’d, 506 U.S. 1017,113 S.Ct. 650, 121 L.Ed.2d 577 (1992); Hol-loway v. Hechler, 817 F.Supp. 617 (S.D.W.Va.1992) (three-judge panel), summarily aff’d,507 U.S. 956, 113 S.Ct. 1378, 122 L.Ed.2d754 (1993); Hastert v. State Bd. of Elections,777 F.Supp. 634 (N.D.Ill.1991) (three-judgepanel); Anne Arundel County Republican Cen-tral Comm. v. State Administrative Bd. of Elec-tion Laws, 781 F.Supp. 394 (D.Md.1991)(three-judge panel), summarily aff’d, 504 U.S.938, 112 S.Ct. 2269, 119 L.Ed.2d 197 (1992);Republican Party of Virginia v. Wilder, 774F.Supp. 400 (W.D.Va.1991) (three-judge pan-el); Badham v. Eu, 694 F.Supp. 664, 670(N.D.Cal.1988), summarily aff’d, 488 U.S.1024, 109 S.Ct. 829, 102 L.Ed.2d 962 (1989);In re 2003 Legislative Apportionment of Houseof Representatives, 2003 ME 81, 827 A.2d 810;McClure v. Secretary of Commonwealth, 436Mass. 614, 766 N.E.2d 847 (2002); LegislativeRedistricting Cases, 331 Md. 574, 629 A.2d646 (1993); Kenai Peninsula Borough v. State,743 P.2d 1352 (Alaska 1987).


541 U.S. 283

tricting is done by a legislature, it shouldnot be very difficult to prove that thelikely political consequences of the reap-portionment were intended.’’ Id., at 129,106 S.Ct. 2797. However, the effectsprong was significantly harder to satisfy.Relief could not be based merely upon thefact that a group of persons banded to-gether for political purposes had failed toachieve representation commensurate withits numbers, or that the apportionmentscheme made its winning of elections moredifficult. Id., at 132, 106 S.Ct. 2797.Rather, S 282it would have to be shown that,taking into account a variety of historicfactors and projected election results, thegroup had been ‘‘denied its chance to effec-tively influence the political process’’ as awhole, which could be achieved even with-out electing a candidate. Id., at 132–133,106 S.Ct. 2797. It would not be enough toestablish, for example, that Democrats hadbeen ‘‘placed in a district with a superma-jority of other Democratic voters’’ or thatthe district ‘‘departs from pre-existing po-litical boundaries.’’ Id., at 140–141, 106S.Ct. 2797. Rather, in a challenge to anindividual district the inquiry would focus‘‘on the opportunity of members of thegroup to participate in party deliberationsin the slating and nomination of candi-dates, their opportunity to register andvote, and hence their chance to directlyinfluence the election returns and to se-cure the attention of the winning candi-date.’’ Id., at 133, 106 S.Ct. 2797. Astatewide challenge, by contrast, would in-volve an analysis of ‘‘the voters’ direct orindirect influence on the elections of thestate legislature as a whole.’’ Ibid. (em-phasis added). With what has proved tobe a gross understatement, the pluralityacknowledged this was ‘‘of necessity a dif-ficult inquiry.’’ Id., at 143, 106 S.Ct. 2797.

In her Bandemer concurrence, JusticeO’CONNOR predicted that the plurality’sstandard ‘‘will over time either prove un-

manageable and arbitrary or else evolvetowards some loose form of proportionali-ty.’’ Id., at 155, 106 S.Ct. 2797 (opinionconcurring in judgment, joined by Burger,C. J., and REHNQUIST, J.). A similarprediction of unmanageability was ex-pressed in Justice Powell’s opinion, makingit the prognostication of a majority of theCourt. See id., at 171, 106 S.Ct. 2797(‘‘The TTT most basic flaw in the plurality’sopinion is its failure to enunciate any stan-dard that affords guidance to legislaturesand courts’’). That prognostication hasbeen amply fulfilled.

In the lower courts, the legacy of theplurality’s test is one long record of puzzle-ment and consternation. See, e.g., Ses-sion, supra, at 474 (‘‘Throughout this casewe have borne S 283witness to the powerful,conflicting forces nurtured by Bandemer’sholding that the judiciary is to address‘excessive’ partisan line-drawing, whileleaving the issue virtually unenforceable’’);Vieth I, 188 F.Supp.2d, at 544 (noting thatthe ‘‘recondite standard enunciated inBandemer offers little concrete guidance’’);Martinez v. Bush, 234 F.Supp.2d 1275,1352 (S.D.Fla.2002) (three-judge court)(Jordan, J., concurring) (the ‘‘lower courtscontinue to struggle in an attempt to inter-pret and apply the ‘discriminatory effect’prong of the [Bandemer] standard’’);O’Lear, supra, at 855 (describing Bandem-er’s standard for assessing discriminatoryeffect as ‘‘somewhat murky’’). The testhas been criticized for its indeterminacy bya host of academic commentators. See,e.g., L. Tribe, American ConstitutionalLaw § 13–9, p. 1083 (2d ed. 1988) (‘‘Nei-ther Justice White’s nor Justice Powell’sapproach to the question of partisan ap-portionment gives any real guidance tolower courts forced to adjudicate this issueTTT’’); Still, Hunting of the Gerrymander,38 UCLA L.Rev. 1019, 1020 (1991) (notingthat the plurality opinion has ‘‘confounded


legislators, practitioners, and academicsalike’’); Schuck, The Thickest Thicket:Partisan Gerrymandering and JudicialRegulation of Politics, 87 Colum. L.Rev.1325, 1365 (1987) (noting that the Bandem-er plurality’s standard requires judgmentsthat are ‘‘largely subjective and beg ques-tions that lie at the heart of political com-petition in a democracy’’); Issacharoff,Judging Politics: The Elusive Quest forJudicial Review of Political Fairness, 71Texas L.Rev. 1643, 1671 (1993) (‘‘Bandem-er begot only confusion’’); Grofman, AnExpert Witness Perspective on Continuingand Emerging Voting Rights Controver-sies, 21 Stetson L.Rev. 783, 816 (1992)(‘‘[A]s far as I am aware I am one of onlytwo people who believe that Bandemermakes sense. Moreover, the other person,Daniel Lowenstein, has a diametrically op-posed view as to what the plurality opinionmeans’’). Because this standard was mis-guided when proposed, has not been im-proved in subsequent application, and isnot even deSfended284 before us today bythe appellants, we decline to affirm it as aconstitutional requirement.

B

Appellants take a run at enunciatingtheir own workable standard based on Ar-ticle I, § 2, and the Equal ProtectionClause. We consider it at length not onlybecause it reflects the litigant’s view as tothe best that can be derived from 18 yearsof experience, but also because it sharesmany features with other proposed stan-dards, so that what is said of it may besaid of them as well. Appellants’ proposedstandard retains the two-pronged frame-work of the Bandemer plurality—intentplus effect—but modifies the type of show-ing sufficient to satisfy each.

To satisfy appellants’ intent standard, aplaintiff must ‘‘show that the mapmakersacted with a predominant intent to

achieve partisan advantage,’’ which can beshown ‘‘by direct evidence or by circum-stantial evidence that other neutral andlegitimate redistricting criteria were sub-ordinated to the goal of achieving partisanadvantage.’’ Brief for Appellants 19 (em-phasis added). As compared with theBandemer plurality’s test of mere intent todisadvantage the plaintiff’s group, this pro-posal seemingly makes the standard moredifficult to meet—but only at the expenseof making the standard more indetermi-nate.

‘‘Predominant intent’’ to disadvantagethe plaintiff ’s political group refers to therelative importance of that goal as com-pared with all the other goals that the mapseeks to pursue—contiguity of districts,compactness of districts, observance of thelines of political subdivision, protection ofincumbents of all parties, cohesion of natu-ral racial and ethnic neighborhoods, com-pliance with requirements of the VotingRights Act of 1965 regarding racial distri-bution, etc. Appellants contend that theirintent test must be discernible and man-ageable because it has been borrowedfrom our racial gerrymandering cases.See Miller v. Johnson, 515 U.S. 900, 115S.Ct. 2475, 132 L.Ed.2d 762 S 285(1995);Shaw v. Reno, 509 U.S. 630, 113 S.Ct.2816, 125 L.Ed.2d 511 (1993). To beginwith, in a very important respect that isnot so. In the racial gerrymandering con-text, the predominant intent test has beenapplied to the challenged district in whichthe plaintiffs voted. See Miller, supra;United States v. Hays, 515 U.S. 737, 115S.Ct. 2431, 132 L.Ed.2d 635 (1995). Here,however, appellants do not assert that anapportionment fails their intent test if anysingle district does so. Since ‘‘it would bequixotic to attempt to bar state legisla-tures from considering politics as they re-draw district lines,’’ Brief for Appellants 3,appellants propose a test that is satisfiedonly when ‘‘partisan advantage was the


541 U.S. 287

predominant motivation behind the entirestatewide plan,’’ id., at 32 (emphasis add-ed). Vague as the ‘‘predominant motiva-tion’’ test might be when used to evaluatesingle districts, it all but evaporates whenapplied statewide. Does it mean, for in-stance, that partisan intent must outweighall other goals—contiguity, compactness,preservation of neighborhoods, etc.—state-wide? And how is the statewide ‘‘out-weighing’’ to be determined? If three-fifths of the map’s districts forgo the pur-suit of partisan ends in favor of strictlyobserving political-subdivision lines, andonly two-fifths ignore those lines to disad-vantage the plaintiffs, is the observance ofpolitical subdivisions the ‘‘predominant’’goal between those two? We are sureappellants do not think so.

Even within the narrower compass ofchallenges to a single district, applying a‘‘predominant intent’’ test to racial gerry-mandering is easier and less disruptive.The Constitution clearly contemplates dis-tricting by political entities, see Article I,§ 4, and unsurprisingly that turns out tobe root-and-branch a matter of politics.See Miller, supra, at 914, 115 S.Ct. 2475(‘‘[R]edistricting in most cases will impli-cate a political calculus in which variousinterests compete for recognition TTT’’);Shaw, supra, at 662, 113 S.Ct. 2816 (White,J., dissenting) (‘‘[D]istricting inevitably isthe expression of interest group politicsTTT’’); Gaffney v. Cummings, 412 U.S.735, 753, 93 S.Ct. 2321, 37 L.Ed.2d 298(1973) (‘‘The reality is that districting inev-itably has and is intended to have subSstan-tial286 political consequences’’). By con-trast, the purpose of segregating voters onthe basis of race is not a lawful one, and ismuch more rarely encountered. Deter-mining whether the shape of a particular

district is so substantially affected by thepresence of a rare and constitutionally sus-pect motive as to invalidate it is quitedifferent from determining whether it is sosubstantially affected by the excess of anordinary and lawful motive as to invalidateit. Moreover, the fact that partisan dis-tricting is a lawful and common practicemeans that there is almost always roomfor an election-impeding lawsuit contend-ing that partisan advantage was the pre-dominant motivation; not so for claims ofracial gerrymandering. Finally, courtsmight be justified in accepting a modestdegree of unmanageability to enforce aconstitutional command which (like theFourteenth Amendment obligation to re-frain from racial discrimination) is clear;whereas they are not justified in inferringa judicially enforceable constitutional obli-gation (the obligation not to apply toomuch partisanship in districting) which isboth dubious and severely unmanageable.For these reasons, to the extent that ourracial gerrymandering cases represent amodel of discernible and manageable stan-dards, they provide no comfort here.

The effects prong of appellants’ proposalreplaces the Bandemer plurality’s vaguetest of ‘‘denied its chance to effectivelyinfluence the political process,’’ 478 U.S., at132–133, 106 S.Ct. 2797, with criteria thatare seemingly more specific. The requi-site effect is established when ‘‘(1) theplaintiffs show that the districts systemat-ically ‘pack’ and ‘crack’ the rival party’svoters,7 and (2) the court’s examination ofthe ‘totality of circumstances’ confirmsthat the map can thwart the plaintiffs’ability to translate a majority of votes intoa majority S 287of seats.’’ Brief for Appel-lants 20 (emphasis and footnote added).This test is loosely based on our cases

7. ‘‘Packing’’ refers to the practice of filling adistrict with a supermajority of a given groupor party. ‘‘Cracking’’ involves the splitting of

a group or party among several districts todeny that group or party a majority in any ofthose districts.


applying § 2 of the Voting Rights Act of1965, 42 U.S.C. § 1973, to discriminationby race, see, e.g., Johnson v. De Grandy,512 U.S. 997, 114 S.Ct. 2647, 129 L.Ed.2d775 (1994). But a person’s politics is rare-ly as readily discernible—and never aspermanently discernible—as a person’srace. Political affiliation is not an immuta-ble characteristic, but may shift from oneelection to the next; and even within agiven election, not all voters follow theparty line. We dare say (and hope) thatthe political party which puts forward anutterly incompetent candidate will loseeven in its registration stronghold. Thesefacts make it impossible to assess the ef-fects of partisan gerrymandering, to fash-ion a standard for evaluating a violation,and finally to craft a remedy. See Ban-demer, supra, at 156, 106 S.Ct. 2797(O’CONNOR, J., concurring in judgment).8

Assuming, however, that the effects ofpartisan gerrymandering can be deter-mined, appellants’ test would invalidate thedistricting only when it prevents a majori-ty of the electorate from electing a majori-ty of representatives. Before considering

whether this particular standard is judi-cially S 288manageable we question whetherit is judicially discernible in the sense ofbeing relevant to some constitutional viola-tion. Deny it as appellants may (and do),this standard rests upon the principle thatgroups (or at least political-action groups)have a right to proportional representa-tion. But the Constitution contains nosuch principle. It guarantees equal pro-tection of the law to persons, not equalrepresentation in government to equiva-lently sized groups. It nowhere says thatfarmers or urban dwellers, Christian fun-damentalists or Jews, Republicans orDemocrats, must be accorded politicalstrength proportionate to their numbers.9

Even if the standard were relevant,however, it is not judicially manageable.To begin with, how is a party’s majoritystatus to be established? Appellants pro-pose using the results of statewide racesas the benchmark of party support. Butas their own complaint describes, in the2000 Pennsylvania statewide electionssome Republicans won and some Demo-

8. A delicious illustration of this is the one casewe have found—alluded to above—that pro-vided relief under Bandemer. See n. 5, supra.In Republican Party of North Carolina v.Hunt, No. 94–2410, 1996 WL 60439 (C.A.4,Feb.12, 1996) (per curiam) (unpublished),judgt. order reported at 77 F.3d 470, theDistrict Court, after a trial with no less than311 stipulations by the parties, 132 witnessstatements, approximately 300 exhibits, and 2days of oral argument, concluded that NorthCarolina’s system of electing superior courtjudges on a statewide basis ‘‘had resulted inRepublican candidates experiencing a consis-tent and pervasive lack of success and exclu-sion from the electoral process as a wholeand that these effects were likely to continueunabated into the future.’’ 77 F.3d, at 470,1996 WL 60439, at *1. In the elections forsuperior court judges conducted just five daysafter this pronouncement, ‘‘every Republicancandidate standing for the office of superiorcourt judge was victorious at the state level,’’ibid., a result which the Fourth Circuit

thought (with good reason) ‘‘directly at oddswith the recent prediction by the districtcourt,’’ id., at 470, 1996 WL 60439, at *2,causing it to remand the case for reconsidera-tion.

9. The Constitution also does not share appel-lants’ alarm at the asserted tendency of parti-san gerrymandering to create more partisanrepresentatives. Assuming that assertion tobe true, the Constitution does not answer thequestion whether it is better for Democraticvoters to have their State’s congressional del-egation include 10 wishy-washy Democrats(because Democratic voters are ‘‘effectively’’distributed so as to constitute bare majoritiesin many districts), or 5 hardcore Democrats(because Democratic voters are tightly packedin a few districts). Choosing the former ‘‘di-lutes’’ the vote of the radical Democrat;choosing the latter does the same to the mod-erate. Neither Article I, § 2, nor the EqualProtection Clause takes sides in this dispute.


541 U.S. 290

crats won. See Juris. Statement 137a–138a (describing how Democratic candi-dates received more votes for Presidentand auditor general, and Republicans re-ceived more votes for United States Sena-tor, attorney general, and treasurer).Moreover, to think that majority status instatewide races establishes majority statusfor district contests, one would have tobelieve that the only factor determiningvoting behavior at all levels is politicalaffiliation. That is assuredly not true. Asone law review comment has put it:

S 289‘‘There is no statewide vote in thiscountry for the House of Representa-tives or the state legislature. Rather,there are separate elections betweenseparate candidates in separate districts,and that is all there is. If the districtschange, the candidates change, theirstrengths and weaknesses change, theircampaigns change, their ability to raisemoney changes, the issues change—ev-erything changes. Political parties donot compete for the highest statewidevote totals or the highest mean districtvote percentages: They compete forspecific seats.’’ Lowenstein & Stein-berg, The Quest for Legislative District-ing in the Public Interest: Elusive orIllusory, 33 UCLA L.Rev. 1, 59–60(1985).

See also Schuck, Partisan Gerrymander-ing: A Political Problem Without JudicialSolution, in Political Gerrymandering andthe Courts 240, 241 (B. Grofman ed.1990).

But if we could identify a majority party,we would find it impossible to ensure thatthat party wins a majority of seats—unlesswe radically revise the States’ traditionalstructure for elections. In any winner-take-all district system, there can be noguarantee, no matter how the district linesare drawn, that a majority of party votesstatewide will produce a majority of seatsfor that party. The point is proved by the2000 congressional elections in Pennsylva-

nia, which, according to appellants’ ownpleadings, were conducted under a judicial-ly drawn district map ‘‘free from partisangerrymandering.’’ Juris. Statement 137a.On this ‘‘neutral playing fiel[d],’’ the Dem-ocrats’ statewide majority of the major-party vote (50.6%) translated into a minor-ity of seats (10, versus 11 for the Republi-cans). Id., at 133a, 137a. Whether byreason of partisan districting or not, partyconstituents may always wind up ‘‘packed’’in some districts and ‘‘cracked’’ throughoutothers. See R. Dixon, Democratic Repre-sentation 462 (1968) (‘‘All Districting Is‘Gerrymandering’ ’’); Schuck, 87 Colum.L.Rev., at 1359. Consider, for S 290example,a legislature that draws district lines withno objectives in mind except compactnessand respect for the lines of political subdi-visions. Under that system, politicalgroups that tend to cluster (as is the casewith Democratic voters in cities) would besystematically affected by what might becalled a ‘‘natural’’ packing effect. SeeBandemer, 478 U.S., at 159, 106 S.Ct. 2797(O’CONNOR, J., concurring in judgment).

Our one-person, one-vote cases, seeReynolds v. Sims, 377 U.S. 533, 84 S.Ct.1362, 12 L.Ed.2d 506 (1964); Wesberry v.Sanders, 376 U.S. 1, 84 S.Ct. 526, 11L.Ed.2d 481 (1964), have no bearing uponthis question, neither in principle nor inpracticality. Not in principle, because tosay that each individual must have anequal say in the selection of representa-tives, and hence that a majority of individ-uals must have a majority say, is not at allto say that each discernible group, wheth-er farmers or urban dwellers or politicalparties, must have representation equiva-lent to its numbers. And not in practicali-ty, because the easily administrable stan-dard of population equality adopted byWesberry and Reynolds enables judges todecide whether a violation has occurred(and to remedy it) essentially on the basisof three readily determined factors—


where the plaintiff lives, how many votersare in his district, and how many votersare in other districts; whereas requiringjudges to decide whether a districting sys-tem will produce a statewide majority for amajority party casts them forth upon a seaof imponderables, and asks them to makedeterminations that not even election ex-perts can agree upon.

For these reasons, we find appellants’proposed standards neither discernible normanageable.

C

For many of the same reasons, we alsoreject the standard suggested by JusticePowell in Bandemer. He agreed with theplurality that a plaintiff should show intentand effect, but believed that the ultimateinquiry ought to focus on whether districtboundaries had been drawn solely for par-ti Ssan291 ends to the exclusion of ‘‘all otherneutral factors relevant to the fairness ofredistricting.’’ 478 U.S., at 161, 106 S.Ct.2797 (opinion concurring in part and dis-senting in part); see also id., at 164–165,106 S.Ct. 2797. Under that inquiry, thecourts should consider numerous factors,though ‘‘[n]o one factor should be disposi-tive.’’ Id., at 173, 106 S.Ct. 2797. Themost important would be ‘‘the shapes ofvoting districts and adherence to estab-lished political subdivision boundaries.’’Ibid. ‘‘Other relevant considerations in-clude the nature of the legislative proce-dures by which the apportionment law wasadopted and legislative history reflectingcontemporaneous legislative goals.’’ Ibid.These factors, which ‘‘bear directly on thefairness of a redistricting plan,’’ combinedwith ‘‘evidence concerning population dis-parities and statistics tending to show votedilution,’’ make out a claim of unconstitu-tional partisan gerrymandering. Ibid.

While Justice Powell rightly criticizedthe Bandemer plurality for failing to sug-

gest a constitutionally based, judiciallymanageable standard, the standard pro-posed in his opinion also falls short of themark. It is essentially a totality-of-the-circumstances analysis, where all conceiva-ble factors, none of which is dispositive,are weighed with an eye to ascertainingwhether the particular gerrymander hasgone too far—or, in Justice Powell’s termi-nology, whether it is not ‘‘fair.’’ ‘‘Fair-ness’’ does not seem to us a judiciallymanageable standard. Fairness is com-patible with noncontiguous districts, it iscompatible with districts that straddle po-litical subdivisions, and it is compatiblewith a party’s not winning the number ofseats that mirrors the proportion of itsvote. Some criterion more solid and moredemonstrably met than that seems to usnecessary to enable the state legislaturesto discern the limits of their districtingdiscretion, to meaningfully constrain thediscretion of the courts, and to win publicacceptance for the courts’ intrusion into aprocess that is the very foundation of dem-ocratic decisionmaking.

S 292IV

We turn next to consideration of thestandards proposed by today’s dissenters.We preface it with the observation that themere fact that these four dissenters comeup with three different standards—all ofthem different from the two proposed inBandemer and the one proposed here byappellants—goes a long way to establish-ing that there is no constitutionally discer-nible standard.

A

Justice STEVENS concurs in the judg-ment that we should not address plaintiffs’statewide political gerrymandering chal-lenges. Though he reaches that result viastanding analysis, post, at 1805, 1806 (dis-senting opinion), while we reach it through


541 U.S. 294

political-question analysis, our conclusionsare the same: these statewide claims arenonjusticiable.

Justice STEVENS would, however, re-quire courts to consider political gerry-mandering challenges at the individual-district level. Much of his dissent is ad-dressed to the incompatibility of severepartisan gerrymanders with democraticprinciples. We do not disagree with thatjudgment, any more than we disagreewith the judgment that it would be uncon-stitutional for the Senate to employ, in im-peachment proceedings, procedures thatare incompatible with its obligation to‘‘try’’ impeachments. See Nixon v. Unit-ed States, 506 U.S. 224, 113 S.Ct. 732, 122L.Ed.2d 1 (1993). The issue we have dis-cussed is not whether severe partisan ger-rymanders violate the Constitution, butwhether it is for the courts to say when aviolation has occurred, and to design aremedy. On that point, Justice STE-VENS’s dissent is less helpful, saying, es-sentially, that if we can do it in the racialgerrymandering context we can do it here.

We have examined, supra, at 1780–1782,the many reasons why that is not so. Onlya few of them are challenged by JusticeSTEVENS. He says that we ‘‘mistakenlyassum[e] that race cannot provide a legiti-mate basis for making politiScal293 judg-ments.’’ Post, at 1811. But we do not saythat race-conscious decisionmaking is al-ways unlawful. Race can be used, forexample, as an indicator to achieve thepurpose of neighborhood cohesiveness indistricting. What we have said is imper-missible is ‘‘the purpose of segregatingvoters on the basis of race,’’ supra, at1781—that is to say, racial gerrymander-ing for race’s sake, which would be theequivalent of political gerrymandering forpolitics’ sake. Justice STEVENS says we‘‘er[r] in assuming that politics is ‘an ordi-nary and lawful motive’ ’’ in districting,

post, at 1803—but all he brings forward tocontest that is the argument that an exces-sive injection of politics is unlawful. So itis, and so does our opinion assume. Thatdoes not alter the reality that setting outto segregate voters by race is unlawful andhence rare, and setting out to segregatethem by political affiliation is (so long asone doesn’t go too far) lawful and henceordinary.

Justice STEVENS’s confidence thatwhat courts have done with racial gerry-mandering can be done with political ger-rymandering rests in part upon his beliefthat ‘‘the same standards should apply,’’post, at 1810. But in fact the standardsare quite different. A purpose to discrimi-nate on the basis of race receives thestrictest scrutiny under the Equal Protec-tion Clause, while a similar purpose todiscriminate on the basis of politics doesnot. ‘‘[N]othing in our case law compelsthe conclusion that racial and political ger-rymanders are subject to precisely thesame constitutional scrutiny. In fact, ourcountry’s long and persistent history ofracial discrimination in voting—as well asour Fourteenth Amendment jurispru-dence, which always has reserved thestrictest scrutiny for discrimination on thebasis of race—would seem to compel theopposite conclusion.’’ Shaw, 509 U.S., at650, 113 S.Ct. 2816 (citation omitted).That quoted passage was in direct re-sponse to (and rejection of) the suggestionmade by Justices White and STEVENS indissent that ‘‘a racial gerrymander of thesort alleged here is functionally equivalentto S 294gerrymanders for nonracial purposes,such as political gerrymanders.’’ Ibid.See also Bush v. Vera, 517 U.S. 952, 964,116 S.Ct. 1941, 135 L.Ed.2d 248 (1996)(plurality opinion) (‘‘We have not subjectedpolitical gerrymandering to strict scruti-ny’’).


Justice STEVENS relies on FirstAmendment cases to suggest that political-ly discriminatory gerrymanders are sub-ject to strict scrutiny under the EqualProtection Clause. See post, at 1803. Itis elementary that scrutiny levels are claimspecific. An action that triggers a height-ened level of scrutiny for one claim mayreceive a very different level of scrutinyfor a different claim because the underly-ing rights, and consequently constitutionalharms, are not comparable. To say thatsuppression of political speech (a claimedFirst Amendment violation) triggers strictscrutiny is not to say that failure to givepolitical groups equal representation (aclaimed equal protection violation) triggersstrict scrutiny. Only an equal protectionclaim is before us in the present case—perhaps for the very good reason that aFirst Amendment claim, if it were sus-tained, would render unlawful all consider-ation of political affiliation in districting,just as it renders unlawful all consider-ation of political affiliation in hiring fornon-policy-level government jobs. Whatcases such as Elrod v. Burns, 427 U.S. 347,96 S.Ct. 2673, 49 L.Ed.2d 547 (1976), re-quire is not merely that Republicans begiven a decent share of the jobs in aDemocratic administration, but that politi-cal affiliation be disregarded.

Having failed to make the case for strictscrutiny of political gerrymandering, Jus-tice STEVENS falls back on the argumentthat scrutiny levels simply do not matterfor purposes of justiciability. He assertsthat a standard imposing a strong pre-sumption of invalidity (strict scrutiny) is nomore discernible and manageable than astandard requiring an evenhanded balanc-ing of all considerations with no thumb onthe scales (ordinary scrutiny). To statethis is to refute it. As is well known, strictscrutiny readily, and almost always, re-sults in invalidation. Moreover, the mere

fact that there S 295exist standards whichthis Court could apply—the propositionwhich much of Justice STEVENS’s opin-ion is devoted to establishing, see, e.g.,post, at 1801–1804, 1812–1813—does notmean that those standards are discerniblein the Constitution. This Court may notwilly-nilly apply standards—even manage-able standards—having no relation to con-stitutional harms. Justice STEVENSpoints out, see post, at 1804–1805, n. 15,that Bandemer said differences betweenracial and political groups ‘‘may be rele-vant to the manner in which the case isadjudicated, but these differences do notjustify a refusal to entertain such a case.’’478 U.S., at 125, 106 S.Ct. 2797. As 18years have shown, Bandemer was wrong.

B

Justice SOUTER, like Justice STE-VENS, would restrict these plaintiffs, onthe allegations before us, to district-specif-ic political gerrymandering claims. Post,at 1817, 1820 (dissenting opinion). UnlikeJustice STEVENS, however, JusticeSOUTER recognizes that there is no exist-ing workable standard for adjudicatingsuch claims. He proposes a ‘‘fresh start,’’post, at 1816: a newly constructed stan-dard loosely based in form on our Title VIIcases, see McDonnell Douglas Corp. v.Green, 411 U.S. 792, 93 S.Ct. 1817, 36L.Ed.2d 668 (1973), and complete with afive-step prima facie test sewn togetherfrom parts of, among other things, ourVoting Rights Act jurisprudence, law re-view articles, and apportionment cases.Even if these self-styled ‘‘clues’’ to uncon-stitutionality could be manageably applied,which we doubt, there is no reason to thinkthey would detect the constitutional crimewhich Justice SOUTER is investigating—an ‘‘extremity of unfairness’’ in partisancompetition. Post, at 1815.


541 U.S. 297

Under Justice SOUTER’s proposedstandard, in order to challenge a particulardistrict, a plaintiff must show (1) that he isa member of a ‘‘cohesive political group’’;(2) ‘‘that the district of his residence TTT

paid little or no heed’’ to traditional dis-tricting principles; (3) that there were‘‘specific corSrelations296 between the dis-trict’s deviations from traditional district-ing principles and the distribution of thepopulation of his group’’; (4) that a hypo-thetical district exists which includes theplaintiff’s residence, remedies the packingor cracking of the plaintiff’s group, anddeviates less from traditional districtingprinciples; and (5) that ‘‘the defendantsacted intentionally to manipulate the shapeof the district in order to pack or crack hisgroup.’’ Post, at 1817–1819. When thoseshowings have been made, the burdenwould shift to the defendants to justify thedistrict ‘‘by reference to objectives otherthan naked partisan advantage.’’ Post, at1819.

While this five-part test seems eminent-ly scientific, upon analysis one finds thateach of the last four steps requires a quan-tifying judgment that is unguided and illsuited to the development of judicial stan-dards: How much disregard of traditionaldistricting principles? How many correla-tions between deviations and distribution?How much remedying of packing or crack-ing by the hypothetical district? Howmany legislators must have had the intentto pack and crack—and how efficaciousmust that intent have been (must it havebeen, for example, a sine qua non cause ofthe districting, or a predominant cause)?At step two, for example, Justice SOUT-ER would require lower courts to assesswhether mapmakers paid ‘‘little or no heedto TTT traditional districting principles.’’Post, at 1817. What is a lower court to dowhen, as will often be the case, the districtadheres to some traditional criteria but notothers? Justice SOUTER’s only response

to this question is to evade it: ‘‘It is notnecessary now to say exactly how a districtcourt would balance a good showing on oneof these indices against a poor showing onanother, for that sort of detail is bestworked out case by case.’’ Post, at 1818.But the devil lurks precisely in such detail.The central problem is determining whenpolitical gerrymandering has gone too far.It does not solve that problem to breakdown the original unanswerable quesStion297

(How much political motivation and effectis too much?) into four more discrete butequally unanswerable questions.

Justice SOUTER’s proposal is doomedto failure for a more basic reason: Notest—yea, not even a five-part test—canpossibly be successful unless one knowswhat he is testing for. In the presentcontext, the test ought to identify depriva-tion of that minimal degree of representa-tion or influence to which a political groupis constitutionally entitled. As we haveseen, the Bandemer test sought (unhelp-fully, but at least gamely) to specify whatthat minimal degree was: ‘‘[a] chance toeffectively influence the political process.’’478 U.S., at 133, 106 S.Ct. 2797. So didthe appellants’ proposed test: ‘‘[the] abilityto translate a majority of votes into amajority of seats.’’ Brief for Appellants20. Justice SOUTER avoids the difficul-ties of those formulations by never tellingus what his test is looking for, other thanthe utterly unhelpful ‘‘extremity of unfair-ness.’’ He vaguely describes the harm heis concerned with as vote dilution, post, at1820, a term which usually implies someactual effect on the weight of a vote. Butno element of his test looks to the effect ofthe gerrymander on the electoral success,the electoral opportunity, or even the polit-ical influence, of the plaintiff ’s group. Wedo not know the precise constitutional de-


privation his test is designed to identifyand prevent.

Even if (though it is implausible) JusticeSOUTER believes that the constitutionaldeprivation consists of merely ‘‘vote dilu-tion,’’ his test would not even identify thateffect. Despite his claimed reliance on theMcDonnell Douglas framework, JusticeSOUTER would allow the plaintiff no op-portunity to show that the mapmakers’compliance with traditional districting fac-tors is pretextual.10 His reason for S 298thisis never stated, but it certainly cannot bethat adherence to traditional districtingfactors negates any possibility of intention-al vote dilution. As we have explainedabove, packing and cracking, whether in-tentional or no, are quite consistent withadherence to compactness and respect forpolitical subdivision lines. See supra, at1783. An even better example is the tradi-tional criterion of incumbency protection.Justice SOUTER has previously acknowl-edged it to be a traditional and constitu-tionally acceptable districting principle.See Vera, 517 U.S., at 1047–1048, 116 S.Ct.1941 (dissenting opinion). Since that is so,his test would not protect those who arepacked, and often tightly so, to ensure thereelection of representatives of either par-ty. Indeed, efforts to maximize partisanrepresentation statewide might well beginwith packing voters of the opposing partyinto the districts of existing incumbents ofthat party. By this means an incumbent isprotected, a potential adversary to the dis-tricting mollified, and votes of the oppos-ing party are diluted.

Like us, Justice SOUTER acknowledgesand accepts that ‘‘some intent to gain polit-ical advantage is inescapable whenever po-litical bodies devise a district plan, and

some effect results from the intent.’’ Post,at 1815. Thus, again like us, he recognizesthat ‘‘the issue is one of how much is toomuch.’’ Ibid. And once those premises areconceded, the only line that can be drawnmust be based, as Justice SOUTER againcandidly admits, upon a substantive ‘‘no-tio[n] of fairness.’’ Ibid. This is the sameflabby goal that deprived Justice Powell’stest of all determinacy. To be sure, Jus-tice SOUTER frames it somewhat differ-ently: Courts must intervene, he says,when ‘‘partisan competition has reached anextremity of unfairness.’’ Post, at 1815(emphasis added). We do not think theproblem is solved by adding the modifier.

S 299C

We agree with much of Justice BREY-ER’s dissenting opinion, which convincing-ly demonstrates that ‘‘political consider-ations will likely play an important, andproper, role in the drawing of districtboundaries.’’ Post, at 1823. This placesJustice BREYER, like the other dissen-ters, in the difficult position of drawing theline between good politics and bad politics.Unlike them, he would tackle this problemat the statewide level.

The criterion Justice BREYER propos-es is nothing more precise than ‘‘the un-justified use of political factors to entrencha minority in power.’’ Post, at 1825 (em-phasis in original). While he invokes inpassing the Equal Protection Clause, itshould be clear to any reader that whatconstitutes unjustified entrenchment de-pends on his own theory of ‘‘effective gov-ernment.’’ Post, at 1823. While one must

10. Justice SOUTER would allow a State, inproving its affirmative defense, to demon-strate that the reasons given for the district’sshape ‘‘were more than a mere pretext for anold-fashioned gerrymander.’’ Post, at 1820.But the need to establish that affirmative de-

fense does not arise until the plaintiff hasestablished his prima facie case. And thatprima facie case fails when, under step two,the district on its face complies with tradition-al districting criteria.


541 U.S. 301

agree with Justice BREYER’s incrediblyabstract starting point that our Constitu-tion sought to create a ‘‘basically demo-cratic’’ form of government, post, at 1822,that is a long and impassable distanceaway from the conclusion that the Judicia-ry may assess whether a group (somehowdefined) has achieved a level of politicalpower (somehow defined) commensuratewith that to which they would be entitledabsent unjustified political machinations(whatever that means).

Justice BREYER provides no real guid-ance for the journey. Despite his promiseto do so, ibid., he never tells us what he istesting for, beyond the unhelpful ‘‘unjusti-fied entrenchment.’’ Post, at 1825. In-stead, he ‘‘set[s] forth several sets of cir-cumstances that lay out the indicia ofabuse,’’ ‘‘along a continuum,’’ post, at 1828,proceeding (presumably) from the mostclearly unconstitutional to the possibly un-constitutional. With regard to the first‘‘scenario,’’ he is willing to assert that theindicia ‘‘would be sufficient to support aclaim.’’ Post, at 1828. This seems re-freshingly categorical, until one realizesthat the indicia consist not merely of thefailure of the party receiving the majorityof votes to acquire S 300a majority of seats intwo successive elections, but also of thefact that there is no ‘‘neutral’’ explanationfor this phenomenon. Ibid. But of coursethere always is a neutral explanation—ifonly the time-honored criterion of incum-bent protection. The indicia set forth inJustice BREYER’s second scenario ‘‘couldalso add up to unconstitutional gerryman-dering,’’ post, at 1828 (emphasis added);and for those in the third ‘‘a court mayconclude that the map crosses the constitu-tional line,’’ post, at 1828 (emphasis added).We find none of this helpful. Each scenar-io suffers from at least one of the problemswe have previously identified, most notablythe difficulties of assessing partisanstrength statewide and of ascertaining

whether an entire statewide plan is moti-vated by political or neutral justifications,see supra, at 1781, 1783. And even atthat, the last two scenarios do not evenpurport to provide an answer, presumablyleaving it to each district court to deter-mine whether, under those circumstances,‘‘unjustified entrenchment’’ has occurred.In sum, we neither know precisely whatJustice BREYER is testing for, nor pre-cisely what fails the test.

But perhaps the most surprising omis-sion from Justice BREYER’s dissent, giv-en his views on other matters, is the ab-sence of any cost-benefit analysis. JusticeBREYER acknowledges that ‘‘a majoritynormally can work its political will,’’ post,at 1826, and well describes the number ofactors, from statewide executive officers,to redistricting commissions, to Congress,to the People in ballot initiatives and refer-enda, that stand ready to make that hap-pen. See post, at 1826. He gives noinstance (and we know none) of permanentfrustration of majority will. But wherethe majority has failed to assert itself forsome indeterminate period (two successiveelections, if we are to believe his firstscenario), Justice BREYER simply as-sumes that ‘‘court action may prove neces-sary,’’ post, at 1827. Why so? In the realworld, of course, court action that is avail-able tends to be sought, not just where itis necessary, but where it is in the interestof the seeking party. And the vaguer thetest S 301for availability, the more frequentlyinterest rather than necessity will producelitigation. Is the regular insertion of thejudiciary into districting, with the delayand uncertainty that brings to the politicalprocess and the partisan enmity it bringsupon the courts, worth the benefit to beachieved—an accelerated (by some un-known degree) effectuation of the majoritywill? We think not.


V

Justice KENNEDY recognizes that wehave ‘‘demonstrat[ed] the shortcomings ofthe other standards that have been consid-ered to date,’’ post, at 1794 (opinion con-curring in judgment). He acknowledges,moreover, that we ‘‘lack TTT comprehen-sive and neutral principles for drawingelectoral boundaries,’’ post, at 1793; andthat there is an ‘‘absence of rules to limitand confine judicial intervention,’’ ibid.From these premises, one might think thatJustice KENNEDY would reach the con-clusion that political gerrymanderingclaims are nonjusticiable. Instead, howev-er, he concludes that courts should contin-ue to adjudicate such claims because astandard may one day be discovered.

The first thing to be said about JusticeKENNEDY’s disposition is that it is notlegally available. The District Court inthis case considered the plaintiffs’ claimsjusticiable but dismissed them because thestandard for unconstitutionality had notbeen met. It is logically impossible toaffirm that dismissal without either (1)finding that the unconstitutional-districtingstandard applied by the District Court, orsome other standard that it should haveapplied, has not been met, or (2) finding(as we have) that the claim is nonjusticia-ble. Justice KENNEDY seeks to affirm‘‘[b]ecause, in the case before us, we haveno standard.’’ Post, at 1796. But it is ourjob, not the plaintiffs’, to explicate thestandard that makes the facts alleged bythe plaintiffs adequate or inadequate tostate a claim. We cannot nonsuit them forour failure to do so.

S 302Justice KENNEDY asserts that todeclare nonjusticiability would be incau-tious. Post, at 1795. Our rush to such aholding after a mere 18 years of fruitlesslitigation ‘‘contrasts starkly’’ he says,‘‘with the more patient approach’’ that thisCourt has taken in the past. Post, at

1795. We think not. When it has come todetermining what areas fall beyond ourArticle III authority to adjudicate, thisCourt’s practice, from the earliest days ofthe Republic to the present, has beenmore reminiscent of Hannibal than ofHamlet. On July 18, 1793, Secretary ofState Thomas Jefferson wrote the Justicesat the direction of President Washington,asking whether they might answer ‘‘ques-tions [that] depend for their solution onthe construction of our treaties, on thelaws of nature and nations, and on thelaws of the land,’’ but that arise ‘‘undercircumstances which do not give a cogni-sance of them to the tribunals of the coun-try.’’ 3 Correspondence and Public Pa-pers of John Jay 486–487 (H. Johnston ed.1891) (emphasis in original). The letterspecifically invited the Justices to give lessthan a categorical yes-or-no answer, offer-ing to present the particular questions‘‘from which [the Justices] will themselvesstrike out such as any circumstancesmight, in their opinion, forbid them topronounce on.’’ Id., at 487. On August 8,1793, the Justices responded in a categori-cal and decidedly ‘‘impatient’’ manner,saying that the giving of advisory opin-ions—not just advisory opinions on partic-ular questions but all advisory opinions,presumably even those concerning legisla-tion affecting the Judiciary—was beyondtheir power. ‘‘[T]he lines of separationdrawn by the Constitution between thethree departments of the government’’prevented it. Id., at 488. The Court re-jected the more ‘‘cautious’’ course of not‘‘deny[ing] all hopes of intervention,’’ post,at 1794, but leaving the door open to thepossibility that at least some advisoryopinions (on a theory we could not yetimagine) would not violate the separationof powers. In Gilligan v. Morgan, 413U.S. 1, 7, 93 S.Ct. 2440, 37 L.Ed.2d 407(1973), a case filed after the Ohio NationalGuard’s shooting S 303of students at Kent


541 U.S. 304

State University, the plaintiffs sought ‘‘ini-tial judicial review and continuing surveil-lance by a federal court over the training,weaponry, and orders of the Guard.’’ TheCourt held the suit nonjusticiable; thematter was committed to the politicalbranches because, inter alia, ‘‘it is difficultto conceive of an area of governmentalactivity in which the courts have less com-petence.’’ Id., at 10, 93 S.Ct. 2440. TheCourt did not adopt the more ‘‘cautious’’course of letting the lower courts try theirhand at regulating the military before wedeclared it impossible. Most recently, inNixon v. United States, the Court, joinedby Justice KENNEDY, held that a claimthat the Senate had employed certain im-permissible procedures in trying an im-peachment was a nonjusticiable politicalquestion. Our decision was not limited tothe particular procedures under challenge,and did not reserve the possibility thatsometime, somewhere, technology or thewisdom derived from experience mightmake a court challenge to Senate impeach-ment all right.

The only cases Justice KENNEDY citesin defense of his never-say-never approachare Baker v. Carr and Bandemer. Seepost, at 1795. Bandemer provides no cov-er. There, all of the Justices who conclud-ed that political gerrymandering claimsare justiciable proceeded to describe whatthey regarded as the discernible and man-ageable standard that rendered it so. Thelower courts were set wandering in thewilderness for 18 years not because theBandemer majority thought it a good idea,but because five Justices could not agreeupon a single standard, and because thestandard the plurality proposed turned outnot to work.

As for Baker v. Carr: It is true enoughthat, having had no experience whatever inapportionment matters of any sort, theCourt there refrained from spelling out theequal protection standard. (It did so a

mere two years later in Reynolds v. Sims,377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506(1964).) But the judgment under review inBaker, unlike the one under review here,did not demand the determination of astandard. The lower S 304court in Bakerhad held the apportionment claim of theplaintiffs nonjusticiable, and so it was logi-cally possible to dispose of the appeal bysimply disagreeing with the nonjusticiabili-ty determination. As we observed earlier,that is not possible here, where the lowercourt has held the claim justiciable butunsupported by the facts. We must eitherenunciate the standard that causes us toagree or disagree with that merits judg-ment, or else affirm that the claim is be-yond our competence to adjudicate.

Justice KENNEDY worries that ‘‘[a] de-termination by the Court to deny all hopesof intervention could erode confidence inthe courts as much as would a prematuredecision to intervene.’’ Post, at 1794–1795.But it is the function of the courts toprovide relief, not hope. What we thinkwould erode confidence is the Court’s re-fusal to do its job—announcing that theremay well be a valid claim here, but we arenot yet prepared to figure it out. More-over, that course does more than erodeconfidence; by placing the district courtsback in the business of pretending to af-ford help when they in fact can give none,it deters the political process from afford-ing genuine relief. As was noted by alower court confronted with a political ger-rymandering claim:

‘‘When the Supreme Court resolves Vi-eth, it may choose to retreat from itsdecision that the question is justiciable,or it may offer more guidance on thenature of the required effectTTTT Wehave learned firsthand what will result ifthe Court chooses to do neither.Throughout this case we have borne wit-ness to the powerful, conflicting forcesnurtured by Bandemer’s holding that


the judiciary is to address ‘excessive’partisan line-drawing, while leaving theissue virtually unenforceable. Inevita-bly, as the political party in power usesdistrict lines to lock in its present advan-tage, the party out of power attempts tostretch the protective cover of the Vot-ing Rights Act, urging dilution of criticalstandards that may, if accepted, aidtheir party in the short-run but S 305workto the detriment of persons now protect-ed by the Act in the long-run. Castingthe appearance both that there is awrong and that the judiciary standsready with a remedy, Bandemer as ap-plied steps on legislative incentives forself-correction.’’ Session, 298F.Supp.2d, at 474.

But the conclusive refutation of JusticeKENNEDY’s position is the point we firstmade: it is not an available disposition.We can affirm because political districtingpresents a nonjusticiable question; or wecan affirm because we believe the correctstandard which identifies unconstitutionalpolitical districting has not been met; wecannot affirm because we do not knowwhat the correct standard is. Reduced toits essence, Justice KENNEDY’s opinionboils down to this: ‘‘As presently advised,I know of no discernible and manageablestandard that can render this claim justici-able. I am unhappy about that, and hopethat I will be able to change my opinion inthe future.’’ What are the lower courts tomake of this pronouncement? We suggestthat they must treat it as a reluctant fifthvote against justiciability at district andstatewide levels—a vote that may changein some future case but that holds, for thetime being, that this matter is nonjusticia-ble.

VI

We conclude that neither Article I, § 2,nor the Equal Protection Clause, nor (whatappellants only fleetingly invoke) Article I,

§ 4, provides a judicially enforceable limiton the political considerations that theStates and Congress may take into accountwhen districting.

Considerations of stare decisis do notcompel us to allow Bandemer to stand.That case involved an interpretation of theConstitution, and the claims of stare deci-sis are at their weakest in that field, whereour mistakes cannot be corrected by Con-gress. See Payne v. Tennessee, 501 U.S.808, 828, 111 S.Ct. 2597, 115 L.Ed.2d 720(1991). They are doubly weak in Bandem-er because the maSjority’s306 inability toenunciate the judicially discernible andmanageable standard that it thought exist-ed (or did not think did not exist) presagedthe need for reconsideration in light ofsubsequent experience. And they are tri-ply weak because it is hard to imagine howany action taken in reliance upon Bandem-er could conceivably be frustrated—exceptthe bringing of lawsuits, which is not thesort of primary conduct that is relevant.

While we do not lightly overturn one ofour own holdings, ‘‘when governing deci-sions are unworkable or are badly rea-soned, ‘this Court has never felt con-strained to follow precedent.’ ’’ 501 U.S.,at 827, 111 S.Ct. 2597 (quoting Smith v.Allwright, 321 U.S. 649, 665, 64 S.Ct. 757,88 L.Ed. 987 (1944)). Eighteen years ofessentially pointless litigation have per-suaded us that Bandemer is incapable ofprincipled application. We would there-fore overrule that case, and decline to ad-judicate these political gerrymanderingclaims.

The judgment of the District Court isaffirmed.

It is so ordered.

Justice KENNEDY, concurring in thejudgment.

A decision ordering the correction of allelection district lines drawn for partisanreasons would commit federal and state


541 U.S. 308

courts to unprecedented intervention inthe American political process. The Courtis correct to refrain from directing thissubstantial intrusion into the Nation’s po-litical life. While agreeing with the plural-ity that the complaint the appellants filedin the District Court must be dismissed,and while understanding that great cautionis necessary when approaching this sub-ject, I would not foreclose all possibility ofjudicial relief if some limited and preciserationale were found to correct an estab-lished violation of the Constitution in someredistricting cases.

When presented with a claim of injuryfrom partisan gerrymandering, courts con-front two obstacles. First is the lack ofcomprehensive and neutral principles fordrawing S 307electoral boundaries. No sub-stantive definition of fairness in districtingseems to command general assent. Sec-ond is the absence of rules to limit andconfine judicial intervention. With uncer-tain limits, intervening courts—even whenproceeding with best intentions—wouldrisk assuming political, not legal, responsi-bility for a process that often produces illwill and distrust.

That courts can grant relief in district-ing cases where race is involved does notanswer our need for fairness principleshere. Those controversies implicate a dif-ferent inquiry. They involve sorting per-missible classifications in the redistrictingcontext from impermissible ones. Race isan impermissible classification. See Shawv. Reno, 509 U.S. 630, 113 S.Ct. 2816, 125L.Ed.2d 511 (1993). Politics is quite adifferent matter. See Gaffney v. Cum-mings, 412 U.S. 735, 752, 93 S.Ct. 2321, 37L.Ed.2d 298 (1973) (‘‘It would be idle, wethink, to contend that any political consid-eration taken into account in fashioning areapportionment plan is sufficient to invali-date it’’).

A determination that a gerrymander vio-lates the law must rest on something morethan the conclusion that political classifica-tions were applied. It must rest insteadon a conclusion that the classifications,though generally permissible, were appliedin an invidious manner or in a way unrelat-ed to any legitimate legislative objective.

The object of districting is to establish‘‘fair and effective representation for allcitizens.’’ Reynolds v. Sims, 377 U.S. 533,565–568, 84 S.Ct. 1362, 12 L.Ed.2d 506(1964). At first it might seem that courtscould determine, by the exercise of theirown judgment, whether political classifica-tions are related to this object or insteadburden representational rights. The lack,however, of any agreed upon model of fairand effective representation makes thisanalysis difficult to pursue.

The second obstacle—the absence ofrules to confine judicial intervention—isrelated to the first. Because there are yetno agreed upon substantive principles offairness in districting, we have no basis onwhich to define clear, manageSable,308 andpolitically neutral standards for measuringthe particular burden a given partisanclassification imposes on representationalrights. Suitable standards for measuringthis burden, however, are critical to ourintervention. Absent sure guidance, theresults from one gerrymandering case tothe next would likely be disparate andinconsistent.

In this case, we have not overcome theseobstacles to determining that the chal-lenged districting violated appellants’rights. The fairness principle appellantspropose is that a majority of voters in theCommonwealth should be able to elect amajority of the Commonwealth’s congres-sional delegation. There is no authorityfor this precept. Even if the novelty ofthe proposed principle were accompaniedby a convincing rationale for its adoption,


there is no obvious way to draw a satisfac-tory standard from it for measuring analleged burden on representational rights.The plurality demonstrates the shortcom-ings of the other standards that have beenconsidered to date. See ante, at Parts IIIand IV (demonstrating that the standardsproposed in Davis v. Bandemer, 478 U.S.109, 106 S.Ct. 2797, 92 L.Ed.2d 85 (1986),by the parties before us, and by our dis-senting colleagues are either unmanagea-ble or inconsistent with precedent, orboth). I would add two comments to theplurality’s analysis. The first is that theparties have not shown us, and I have notbeen able to discover, helpful discussionson the principles of fair districting dis-cussed in the annals of parliamentary orlegislative bodies. Our attention has notbeen drawn to statements of principled,well-accepted rules of fairness that shouldgovern districting, or to helpful formula-tions of the legislator’s duty in drawingdistrict lines.

Second, even those criteria that mightseem promising at the outset (e.g., contigu-ity and compactness) are not altogethersound as independent judicial standardsfor measuring a burden on representation-al rights. They cannot promise politicalneutrality when used as the basis for re-lief. Instead, it seems, a decision underthese standards would S 309unavoidably havesignificant political effect, whether intend-ed or not. For example, if we were todemand that congressional districts take aparticular shape, we could not assure theparties that this criterion, neutral enoughon its face, would not in fact benefit onepolitical party over another. See Gaffney,supra, at 753, 93 S.Ct. 2321 (‘‘District linesare rarely neutral phenomena. They canwell determine what district will be pre-dominantly Democratic or predominantlyRepublican, or make a close race likely’’);see also R. Bork, The Tempting of Amer-ica: The Political Seduction of the Law 88–

89 (1990) (documenting the author’s ser-vice as a special master responsible forredistricting Connecticut and noting thathis final plan so benefited the DemocraticParty, albeit unintentionally, that the partychairman personally congratulated him);M. Altman, Modeling the Effect of Manda-tory District Compactness on PartisanGerrymanders, 17 Pol. Geography 989,1000–1006 (1998) (explaining that compact-ness standards help Republicans becauseDemocrats are more likely to live in highdensity regions).

The challenge in finding a manageablestandard for assessing burdens on repre-sentational rights has long been recog-nized. See Lowenstein & Steinberg, TheQuest for Legislative Districting in thePublic Interest: Elusive or Illusory? 33UCLA L.Rev. 1, 74 (1985) (‘‘[W]hat mat-ters to us, and what we think matters toalmost all Americans when district linesare drawn, is how the fortunes of theparties and the policies the parties standfor are affected. When such things are atstake there is no neutrality. There is onlypolitical contest’’). The dearth of helpfulhistorical guidance must, in part, causethis uncertainty.

There are, then, weighty arguments forholding cases like these to be nonjusticia-ble; and those arguments may prevail inthe long run. In my view, however, thearguments are not so compelling that theyrequire us now to bar all future claims ofinjury from a partisan gerrymander. It isnot in our tradition to foreclose the judicialprocess from the atStempt310 to define stan-dards and remedies where it is allegedthat a constitutional right is burdened ordenied. Nor is it alien to the Judiciary todraw or approve election district lines.Courts, after all, already do so in manyinstances. A determination by the Courtto deny all hopes of intervention could


541 U.S. 311

erode confidence in the courts as much aswould a premature decision to intervene.

Our willingness to enter the politicalthicket of the apportionment process withrespect to one-person, one-vote claimsmakes it particularly difficult to justify acategorical refusal to entertain claimsagainst this other type of gerrymandering.The plurality’s conclusion that absent an‘‘easily administrable standard,’’ ante, at1783, the appellants’ claim must be nonjus-ticiable contrasts starkly with the morepatient approach of Baker v. Carr, 369U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d 663(1962), not to mention the controlling prec-edent on the question of justiciability ofDavis v. Bandemer, supra, the case theplurality would overrule. See ante, at1792.

In Baker the Court made clear that themore abstract standards that guide analy-sis of all Fourteenth Amendment claimssufficed to ensure justiciability of a one-person, one-vote claim. See 369 U.S., at226, 82 S.Ct. 691.

‘‘Nor need the appellants, in order tosucceed in this action, ask the Court toenter upon policy determinations forwhich judicially manageable standardsare lacking. Judicial standards underthe Equal Protection Clause are welldeveloped and familiar, and it has beenopen to courts since the enactment ofthe Fourteenth Amendment to deter-mine, if on the particular facts theymust, that a discrimination reflects nopolicy, but simply arbitrary and capri-cious action.’’ Ibid.

The Court said this before the more specif-ic standard with which we are now familiaremerged to measure the burden nonequi-populous districting causes on representa-tional rights. See Reynolds, 377 U.S., at565–568, 84 S.Ct. 1362 (concluding thatS 311‘‘[s]ince the achieving of fair and effec-tive representation for all citizens is con-

cededly the basic aim of legislative appor-tionment,’’ a legislature’s reliance on otherapportionment interests is invalid, arbi-trary, and capricious action if it leads tounequal populations among districts). Theplurality’s response that in Baker thisCourt sat in review only of a nonjusticiabil-ity holding is wide of the mark. See ante,at 1791. As the plurality itself instructs:Before a court can conclude that it ‘‘has[any] business entertaining [a] claim,’’ itmust conclude that some ‘‘judicially en-forceable righ[t]’’ is at issue. Ante, at1776. Whether a manageable standardmade the right at issue in Baker enforce-able was as much a necessary inquirythere as it is here. In light of Baker andDavis v. Bandemer, which directly addressthe question of nonjusticiability in the spe-cific context of districting and of assertedviolations of the Fourteenth Amendment,the plurality’s further survey of cases in-volving different approaches to the justici-ability of different claims cannot bethought controlling. See ante, at 1790.

Even putting Baker to the side—and soassuming that the existence of a workablestandard for measuring a gerrymander’sburden on representational rights distin-guishes one-person, one-vote claims frompartisan gerrymandering claims for justici-ability purposes—I would still reject theplurality’s conclusions as to nonjusticiabili-ty. Relying on the distinction between aclaim having or not having a workablestandard of that sort involves a difficultproof: proof of a categorical negative.That is, the different treatment of claimsotherwise so alike hinges entirely on proofthat no standard could exist. This is adifficult proposition to establish, for prov-ing a negative is a challenge in any con-text.

That no such standard has emerged inthis case should not be taken to provethat none will emerge in the future.


Where important rights are involved, theimpossibility of full analytical satisfactionis reason to err on the side of caution.Allegations of unconstitutional bias in ap-portionment are S 312most serious claims,for we have long believed that ‘‘the rightto vote’’ is one of ‘‘those political processesordinarily to be relied upon to protect mi-norities.’’ United States v. CaroleneProducts Co., 304 U.S. 144, 153, n. 4, 58S.Ct. 778, 82 L.Ed. 1234 (1938). If aState passed an enactment that declared‘‘All future apportionment shall be drawnso as most to burden Party X’s rights tofair and effective representation, thoughstill in accord with one-person, one-voteprinciples,’’ we would surely conclude theConstitution had been violated. If that isso, we should admit the possibility re-mains that a legislature might attempt toreach the same result without that ex-press directive. This possibility suggeststhat in another case a standard mightemerge that suitably demonstrates how anapportionment’s de facto incorporation ofpartisan classifications burdens rights offair and effective representation (and soestablishes the classification is unrelatedto the aims of apportionment and thus isused in an impermissible fashion).

The plurality says that 18 years, in ef-fect, prove the negative. Ante, at 1792(‘‘Eighteen years of essentially pointlesslitigation have persuaded us’’). As JusticeSOUTER is correct to point out, however,during these past 18 years the lowercourts could do no more than follow Davisv. Bandemer, which formulated a single,apparently insuperable standard. Seepost, at 1816 (dissenting opinion). More-over, by the timeline of the law 18 years israther a short period. In addition, therapid evolution of technologies in the ap-portionment field suggests yet unexploredpossibilities. Computer assisted district-ing has become so routine and sophisticat-ed that legislatures, experts, and courts

can use databases to map electoral dis-tricts in a matter of hours, not months.See, e.g., Larios v. Cox, 305 F.Supp.2d1335 (N.D.Ga.2004) (per curiam). Tech-nology is both a threat and a promise. Onthe one hand, if courts refuse to entertainany claims of partisan gerrymandering,the temptation to use partisan favoritismin districting in an unconstitutional man-ner will grow. On the other hand, thesenew S 313technologies may produce newmethods of analysis that make more evi-dent the precise nature of the burdensgerrymanders impose on the representa-tional rights of voters and parties. Thatwould facilitate court efforts to identifyand remedy the burdens, with judicial in-tervention limited by the derived stan-dards.

If suitable standards with which to mea-sure the burden a gerrymander imposeson representational rights did emerge,hindsight would show that the Court pre-maturely abandoned the field. That is arisk the Court should not take. Instead,we should adjudicate only what is in thepapers before us. See Baker, 369 U.S., at331, 82 S.Ct. 691 (Harlan, J., dissenting)(concluding that the malapportionmentclaim ‘‘should have been dismissed for ‘fail-ure to state a claim upon which relief canbe granted’ ’’ because ‘‘[u]ntil it is firstdecided to what extent [the] right [to ap-portion] is limited by the Federal Constitu-tion, and whether what [a State] has doneor failed to do TTT runs afoul of any suchlimitation, we need not reach the issues of‘justiciability’ or ‘political question’ ’’).

Because, in the case before us, we haveno standard by which to measure the bur-den appellants claim has been imposed ontheir representational rights, appellantscannot establish that the alleged politicalclassifications burden those same rights.Failing to show that the alleged classifica-tions are unrelated to the aims of appor-tionment, appellants’ evidence at best dem-


541 U.S. 315

onstrates only that the legislature adoptedpolitical classifications. That describes noconstitutional flaw, at least under the gov-erning Fourteenth Amendment standard.See Gaffney, 412 U.S., at 752, 93 S.Ct.2321. As a consequence, appellants’ com-plaint alleges no impermissible use of po-litical classifications and so states no validclaim on which relief may be granted. Itmust be dismissed as a result. See Fed.Rule Civ. Proc. 12(b)(6); see also Davis v.Bandemer, 478 U.S., at 134, 106 S.Ct.2797.

The plurality thinks I resolve this casewith reference to no standard, see ante, at1790, but that is wrong. TheFourSteenth314 Amendment standard gov-erns; and there is no doubt of that. Myanalysis only notes that if a subsidiarystandard could show how an otherwisepermissible classification, as applied, bur-dens representational rights, we could con-clude that appellants’ evidence states aprovable claim under the FourteenthAmendment standard.

Though in the briefs and at argumentthe appellants relied on the Equal Protec-tion Clause as the source of their substan-tive right and as the basis for relief, I notethat the complaint in this case also allegeda violation of First Amendment rights.See Amended Complaint ¶ 48; Juris.Statement 145a. The First Amendmentmay be the more relevant constitutionalprovision in future cases that allege uncon-stitutional partisan gerrymandering. Af-ter all, these allegations involve the FirstAmendment interest of not burdening orpenalizing citizens because of their partic-ipation in the electoral process, their vot-ing history, their association with a politi-cal party, or their expression of politicalviews. See Elrod v. Burns, 427 U.S. 347,96 S.Ct. 2673, 49 L.Ed.2d 547 (1976) (plu-rality opinion). Under general FirstAmendment principles those burdens in

other contexts are unconstitutional absenta compelling government interest. Seeid., at 362, 96 S.Ct. 2673. ‘‘Representativedemocracy in any populous unit of gover-nance is unimaginable without the abilityof citizens to band together in promotingamong the electorate candidates who es-pouse their political views.’’ CaliforniaDemocratic Party v. Jones, 530 U.S. 567,574, 120 S.Ct. 2402, 147 L.Ed.2d 502(2000). As these precedents show, FirstAmendment concerns arise where a Stateenacts a law that has the purpose andeffect of subjecting a group of voters ortheir party to disfavored treatment by rea-son of their views. In the context of parti-san gerrymandering, that means that FirstAmendment concerns arise where an ap-portionment has the purpose and effect ofburdening a group of voters’ representa-tional rights.

The plurality suggests there is no placefor the First Amendment in this area.See ante, at 1786. The implication S 315isthat under the First Amendment any andall consideration of political interests in anapportionment would be invalid. Ibid.(‘‘Only an equal protection claim is beforeus in the present case—perhaps for thevery good reason that a First Amendmentclaim, if it were sustained, would renderunlawful all consideration of political affili-ation in districting’’). That misrepresentsthe First Amendment analysis. The inqui-ry is not whether political classificationswere used. The inquiry instead is whetherpolitical classifications were used to bur-den a group’s representational rights. If acourt were to find that a State did imposeburdens and restrictions on groups or per-sons by reason of their views, there wouldlikely be a First Amendment violation, un-less the State shows some compelling in-terest. Of course, all this depends first oncourts’ having available a manageablestandard by which to measure the effect ofthe apportionment and so to conclude that


the State did impose a burden or restric-tion on the rights of a party’s voters.

Where it is alleged that a gerrymanderhad the purpose and effect of imposingburdens on a disfavored party and its vot-ers, the First Amendment may offer asounder and more prudential basis for in-tervention than does the Equal ProtectionClause. The equal protection analysisputs its emphasis on the permissibility ofan enactment’s classifications. This workswhere race is involved since classifying byrace is almost never permissible. It pres-ents a more complicated question when theinquiry is whether a generally permissibleclassification has been used for an imper-missible purpose. That question can onlybe answered in the affirmative by the sub-sidiary showing that the classification asapplied imposes unlawful burdens. TheFirst Amendment analysis concentrates onwhether the legislation burdens the repre-sentational rights of the complaining par-ty’s voters for reasons of ideology, beliefs,or political association. The analysis al-lows a pragmatic or functional assessmentthat accords some latitude to the States.See Eu v. San Francisco S 316County Demo-cratic Central Comm., 489 U.S. 214, 109S.Ct. 1013, 103 L.Ed.2d 271 (1989);Anderson v. Celebrezze, 460 U.S. 780, 103S.Ct. 1564, 75 L.Ed.2d 547 (1983).

Finally, I do not understand the plurali-ty to conclude that partisan gerrymander-ing that disfavors one party is permissible.Indeed, the plurality seems to acknowl-edge it is not. See ante, at 1785 (‘‘We donot disagree with [the] judgment’’ that‘‘partisan gerrymanders [are incompatible]with democratic principles’’); ibid. (notingthat it is the case, and that the pluralityopinion assumes it to be the case, that ‘‘anexcessive injection of politics [in district-ing] is unlawful’’). This is all the morereason to admit the possibility of latersuits, while holding just that the parties

have failed to prove, under our ‘‘well devel-oped and familiar’’ standard, that theselegislative classifications ‘‘reflec[t] no poli-cy, but simply arbitrary and capriciousaction.’’ Baker, 369 U.S., at 226, 82 S.Ct.691. That said, courts must be cautiousabout adopting a standard that turns onwhether the partisan interests in the redis-tricting process were excessive. Exces-siveness is not easily determined. Consid-er these apportionment schemes: In oneState, Party X controls the apportionmentprocess and draws the lines so it capturesevery congressional seat. In three otherStates, Party Y controls the apportionmentprocess. It is not so blatant or egregious,but proceeds by a more subtle effort, cap-turing less than all the seats in each State.Still, the total effect of Party Y’s effort isto capture more new seats than Party Xcaptured. Party X’s gerrymander wasmore egregious. Party Y’s gerrymanderwas more subtle. In my view, however,each is culpable.

* * *

The ordered working of our Republic,and of the democratic process, depends ona sense of decorum and restraint in allbranches of government, and in the citi-zenry itself. Here, one has the sense thatlegislative restraint was abandoned. Thatshould not be thought to serve the inter-ests of our political order. Nor should itbe thought to serve S 317our interest in dem-onstrating to the world how democracyworks. Whether spoken with concern orpride, it is unfortunate that our legislatorshave reached the point of declaring that,when it comes to apportionment: ‘‘ ‘We arein the business of rigging elections.’ ’’Hoeffel, Six Incumbents Are a Week Awayfrom Easy Election, Winston–Salem Jour-nal, Jan. 27, 1998, p. B1 (quoting a NorthCarolina state senator).

Still, the Court’s own responsibilities re-quire that we refrain from intervention in


541 U.S. 318

this instance. The failings of the manyproposed standards for measuring the bur-den a gerrymander imposes on representa-tional rights make our intervention im-proper. If workable standards do emergeto measure these burdens, however, courtsshould be prepared to order relief. Withthese observations, I join the judgment ofthe Court.

Justice STEVENS, dissenting.

The central question presented by thiscase is whether political gerrymanderingclaims are justiciable. Although our rea-sons for coming to this conclusion differ,five Members of the Court are convincedthat the plurality’s answer to that questionis erroneous. Moreover, as is apparentfrom our separate writings today, we sharethe view that, even if these appellants arenot entitled to prevail, it would be contraryto precedent and profoundly unwise toforeclose all judicial review of similarclaims that might be advanced in the fu-ture. That we presently have somewhatdiffering views—concerning both the prec-edential value of some of our recent casesand the standard that should be applied infuture cases—should not obscure the factthat the areas of agreement set forth inthe separate opinions are of far greatersignificance.

The concept of equal justice under lawrequires the State to govern impartially.See Romer v. Evans, 517 U.S. 620, 623,116 S.Ct. 1620, 134 L.Ed.2d 855 (1996);Lehr v. Robertson, 463 U.S. 248, 265, 103S.Ct. 2985, 77 L.Ed.2d 614 (1983); NewYork City Transit Authority v. Beazer,440 U.S. 568, 587, 99 S.Ct. 1355, 59L.Ed.2d 587 S 318(1979). Today’s pluralityopinion would exempt governing officialsfrom that duty in the context of legislative

redistricting and would give license, forthe first time, to partisan gerrymandersthat are devoid of any rational justification.In my view, when partisanship is the legis-lature’s sole motivation—when any pre-tense of neutrality is forsaken unabashedlyand all traditional districting criteria aresubverted for partisan advantage—thegoverning body cannot be said to haveacted impartially.

Although we reaffirm the central hold-ing of the Court in Davis v. Bandemer,478 U.S. 109, 106 S.Ct. 2797, 92 L.Ed.2d 85(1986), we have not reached agreement onthe standard that should govern partisangerrymandering claims. I would decidethis case on a narrow ground. Plaintiffs-appellants urge us to craft new rules thatin effect would authorize judicial review ofstatewide election results to protect thedemocratic process from a transient ma-jority’s abuse of its power to define votingdistricts. I agree with the plurality’s re-fusal to undertake that ambitious project.Ante, at 1780–1784. I am persuaded, how-ever, that the District Court failed to applywell-settled propositions of law when itgranted the defendants’ motion to dismissplaintiff-appellant Susan Furey’s gerry-mandering claim.

According to the complaint, Furey is aregistered Democrat who resides at anaddress in Montgomery County, Pennsyl-vania, that was located under the 1992districting plan in Congressional District13.1 Under the new plan adopted by theGeneral Assembly in 2002, Furey’s addressnow places her in the ‘‘non-compact’’ Dis-trict 6.2 Furey alleges that the new dis-tricting plan was created ‘‘solely’’ to effec-tuate the interests of Republicans,3 andthat the General Assembly relied ‘‘exclu-sively’’ on a principle of ‘‘maximum parti-

1. App. to Juris. Statement 129a.

2. Ibid.

3. Id., at 142a.


san advantage’’ when drawing the plan.4

In my judgment, Furey’s S 319allegations areplainly sufficient to establish: (1) that shehas standing to challenge the constitution-ality of District 6; (2) that her district-specific claim is not foreclosed by the Ban-demer plurality’s rejection of a statewideclaim of political gerrymandering; and (3)that she has stated a claim that, at leastwith respect to District 6, Pennsylvania’sredistricting plan violates the equal protec-tion principles enunciated in our votingrights cases both before and after Ban-demer. The District Court thereforeerred when it granted the defendants’ mo-tion to dismiss Furey’s claim.

I

Prior to our seminal decision in Baker v.Carr, 369 U.S. 186, 82 S.Ct. 691, 7 L.Ed.2d663 (1962), a majority of this Court hadheeded Justice Frankfurter’s repeatedwarnings about the dire consequences ofentering the ‘‘political thicket’’ of legisla-tive districting. Colegrove v. Green, 328U.S. 549, 556, 66 S.Ct. 1198, 90 L.Ed. 1432(1946). As a result, even the most egre-gious gerrymanders were sheltered fromjudicial review.5 It was after Baker thatwe first decided that the Constitution pro-hibits legislators from drawing districtlines that diminish the value of individualvotes in overpopulated districts. In reach-

ing that concluSsion,320 we explained that‘‘legislatures TTT should be bodies whichare collectively responsive to the popularwill,’’ Reynolds v. Sims, 377 U.S. 533, 565,84 S.Ct. 1362, 12 L.Ed.2d 506 (1964), andwe accordingly described ‘‘the basic aim oflegislative apportionment’’ as ‘‘achievingTTT fair and effective representation for allcitizens,’’ id., at 565–566, 84 S.Ct. 1362.Consistent with that goal, we also re-viewed claims that the majority had dis-criminated against particular groups ofvoters by drawing multimember districtsthat threatened ‘‘to minimize or cancel outthe voting strength of racial or politicalelements of the voting population.’’ Fort-son v. Dorsey, 379 U.S. 433, 439, 85 S.Ct.498, 13 L.Ed.2d 401 (1965). Such districtswere ‘‘vulnerable’’ to constitutional chal-lenge ‘‘if racial or political groups ha[d]been fenced out of the political process andtheir voting strength invidiously mini-mized.’’ Gaffney v. Cummings, 412 U.S.735, 754, 93 S.Ct. 2321, 37 L.Ed.2d 298(1973). See also Whitcomb v. Chavis, 403U.S. 124, 143, 91 S.Ct. 1858, 29 L.Ed.2d363 (1971); Burns v. Richardson, 384 U.S.73, 88, 86 S.Ct. 1286, 16 L.Ed.2d 376(1966).

Our holding in Bandemer, 478 U.S., at118–127, 106 S.Ct. 2797, that partisan ger-rymandering claims are justiciable fol-lowed ineluctably from the central reason-

4. Id., at 143a.

5. In Colegrove, for example, the Illinois Legis-lature had drawn the State’s district linesunder the 1901 State Apportionment Act andhad not reapportioned in the four ensuingdecades, ‘‘despite census figures indicatinggreat changes in the distribution of the popu-lation.’’ 328 U.S., at 569, 66 S.Ct. 1198(Black, J., dissenting). The populations ofIllinois’ districts in 1945 consequently rangedfrom 112,000 in the least populous district to900,000 in the most. Ibid. Nonetheless, theCourt, per Justice Frankfurter, concluded that‘‘due regard for the effective working of ourGovernment revealed this issue to be of a

peculiarly political nature and therefore notmeet for judicial determination.’’ Id., at 552,66 S.Ct. 1198. Fewer than 20 years later, theCourt, confronted with a strikingly similar setof facts—a Tennessee apportionment plan setby a 1901 statute that had remained virtuallyunchanged despite dramatic populationgrowth—held, in obvious tension with Cole-grove, that the complaint stated a justiciablecause of action. Baker, 369 U.S., at 192, 197–198, 82 S.Ct. 691. The Court distinguishedColegrove as simply ‘‘a refusal to exercise eq-uity’s powers.’’ 369 U.S., at 235, 82 S.Ct.691.


541 U.S. 322

ing in Baker, 369 U.S. 186, 82 S.Ct. 691, 7L.Ed.2d 663. What was true in Baker isno less true in this context:

‘‘The question here is the consistency ofstate action with the Federal Constitu-tion. We have no question decided, orto be decided, by a political branch ofgovernment coequal with this Court.Nor do we risk embarrassment of ourgovernment abroad, or grave distur-bance at home if we take issue with[Pennsylvania] as to the constitutionalityof her action here challenged. Nor needthe appellants, in order to succeed inthis action, ask the Court to enter uponpolicy determinations for which judicial-ly manageable standards are lacking.Judicial standards under the Equal Pro-tection Clause are well developed andfamiliar, and it has been open to courtssince the enactment of the FourteenthAmendment to determine, if on the par-ticular facts they must, that a discrimi-nation reflects no policy, but simplyarSbitrary321 and capricious action.’’ Id.,at 226, 82 S.Ct. 691 (footnote omitted).

‘‘[T]hat the [gerrymandering] claim is sub-mitted by a political group, rather than aracial group, does not distinguish [thecases] in terms of justiciability.’’ Bandem-er, 478 U.S., at 125, 106 S.Ct. 2797.

At issue in this case, as the pluralitystates, ante, at 1776, is Baker’s secondtest—the presence or absence of judiciallymanageable standards. The judicial stan-dards applicable to gerrymandering claimsare deeply rooted in decisions that longpreceded Bandemer and have been refinedin later cases. Among those well-settledprinciples is the understanding that a dis-trict’s peculiar shape might be a symptomof an illicit purpose in the line-drawingprocess. Most notably, in Gomillion v.Lightfoot, 364 U.S. 339, 340, 81 S.Ct. 125, 5L.Ed.2d 110 (1960), the Court invalidatedan Alabama statute that altered the

boundaries of the city of Tuskegee ‘‘from asquare to an uncouth twenty-eight-sidedfigure’’ for the sole purpose of preventingAfrican–Americans from voting in munici-pal elections. The allegations of bizarreshape and improper motive, ‘‘if proven,would abundantly [have] establish[ed] thatAct 140 was not an ordinary geographicredistricting measure even within familiarabuses of gerrymandering.’’ Id., at 341, 81S.Ct. 125. Justice Fortas’ concurringopinion in Kirkpatrick v. Preisler, 394 U.S.526, 538, 89 S.Ct. 1225, 22 L.Ed.2d 519(1969), which referred to gerrymanderingas ‘‘the deliberate and arbitrary distortionof district boundaries and populations forpartisan or personal political purposes,’’also identified both shape and purpose asrelevant standards. The maps attached asexhibits in Gomillion, 364 U.S., at 348, 81S.Ct. 125 (Appendix to opinion of theCourt), and in subsequent voting rightscases demonstrate that an ‘‘uncouth’’ orbizarre shape can easily identify a districtdesigned for a single-minded, nonneutralpurpose.

With purpose as the ultimate inquiry,other considerations have supplied readystandards for testing the lawfulness of agerrymander. In his dissent in Bandem-er, Justice Powell S 322explained that ‘‘themerits of a gerrymandering claim must bedetermined by reference to the configura-tions of the districts, the observance ofpolitical subdivision lines, and other crite-ria that have independent relevance to thefairness of redistricting.’’ 478 U.S., at 165,106 S.Ct. 2797. Applying this three-partstandard, Justice Powell first reviewed theprocedures used in Indiana’s redistrictingprocess and noted that the party in powerhad excluded the opposition from its delib-erations and had placed excessive weighton data concerning party voting trends.Id., at 175–176, 106 S.Ct. 2797. Second,Justice Powell pointed to the strangeshape of districts that conspicuously ig-


nored traditional districting principles.Id., at 176–177, 106 S.Ct. 2797. He notedthe impact of such shapes on residents ofthe uncouth districts,6 and he included inhis opinion maps that illustrated the irreg-ularity of the district shapes, id., at 181,183, 106 S.Ct. 2797. Third and finally,Justice Powell reviewed other ‘‘substantialevidence,’’ including contemporaneousstatements and press accounts, demon-strating that the architects of the districts‘‘were motivated solely by partisan consid-erations.’’ Id., at 177, 106 S.Ct. 2797.

The Court has made use of all threeparts of Justice Powell’s standard in itsrecent racial gerrymandering jurispru-dence. In those cases, the Court has ex-amined claims that redistricting schemesviolate the equal protection guaranteewhere they are ‘‘so highly irregular’’ ontheir face that they ‘‘rationally cannot beunderstood as anything other than an ef-fort’’ to segregate voters by race, Shaw v.Reno, 509 U.S. 630, 646–647, 113 S.Ct.2816, 125 L.Ed.2d 511 (1993) (Shaw I), or

where ‘‘race for its own sake, and notother districting principles, was the legisla-ture’s dominant and controlling rationalein drawing its district lines,’’ Miller v.Johnson, 515 U.S. 900, 913, 115 S.Ct. 2475,132 L.Ed.2d 762 (1995). See S 323also Eas-ley v. Cromartie, 532 U.S. 234, 241, 121S.Ct. 1452, 149 L.Ed.2d 430 (2001); Shawv. Hunt, 517 U.S. 899, 905, 116 S.Ct. 1894,135 L.Ed.2d 207 (1996) (Shaw II).7 TheShaw line of cases has emphasized that‘‘reapportionment is one area in which ap-pearances do matter,’’ Shaw I, 509 U.S., at647, 113 S.Ct. 2816, and has focused bothon the shape of the challenged districtsand the purpose behind the line-drawing inassessing the constitutionality of majority-minority districts under the Equal Protec-tion Clause. These decisions, like JusticePowell’s opinion in Bandemer, have alsoconsidered the process by which the dis-tricting schemes were enacted,8 looked toother evidence demonstrating that purelyimproper considerations motivated the de-cision,9 and included maps illustrating out-landish district shapes.10

6. ‘‘ ‘[T]he potential for voter disillusion andnonparticipation is great,’ as voters are forcedto focus their political activities in artificialelectoral units. Intelligent voters, regardlessof party affiliation, resent this sort of politicalmanipulation of the electorate for no publicpurpose.’’ 478 U.S., at 177, 106 S.Ct. 2797(citation omitted).

7. The reasoning in these decisions followednot only from Gomillion v. Lightfoot, 364 U.S.339, 81 S.Ct. 125, 5 L.Ed.2d 110 (1960), seeShaw I, 509 U.S., at 644–645, 113 S.Ct. 2816(relying on Gomillion ), but also from JusticePowell’s observation in Davis v. Bandemer,478 U.S. 109, 173, n. 12, 106 S.Ct. 2797, 92L.Ed.2d 85 (1986), that ‘‘[i]n some cases,proof of grotesque district shapes may, with-out more, provide convincing proof of uncon-stitutional gerrymandering.’’

8. In Miller v. Johnson, 515 U.S. 900, 917–919,115 S.Ct. 2475, 132 L.Ed.2d 762 (1995), theCourt reviewed the procedures followed bythe Georgia Legislature in responding to the

Justice Department’s objections to its originalplan, and the part that the operator of its‘‘reapportionment computer’’ played in de-signing the districts, to support its conclusion‘‘that the legislature subordinated traditionaldistricting principles to race.’’ See also Bushv. Vera, 517 U.S. 952, 961–962, 116 S.Ct.1941, 135 L.Ed.2d 248 (1996) (plurality opin-ion) (discussing use of computer program tomanipulate district lines).

9. In Shaw II, 517 U.S. 899, 910, 116 S.Ct.1894, 135 L.Ed.2d 207 (1996), for instance,the Court considered the fact that certainreports regarding the effects of past discrimi-nation were not before the legislature andtherefore could not have played a role in thedistricting process.

10. Hunt v. Cromartie, 526 U.S. 541, 554, 119S.Ct. 1545, 143 L.Ed.2d 731 (1999); Bush v.Vera, 517 U.S., at 986, 116 S.Ct. 1941 (plural-ity opinion); Miller, 515 U.S., at 928, 115S.Ct. 2475; Shaw I, 509 U.S., at 659, 113S.Ct. 2816.


541 U.S. 325

Given this clear line of precedents, Ishould have thought the question of justici-ability in cases such as this—where a setof plaintiffs argues that a single motivationresulted in a districting scheme with dis-criminatory effects—to be well settled.The plurality’s contrary conclusion cannotbe S 324squared with our long history ofvoting rights decisions. Especially per-plexing is the plurality’s ipse dixit distinc-tion of our racial gerrymandering cases.Notably, the plurality does not argue thatthe judicially manageable standards thathave been used to adjudicate racial gerry-mandering claims would not be equallymanageable in political gerrymanderingcases. Instead, its distinction of thosecases rests on its view that race as adistricting criterion is ‘‘much more rarelyencountered’’ than partisanship, ante, at1781, and that determining whether race—‘‘a rare and constitutionally suspect mo-tive’’—dominated a districting decision ‘‘isquite different from determining whether[such a decision] is so substantially affect-ed by the excess of an ordinary and lawfulmotive as to [be] invali[d],’’ ibid. Butthose considerations are wholly irrelevantto the issue of justiciability.

To begin with, the plurality errs in as-suming that politics is ‘‘an ordinary andlawful motive.’’ We have squarely reject-ed the notion that a ‘‘purpose to discrimi-nate on the basis of politics,’’ ante, at 1781,1785, is never subject to strict scrutiny.On the contrary, ‘‘political belief and asso-ciation constitute the core of those activi-ties protected by the First Amendment,’’Elrod v. Burns, 427 U.S. 347, 356, 96 S.Ct.

2673, 49 L.Ed.2d 547 (1976) (plurality opin-ion), and discriminatory governmental de-cisions that burden fundamental FirstAmendment interests are subject to strictscrutiny, id., at 363, 96 S.Ct. 2673; cf.Police Dept. of Chicago v. Mosley, 408 U.S.92, 94–95, 92 S.Ct. 2286, 33 L.Ed.2d 212(1972). Thus, unless party affiliation is anappropriate requirement for the position inquestion, government officials may notbase a decision to hire, promote, transfer,recall, discharge, or retaliate against anemployee, or to terminate a contract, onthe individual’s partisan affiliation orspeech. See Board of Comm’rs, Wabaun-see Cty. v. Umbehr, 518 U.S. 668, 674–675,116 S.Ct. 2342, 135 L.Ed.2d 843 (1996);O’Hare Truck Service, Inc. v. City ofNorthlake, 518 U.S. 712, 716–717, 116 S.Ct.2353, 135 L.Ed.2d 874 (1996); Rutan v.Republican Party of Ill., 497 U.S. 62, 64–65, 110 S.Ct. 2729, 111 L.Ed.2d 52 (1990);Branti v. Finkel, 445 U.S. 507, S 325519–520,100 S.Ct. 1287, 63 L.Ed.2d 574 (1980);Elrod, 427 U.S., at 355–363, 96 S.Ct.2673.11 It follows that political affiliation isnot an appropriate standard for excludingvoters from a congressional district.

The plurality argues that our patronagecases do not support the proposition thatstrict scrutiny should be applied in politicalgerrymandering cases because ‘‘[i]t is ele-mentary that scrutiny levels are claim spe-cific.’’ Ante, at 1786. It is also elementa-ry, however, that the level of scrutiny isrelevant to the question whether there hasbeen a constitutional violation, not the

11. The plurality opinion seems to assume thatthe dissenting opinions in Umbehr, 518 U.S.,at 686, 116 S.Ct. 2342 (SCALIA, J.), and Ru-tan, 497 U.S., at 92, 110 S.Ct. 2729 (SCALIA,J.), correctly state the law—namely, that‘‘when a practice not expressly prohibited bythe text of the Bill of Rights bears the en-dorsement of a long tradition of open, wide-spread, and unchallenged use that dates backto the beginning of the Republic, we have no

proper basis for striking it down,’’ id., at 95,110 S.Ct. 2729. Cf. ante, at 1774–1775 (trac-ing the history of political gerrymanders tothe beginning of the 18th century). But‘‘[o]ur inquiry does not begin with the judg-ment of history’’; ‘‘[r]ather, inquiry mustcommence with identification of the constitu-tional limitations implicated by a challengedgovernmental practice.’’ Elrod, 427 U.S., at354–355, 96 S.Ct. 2673.


question of justiciability.12 The standardsoutlined above are discernible and judicial-ly manageable regardless of the number ofcases in which they must be applied or thelevel of scrutiny at which the analysis oc-curs.13 Thus, the dicta from Shaw I andBush v. Vera, 517 U.S. 952, 116 S.Ct. 1941,135 L.Ed.2d 248 (1996), on which the plu-rality relies, ante, at 1785, are beside thepoint, because they speak not at all to thesubject of justiciability. And while ofcourse a difference exists beStween326 theconstitutional interests protected by theFirst and Fourteenth Amendments, therelevant lesson of the patronage cases isthat partisanship is not always as benign aconsideration as the plurality appears toassume. In any event, as I understandthe plurality’s opinion, it seems to agreethat if the State goes ‘‘too far’’—if it en-gages in ‘‘political gerrymandering for pol-itics’ sake’’—it violates the Constitution inthe same way as if it undertakes ‘‘racialgerrymandering for race’s sake.’’ Ibid.But that sort of constitutional violationcannot be touched by the courts, the plu-rality maintains, because the judicial obli-gation to intervene is ‘‘dubious.’’ Ante, at1781.14

State action that discriminates againsta political minority for the sole and una-dorned purpose of maximizing the powerof the majority plainly violates the deci-sionmaker’s duty to remain impartial.See, e.g., Lehr, 463 U.S., at 265, 103S.Ct. 2985. Gerrymanders necessarilyrest on legislators’ predictions that‘‘members of certain identifiable groupsTTT will vote in the same way.’’ Mobilev. Bolden, 446 U.S. 55, 87, 100 S.Ct.1490, 64 L.Ed.2d 47 (1980) (STEVENS,J., concurring in judgment). ‘‘In theline-drawing process, racial, religious,ethnic, and economic gerrymanders areall species of political gerrymanders.’’Id., at 88, 100 S.Ct. 1490. Thus, the crit-ical issue in both racial and political ger-rymandering cases is the same: whethera single nonneutral criterion controlledthe districting process to such an extentthat the Constitution was offended. ThisCourt has treated that precise questionas justiciable in Gomillion and in theShaw line of cases, and today’s pluralityhas supplied no persuasive reason S 327fordistinguishing the justiciability of partisangerrymanders. Those cases confirm andreinforce the holding that partisan gerry-mandering claims are justiciable.15

12. It goes without saying that a claim thatotherwise would trigger strict scrutiny mightnonetheless be nonjusticiable. See, e.g., Allenv. Wright, 468 U.S. 737, 104 S.Ct. 3315, 82L.Ed.2d 556 (1984); DeFunis v. Odegaard,416 U.S. 312, 94 S.Ct. 1704, 40 L.Ed.2d 164(1974) (per curiam).

13. The plurality explains that it is willing to‘‘accep[t] a modest degree of unmanageabili-ty’’ where the ‘‘constitutional command TTT isclear,’’ but not where the ‘‘constitutional obli-gation TTT is both dubious and severely un-manageable.’’ Ante, at 1781. Not only doesthis statement cast doubt on the plurality’sfaith in our racial gerrymandering cases, butits reasoning is clearly tautological.

14. The plurality’s reluctance to recognize thejusticiability of partisan gerrymanders seemsdriven in part by a fear that recognizing such

claims will give rise to a flood of litigation.See ante, at 1781. But the list of cases that itcites in its lengthy footnote 6, ante, at 1778,suggests that in the two decades since Ban-demer, there has been an average of just threeor four partisan gerrymandering cases filedevery year. That volume is obviously trivialwhen compared, for example, to the amountof litigation that followed our adoption of the‘‘one-person, one-vote’’ rule. See Reynolds v.Sims, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d506 (1964).

15. Writing for the Court in Bandemer, JusticeWhite put it well: ‘‘That the characteristics ofthe complaining group are not immutable orthat the group has not been subject to thesame historical stigma may be relevant to themanner in which the case is adjudicated, butthese differences do not justify a refusal to


541 U.S. 328

II

The plurality opinion in Bandemer dealtwith a claim that the Indiana apportion-ment scheme for state legislative districtsdiscriminated against Democratic voterson a statewide basis. 478 U.S., at 127, 106S.Ct. 2797. In my judgment, the Bandem-er Court was correct to entertain thatstatewide challenge, because the plaintiffsin that case alleged a group harm thataffected members of their party through-out the State. In the subsequent line ofracial gerrymandering cases, however, theCourt shifted its focus from statewide chal-lenges and required, as a matter of stand-ing, that plaintiffs stating race-based equalprotection claims actually reside in the dis-tricts they are challenging. See UnitedStates v. Hays, 515 U.S. 737, 745, 115 S.Ct.2431, 132 L.Ed.2d 635 (1995). BecauseHays has altered the standing rules forgerrymandering claims—and because, inmy view, racial and political gerrymandersare species of the same constitutional con-cern—the Hays standing rule requires dis-missal of the statewide claim.16 But thatdoes not S 328end the matter. Challenges tospecific districts, such as those consideredin the Shaw cases, relate to a differenttype of ‘‘representational’’ harm, and thoseallegations necessarily must be consideredon a district-by-district basis. The com-

plaint in this case alleges injuries of bothtypes—a group harm to Democratic votersthroughout Pennsylvania and a more indi-vidualized representational injury to Fureyas a resident of District 6.

In a challenge to a statewide districtingplan, plaintiffs-appellants complain thatthey have been injured because of theirmembership in a particular, identifiablegroup. The plaintiffs-appellees in Ban-demer, for example, alleged ‘‘that Demo-cratic voters over the State as a whole, notDemocratic voters in particular districts,ha[d] been subjected to unconstitutionaldiscrimination.’’ 478 U.S., at 127, 106S.Ct. 2797 (citing complaint). They specif-ically claimed that they were injured asmembers of a group because the numberof Democratic representatives was notcommensurate with the number of Demo-cratic voters throughout Indiana. Muchlike the plaintiffs-appellees in Bandemer,plaintiffs-appellants in this case allege thatthe statewide plan will enable Republicans,who constitute about half of Pennsylvania’svoters, to elect 13 or 14 members of theState’s 19–person congressional delega-tion.17 Under Hays, however, plaintiffs-appellants lack standing to challenge thedistricting plan on a statewide basis. 515U.S., at 744–745, 115 S.Ct. 2431.18

entertain such a case.’’ 478 U.S., at 125, 106S.Ct. 2797.

16. The cases that the plurality cites today,ante, at 1778, n. 6, support the conclusionthat it would have been wise to endorse theviews expressed in Justice Powell’s dissent inBandemer, 478 U.S., at 161, 106 S.Ct. 2797,and my concurrence in Karcher v. Daggett,462 U.S. 725, 744, 103 S.Ct. 2653, 77 L.Ed.2d133 (1983). I remain convinced that ouropinions correctly interpreted the law. If thatstandard were applied to the statewide chal-lenge in this case, a trial of the entire casewould be required. For the purpose of decid-ing this case, even though I dissented from

our decision in Shaw I and remain convincedthat it was incorrectly decided, I would givethe Shaw cases stare decisis effect in the polit-ical gerrymandering context. Given theCourt’s illogical disposition of this case, how-ever, in future cases I would feel free toreexamine the standing issue. I surely wouldnot suggest that a plaintiff would never havestanding to litigate a statewide claim.


18. As the Court explained in Hays, ‘‘[v]otersin [gerrymandered] districts may suffer thespecial representational harms [that constitu-tionally suspect] classifications can cause inthe voting context. On the other hand, where


A challenge to a specific district or dis-tricts, on the other hand, alleges a differ-ent type of injury entirely—one thatS 329our recent racial gerrymandering caseshave recognized as cognizable.19 In ShawI we held that ‘‘a plaintiff challenging a re-apportionment statute under the EqualProtection Clause may state a claim by al-leging that the legislation, though raceneutral on its face, rationally cannot be un-derstood as anything other than an effortto separate voters into different districtson the basis of race.’’ 509 U.S., at 649, 113S.Ct. 2816. After describing the perni-cious consequences of race-conscious dis-tricting—even when designed to enhancethe representation of the minority—andafter explaining why dramatically irregu-lar shapes ‘‘ ‘have sufficient probativeforce to call for an explanation,’ ’’ id., at647, 113 S.Ct. 2816 (quoting Karcher v.Daggett, 462 U.S. 725, 755, 103 S.Ct. 2653,77 L.Ed.2d 133 (1983) (STEVENS, J., con-curring)), we described the message a mis-shapen district sends to elected officials:

‘‘When a district obviously is createdsolely to effectuate the perceived com-mon interests of one racial group, elect-ed officials are more likely to believethat their primary obligation is to repre-sent only the members of that group,rather than their constituency as awhole. This is altogether antithetical toour system of representative democra-cy.’’ Shaw I, 509 U.S., at 648, 113 S.Ct.2816.

Undergirding the Shaw cases is the prem-ise that racial gerrymanders effect a con-stitutional wrong when they disrupt therepresentational norms that ordinarilytether elected officials to their constituen-cies as a whole.

‘‘[L]egislatures,’’ we have explained,‘‘should be bodies which are collectivelyresponsive to the popular will,’’ Reynolds,377 U.S., at 565, 84 S.Ct. 1362, for ‘‘[l]egis-lators are elected by voters, S 330not farmsor cities or economic interests,’’ id., at 562,84 S.Ct. 1362.20 Gerrymanders subvertthat representative norm because the win-ner of an election in a gerrymandereddistrict inevitably will infer that her suc-cess is primarily attributable to the archi-tect of the district rather than to a constit-uency defined by neutral principles. TheShaw cases hold that this disruption of therepresentative process imposes a cogniza-ble ‘‘representational har[m].’’ Hays, 515U.S., at 745, 115 S.Ct. 2431. Because thatharm falls squarely on the voters in thedistrict whose representative might ordoes misperceive the object of her fealty,the injury is cognizable only when statedby voters who reside in that particulardistrict, see Shaw II, 517 U.S., at 904, 116S.Ct. 1894; otherwise the ‘‘plaintiff wouldbe asserting only a generalized grievanceagainst governmental conduct of which heor she does not approve,’’ Hays, 515 U.S.,at 745, 115 S.Ct. 2431. See also Bush v.Vera, 517 U.S., at 957–958, 116 S.Ct. 1941(plurality opinion).

a plaintiff does not live in such a district, heor she does not suffer those special harmsTTTT’’ 515 U.S., at 745, 115 S.Ct. 2431.

19. The plurality in Bandemer, 478 U.S., at127, 106 S.Ct. 2797, itself acknowledged that‘‘the focus of the equal protection inquiry’’ ina statewide challenge ‘‘is necessarily some-what different from that involved in the re-view of individual districts.’’

20. Cf. McConnell v. Federal Election Comm’n,540 U.S. 93, 153, 124 S.Ct. 619, 666, 157L.Ed.2d 491 (2003) (‘‘Just as troubling to afunctioning democracy as classic quid pro quocorruption is the danger that officeholderswill decide issues not on the merits or thedesires of their constituencies, but accordingto the wishes of those who have made largefinancial contributions valued by the office-holder’’).


541 U.S. 331

Although the complaint in this case in-cludes a statewide challenge, plaintiff-ap-pellant Furey states a stronger claim as aresident of the misshapen District 6.21 Shecomplains not merely about the injury re-sulting from the probable election of acongressional delegation that does not fair-ly repreSsent331 the entire State, or aboutthe harm flowing from the probable elec-tion of a Republican to represent District6.22 She also alleges that the grotesqueconfiguration of that district itself imposesa special harm on the members of thepolitical minority residing in District 6 thatdirectly parallels the harm recognized inShaw I. Officials elected by the majorityparty in such a district, she claims, ‘‘aremore likely to believe that their primaryobligation is to represent only the mem-bers of that group, rather than the constit-uency as a whole.’’ 23 This is precisely theharm that the Shaw cases treat as cogniza-

ble in the context of racial gerrymander-ing. The same treatment is warranted inthis case.

The risk of representational harms iden-tified in the Shaw cases is equally great, ifnot greater, in the context of partisan ger-rymanders. Shaw I was borne of the con-cern that an official elected from a raciallygerrymandered district will feel beholdenonly to a portion of her constituents, andthat those constituents will be defined byrace. 509 U.S., at 648, 113 S.Ct. 2816.The parallel danger of a partisan gerry-mander is that the representative will per-ceive that the people who put her in powerare those who drew the map rather thanthose who cast ballots, and she will feelbeholden not to a subset of her constituen-cy, but to no part of her constituency atall.24 The problem, simply put, is that thewill of the cartographers rather than thewill of the people will govern.25 As Judge

21. Plaintiffs-appellants Richard and NormaJean Vieth are registered Democrats who re-side in District 16. App. to Juris. Statement129a. The complaint does not claim that theyresided in a different district under the olddistricting scheme, nor does it anywhere al-lege, as it does on Furey’s behalf, that District16 in particular is irregularly shaped. Aglance at the appended map, infra, revealsthat District 16 is not especially unusual in itscontours. Without more specific allegationsregarding District 16, I would limit the analy-sis to District 6.

22. When her residence was located in District13, Furey was represented by a Democrat.App. 261.


24. ‘‘[A]mple evidence demonstrates that manyof today’s congressional representatives owetheir election not to ‘the People of the severalstates’ but to the mercy of state legislatures.’’Note, 117 Harv. L.Rev. 1196, 1202 (2004).

25. In this sense the partisan gerrymander isthe American cousin of the English ‘‘rottenborough.’’ In the English system, Members

of Parliament were elected from geographicunits that remained unchanged despite popu-lation changes wrought by the Industrial Rev-olution. ‘‘Because representation was notbased on population, vast inequities devel-oped over time in the form of the so-calledrotten boroughs. Old Sarum, for instance,had no human residents—only a few sheep—yet sent the same number of representativesto Parliament as Yorkshire, with nearly a mil-lion inhabitants.’’ R. Zagarri, The Politics ofSize: Representation in the United States,1776–1850, p. 37 (1987). As a result of thissystem, ‘‘many insignificant places returnedmembers, while many important towns didnot,’’ and ‘‘even in large towns the memberswere often elected by a tiny fraction of thepopulation.’’ J. Butler, The Passing of theGreat Reform Bill 176 (1914). Meanwhile,‘‘[t]he Government bribed the patron or mem-ber or both by means of distinctions andoffices or by actual cash,’’ and ‘‘[t]he patronand member bribed the electors in the sameway.’’ Ibid. The rotten boroughs clearlywould violate our familiar one-person, one-vote rule, but they were also troubling be-cause the representative of such a boroughowed his primary loyalty to his patron andthe government rather than to his constitu-ents (if he had any). Similarly, in gerryman-


S 332Ward recently wrote, ‘‘extreme partisangerrymandering leads to a system in whichthe representatives choose their constitu-ents, rather than vice-versa.’’ Session v.Perry, 298 F.Supp.2d 451, 516 (E.D.Tex.2004) (per curiam) (concurring in part anddissenting in part).

III

Elected officials in some sense serve twomasters: the constituents who electedthem and the political sponsors who sup-port them. Their primary obligations are,of course, to the public in general, but it isneither realistic nor fair to expect themwholly to ignore the political consequencesof their decisions. ‘‘It would be idle TTT tocontend that any political considerationtaken into account in fashioning a reappor-tionment plan is sufficient to invalidate it.’’Gaffney, 412 U.S., at 752, 93 S.Ct. 2321.Political factors are common and permissi-ble elements of the art of governing ademocratic society.

But while political considerations mayproperly influence the decisions of ourelected officials, when such decisions disSad-vantage333 members of a minority group—

whether the minority is defined by itsmembers’ race, religion, or political affilia-tion—they must rest on a neutral predi-cate. See Hampton v. Mow Sun Wong,426 U.S. 88, 100, 96 S.Ct. 1895, 48 L.Ed.2d495 (1976) (‘‘The federal sovereign, like theStates, must govern impartially’’); Ban-demer, 478 U.S., at 166, 106 S.Ct. 2797(Powell, J., dissenting). The Constitutionenforces ‘‘a commitment to the law’s neu-trality where the rights of persons are atstake.’’ Romer, 517 U.S., at 623, 116 S.Ct.1620. See also Board of Trustees of Univ.of Ala. v. Garrett, 531 U.S. 356, 375, 121S.Ct. 955, 148 L.Ed.2d 866 (2001) (KEN-NEDY, J., concurring) (‘‘States act as neu-tral entities, ready to take instruction andto enact laws when their citizens so de-mand’’). Thus, the Equal ProtectionClause implements a duty to govern im-partially that requires, at the very least,that every decision by the sovereign servesome nonpartisan public purpose.26

In evaluating a claim that a governmen-tal decision violates the Equal ProtectionClause, we have long required a showingof discriminatory purpose. See Washing-ton v. Davis, S 334426 U.S. 229, 96 S.Ct.

dered districts, instead of local groups definedby neutral criteria selecting their representa-tives, it is the architects of the districts whoselect the constituencies and, in effect, therepresentatives.

26. In the realm of federal elections, the re-quirement of governmental neutrality is but-tressed by this Court’s recognition that theElections Clause is not ‘‘ ‘a source of power todictate electoral outcomes, to favor or disfa-vor a class of candidates, or to evade impor-tant constitutional restraints.’ ’’ Cook v. Gra-like, 531 U.S. 510, 523, 121 S.Ct. 1029, 149L.Ed.2d 44 (2001) (quoting U.S. Term Limits,Inc. v. Thornton, 514 U.S. 779, 833–834, 115S.Ct. 1842, 131 L.Ed.2d 881 (1995)). Andthis duty to govern impartially extends to ex-ecutive and legislative officials alike. Begin-ning as early as its first session in 1789,Congress has passed a number of statutes

designed to guarantee that Executive Branchemployees neutrally carry out their duties.See Ex parte Curtis, 106 U.S. 371, 372–373, 1S.Ct. 381, 27 L.Ed. 232 (1882). Some ofthose laws avoided the danger that ‘‘the gov-ernment itself may be made to furnish indi-rectly the money to defray the expenses ofkeeping the political party in power that hap-pens to have for the time being the control ofthe public patronage.’’ Id., at 375, 1 S.Ct.381. It is ‘‘fundamental’’ that federal employ-ees ‘‘are expected to enforce the law andexecute the programs of the Governmentwithout bias or favoritism for or against anypolitical party or group or the members there-of.’’ Civil Service Comm’n v. Letter Carriers,413 U.S. 548, 564–565, 93 S.Ct. 2880, 37L.Ed.2d 796 (1973). That expectation reflectsthe principle that ‘‘the impartial execution ofthe laws’’ is a ‘‘great end of Government.’’Id., at 565, 93 S.Ct. 2880.


541 U.S. 335

2040, 48 L.Ed.2d 597 (1976).27 That re-quirement applies with full force to dis-tricting decisions. The line that divides aracial or ethnic minority unevenly betweenschool districts can be entirely legitimate ifchosen on the basis of neutral factors—county lines, for example, or a naturalboundary such as a river or major thor-oughfare. But if the district lines werechosen for the purpose of limiting thenumber of minority students in the school,or the number of families holding unpopu-lar religious or political views, that invidi-ous purpose surely would invalidate thedistrict. See Gomillion v. Lightfoot, 364U.S., at 344–345, 81 S.Ct. 125; cf. Boardof Ed. of Kiryas Joel Village School Dist.v. Grumet, 512 U.S. 687, 699–700, 114 S.Ct.2481, 129 L.Ed.2d 546 (1994).

Consistent with that principle, our re-cent racial gerrymandering cases have ex-amined the shape of the district and thepurpose of the districting body to deter-mine whether race, above all other criteria,predominated in the line-drawing process.We began by holding in Shaw I that adistricting scheme could be ‘‘so irrationalon its face that it [could] be understoodonly as an effort to segregate voters intoseparate voting districts because of theirrace.’’ 509 U.S., at 658, 113 S.Ct. 2816.Then, in Miller, we explained that ShawI’s irrational-shape test did not treat thebizarreness of a district’s lines itself as aconstitutional violation; rather, the irregu-larity of the district’s contours in Shaw Iwas ‘‘persuasive circumstantial evidencethat race for its own sake, and not otherdistricting principles, was the legislature’s

dominant and controlling rationale indrawing its district lines.’’ 515 U.S., atS 335913, 115 S.Ct. 2475. Under the Shawcases, then, the use of race as a criterion inredistricting is not per se impermissible,see Shaw I, 509 U.S., at 642, 113 S.Ct.2816; Shaw II, 517 U.S. 899, 116 S.Ct.1894, 135 L.Ed.2d 207, but when race iselevated to paramount status—when it isthe be-all and end-all of the redistrictingprocess—the legislature has gone too far.‘‘Race must not simply have been a moti-vation TTT but the predominant factor mo-tivating the legislature’s districting deci-sion.’’ Easley, 532 U.S., at 241, 121 S.Ct.1452 (internal quotation marks and cita-tions omitted).

Just as irrational shape can serve asan objective indicator of an impermissi-ble legislative purpose, other objectivefeatures of a districting map can savethe plan from invalidation. We haveexplained that ‘‘traditional districtingprinciples,’’ which include ‘‘compactness,contiguity, and respect for political sub-divisions,’’ are ‘‘important not becausethey are constitutionally required TTT

but because they are objective factorsthat may serve to defeat a claim that adistrict has been gerrymandered on ra-cial lines.’’ Shaw I, 509 U.S., at 647,113 S.Ct. 2816 (citing Gaffney, 412 U.S.,at 752, n. 18, 93 S.Ct. 2321; Karcher,462 U.S., at 755, 103 S.Ct. 2653 (STE-VENS, J., concurring)). ‘‘Where theseor other race-neutral considerations arethe basis for redistricting legislation,and are not subordinated to race, aState can ‘defeat a claim that a district

27. In Washington v. Davis, we referred to anearlier challenge to a New York reapportion-ment statute that had failed because the plain-tiffs had not shown that the statute was ‘‘ ‘theproduct of a state contrivance to segregate onthe basis of race or place of origin.’ ’’ 426U.S., at 240, 96 S.Ct. 2040 (quoting Wright v.Rockefeller, 376 U.S. 52, 58, 84 S.Ct. 603, 11

L.Ed.2d 512 (1964)). We emphasized that theCourt in Wright had been unanimous in iden-tifying the issue as ‘‘whether the ‘boundariesTTT were purposefully drawn on raciallines.’ ’’ 426 U.S., at 240, 96 S.Ct. 2040(quoting Wright, 376 U.S., at 67, 84 S.Ct.603).


has been gerrymandered on raciallines.’ ’’ Miller, 515 U.S., at 916, 115S.Ct. 2475 (quoting Shaw I, 509 U.S., at647, 113 S.Ct. 2816).

In my view, the same standards shouldapply to claims of political gerrymander-ing, for the essence of a gerrymander isthe same regardless of whether the groupis identified as political or racial. Gerry-mandering always involves the drawing ofdistrict boundaries to maximize the votingstrength of the dominant political factionand to minimize the strength of one ormore groups of opponents. Mobile, 446U.S., at 87, 100 S.Ct. 1490 (STEVENS, J.,concurring in judgment). In seeking thedesired result, legislators necessarily makejudgments about the probability that themembers of identifiable S 336groups—wheth-er economic, religious, ethnic, or racial—will vote in a certain way. The overridingpurpose of those predictions is political.See Karcher, 462 U.S., at 749–750, 103S.Ct. 2653 (STEVENS, J., concurring);Mobile, 446 U.S., at 88, 100 S.Ct. 1490(STEVENS, J., concurring in judgment).28

It follows that the standards that enablecourts to identify and redress a racial ger-rymander could also perform the samefunction for other species of gerryman-ders. See Bandemer, 478 U.S., at 125, 106

S.Ct. 2797; Cousins v. City Council ofChicago, 466 F.2d 830, 853 (C.A.7 1972)(STEVENS, J., dissenting).

The racial gerrymandering cases there-fore supply a judicially manageable stan-dard for determining when partisanship,like race, has played too great of a role inthe districting process. Just as race canbe a factor in, but cannot dictate the out-come of, the districting process, so too canpartisanship be a permissible considerationin drawing district lines, so long as it doesnot predominate. If, as plaintiff-appellantFurey has alleged, the predominant motiveof the legislators who designed District 6,and the sole justification for its bizarreshape, was a purpose to discriminateagainst a political minority, that invidiouspurpose should invalidate the district.

The plurality reasons that the standardsfor evaluating racial gerrymanders are notworkable in cases such as this becausepartisan considerations, unlike racial ones,are perfectly legitimate. Ante, at 1781.Until today, however, there has not beenthe slightest intimation in any opinionwritten by any Member of this Court thata naked purpose S 337to disadvantage a polit-ical minority would provide a rational basisfor drawing a district line.29 On the con-

28. I have elsewhere explained my view thatrace as a factor in the districting process is nodifferent from any other political consider-ation. Creating a majority-minority district isno better and no worse than creating anIrish–American, or Polish–American, or Ital-ian–American district. In all events the rele-vant question is whether the sovereign abro-gated its obligation to govern neutrally. SeeKarcher, 462 U.S., at 753–754, 103 S.Ct. 2653(STEVENS, J., concurring); Mobile, 446 U.S.,at 88, 100 S.Ct. 1490 (STEVENS, J., concur-ring in judgment); Cousins v. City Council ofChicago, 466 F.2d 830, 850–853 (C.A.7 1972)(STEVENS, J., dissenting).

29. The plurality’s long discussion of the histo-ry of political gerrymanders is interesting,ante, at 1774–1776, but it surely is not intend-

ed to suggest that the vintage of an invidiouspractice—even ‘‘an American political tradi-tion as old as the Republic,’’ Board ofComm’rs, Wabaunsee Cty. v. Umbehr, 518 U.S.668, 688, 116 S.Ct. 2342, 135 L.Ed.2d 843(1996) (SCALIA, J., dissenting)—should insu-late it from constitutional review. Compare,e.g., Bradwell v. State, 16 Wall. 130, 21 L.Ed.442 (1873), with Nevada Dept. of Human Re-sources v. Hibbs, 538 U.S. 721, 729, 123 S.Ct.1972, 155 L.Ed.2d 953 (2003). The historicaldiscussion might be relevant if it attempted tojustify political gerrymandering as an accept-able use of governmental power. In the end,however, the plurality’s defense of its positioncomes down to the unconvincing assertionthat it lacks the juridical capacity to adminis-ter the standards the Court fashioned in itsrecent racial gerrymandering jurisprudence.


541 U.S. 338

trary, our opinions referring to politicalgerrymanders have consistently assumedthat they were at least undesirable, and wealways have indicated that political consid-erations are among those factors that maynot dominate districting decisions.30 Pure-ly partisan motives are ‘‘rational’’ in a lit-eral sense, but there must be a limitingprinciple. ‘‘[T]he word ‘rational’—for meat least—includes elements of legitimacyand neutrality that must always character-ize the performance of the sovereign’s dutyto govern impartially.’’ Cleburne v. Cle-burne Living Center, Inc., 473 U.S. 432,452, 105 S.Ct. 3249, 87 L.Ed.2d 313 (1985)(STEVENS, J., concurring). A legislaturecontrolled by one party could not, for in-stance, impose special taxes on membersof the minority party, or use tax revenuesto pay the majority party’s campaign ex-penses. The rational basis for governmentdecisions must satisfy a standard of legiti-macy and S 338neutrality; an acceptable ra-tional basis can be neither purely personalnor purely partisan. See id., at 452–453,105 S.Ct. 3249.

The Constitution does not, of course,require proportional representation of ra-cial, ethnic, or political groups. In that Iagree with the plurality. Ante, at 1782.We have held, however, that proportionalrepresentation of political groups is a per-missible objective, Gaffney, 412 U.S., at754, 93 S.Ct. 2321, and some of us haveexpressed the opinion that a majority’sdecision to enhance the representation of aracial minority is equally permissible, par-ticularly when the decision is designed tocomply with the Voting Rights Act of1965.31 Thus, the view that the pluralityimplicitly embraces today—that a gerry-mander contrived for the sole purpose ofdisadvantaging a political minority is lessobjectionable than one seeking to benefit aracial minority—is doubly flawed. It dis-regards the obvious distinction between aninvidious and a benign purpose, and itmistakenly assumes that race cannot pro-vide a legitimate basis for making politicaljudgments.32

30. Bandemer, 478 U.S. 109, 106 S.Ct. 2797,92 L.Ed.2d 85 (plurality opinion); Gaffney v.Cummings, 412 U.S. 735, 754, 93 S.Ct. 2321,37 L.Ed.2d 298 (1973); Whitcomb v. Chavis,403 U.S. 124, 143, 91 S.Ct. 1858, 29 L.Ed.2d363 (1971); Burns v. Richardson, 384 U.S. 73,88, 86 S.Ct. 1286, 16 L.Ed.2d 376 (1966);Fortson v. Dorsey, 379 U.S. 433, 439, 85 S.Ct.498, 13 L.Ed.2d 401 (1965). Consistent withthese statements, the District Court in a re-cent case correctly described political gerry-mandering as ‘‘a purely partisan exercise’’and ‘‘an abuse of power that, at its core,evinces a fundamental distrust of voters, serv-ing the self-interest of the political parties atthe expense of the public good.’’ App. toJuris. Statement in Balderas v. Texas, O.T.2001, No. 01–1196, p. 10.

31. See Shaw II, 517 U.S., at 918, 116 S.Ct.1894 (STEVENS, J., dissenting); Bush v.Vera, 517 U.S., at 1033–1034, 116 S.Ct. 1941(STEVENS, J., dissenting); Miller, 515 U.S.,at 947–948, 115 S.Ct. 2475 (GINSBURG, J.,dissenting).

32. Because race so seldom provides a ration-al basis for a governmental decision, racialclassifications almost always fail to survive‘‘rational basis’’ scrutiny. But ‘‘[n]ot everydecision influenced by race is equally objec-tionable.’’ Grutter v. Bollinger, 539 U.S. 306,327, 123 S.Ct. 2325, 156 L.Ed.2d 304 (2003).When race is used as the basis for makingpredictive political judgments, it may be asreliable (or unreliable) as other group charac-teristics, such as political affiliation, econom-ic status, or national origin. The fact thatrace is an immutable characteristic does notmean that there is anything immutable orcertain about the political behavior of themembers of any racial class. See Mobile v.Bolden, 446 U.S. 55, 88, 100 S.Ct. 1490, 64L.Ed.2d 47 (1980) (STEVENS, J., concurringin judgment). Registered Republicans of allraces sometimes vote for Democratic candi-dates, and vice versa.

The plurality asserts that a person’s politics,unlike her race, is not readily ‘‘discernible.’’Ante, at 1782. But that assertion is belied bythe evidence that the architects of political


S 339In sum, in evaluating a challenge to aspecific district, I would apply the stan-dard set forth in the Shaw cases and askwhether the legislature allowed partisanconsiderations to dominate and control thelines drawn, forsaking all neutral princi-ples.33 Under my analysis, if no neutralcriterion can be identified to justify thelines drawn, and if the only possible expla-nation for a district’s bizarre shape is anaked desire to increase partisan strength,then no rational basis exists to save thedistrict from an equal protection challenge.Such a narrow test would cover only a fewmeritorious claims, but it would precludeextreme abuses, such as those disclosed bythe record in Badham v. Eu, 694 F.Supp.664 (N.D.Cal.1988), summarily aff’d, 488U.S. 1024, 109 S.Ct. 829, 102 L.Ed.2d 962(1989),34 and it would perhaps shorten thetime period in which the pernicious effectsof such a gerrymander are felt. This testwould mitigate the current trend underwhich partisan considerations are becom-ing the be-all and end-all in apportioningrepresentatives.

IVPlaintiff-appellant Furey plainly has

stated a claim that District 6 constitutesan unconstitutional partisan gerrymander.According to the complaint, Pennsylva-nia’s 2002 redistricting plan splits ‘‘Mont-gomery County alone TTT into sixS 340different congressional districts.’’ 35

The new District 6 ‘‘looms like a dragondescending on Philadelphia from the west,splitting up towns and communitiesthroughout Montgomery and Berks Coun-ties.’’ 36 Furey alleges that the districtingplan was created ‘‘solely to effectuate theinterests’’ of Republicans,37 and that theGeneral Assembly relied ‘‘exclusively on aprinciple of maximum partisan advantage’’when drawing the plan,38 ‘‘to the exclusionof all other criteria.’’ 39 The 2002 plan ‘‘isso irregular on its face that it rationallycan be viewed only as an effort TTT toadvance the interests of one political par-ty, without regard for traditional redis-tricting principles and without any legiti-mate or compelling justification.’’ 40 ‘‘Theproblem,’’ Furey claims, is that the legis-lature ‘‘subordinated—indeed ignored—alltraditional redistricting principles and all

gerrymanders seem to have no difficulty indiscerning the voters’ political affiliation. Af-ter all, eligibility to vote in primary electionsoften requires the citizen to register her partyaffiliation, but it never requires her to registerher race.

33. The one-person, one-vote rule obviouslyconstitutes a neutral districting criterion, butour gerrymandering cases have never citedthat principle as one of the traditional criteria‘‘that may serve to defeat a claim that adistrict has been gerrymandered on raciallines.’’ Shaw I, 509 U.S., at 647, 113 S.Ct.2816. Thus, I would require that a district bejustified with reference to both the one-per-son, one-vote rule and some other neutralcriterion. See Bandemer, 478 U.S., at 162,168, 106 S.Ct. 2797 (Powell, J., concurring inpart and dissenting in part).

34. The California districting scheme at issuein Badham featured a large number of dis-

tricts with highly irregular shapes, all de-signed, the plaintiffs-appellants alleged, to di-lute Republican voting strength throughoutthe State. See Juris. Statement in Badhamv. Eu, O.T.1987, No. 87–1818, Exh. D, p. 77a.Three Members of this Court dissented fromthe summary affirmance in Badham andwould have noted probable jurisdiction. 488U.S. 1024, 109 S.Ct. 829, 102 L.Ed.2d 962(1989).


36. Id., at 136a.

37. Id., at 142a.

38. Id., at 143a.

39. Id., at 140a.

40. Id., at 143a.


541 U.S. 341

legitimate bases for governmental deci-sionmaking, in order to favor those withone political viewpoint over another.’’ 41

The plan ‘‘ignores all other traditional re-districting criteria,’’ she alleges, ‘‘thusdemonstrating that partisanship—andnothing else—was the rationale behindthe plan.’’ 42 Because this complaintstates a claim under a judicially managea-ble standard for adjudicating partisan ger-rymandering cases, I would reverse thejudgment of the District Court and re-mand for further proceedings consistentwith this opinion.

The plurality candidly acknowledgesthat legislatures can fashion standards toremedy political gerrymandering that areperfectly manageable and, indeed, that thelegislatures in Iowa and elsewhere havedone so. Ante, at 1776, n. 4. If S 341aviolation of the Constitution is found, acourt could impose a remedy patternedafter such a statute. Thus, the problem, inthe plurality’s view, is not that there is nojudicially manageable standard to fix anunconstitutional partisan gerrymander, but

rather that the Judiciary lacks the abilityto determine when a state legislature hasviolated its duty to govern impartially.

Quite obviously, however, several stan-dards for identifying impermissible parti-san influence are available to judges whohave the will to enforce them. We couldhold that every district boundary musthave a neutral justification; we could ap-ply Justice Powell’s three-factor approachin Bandemer; we could apply the predomi-nant motivation standard fashioned by theCourt in its racial gerrymandering cases;or we could endorse either of the ap-proaches advocated today by JusticeSOUTER and Justice BREYER. What isclear is that it is not the unavailability ofjudicially manageable standards thatdrives today’s decision. It is, instead, afailure of judicial will to condemn even themost blatant violations of a state legisla-ture’s fundamental duty to govern impar-tially.

Accordingly, I respectfully dissent.

[Appendix to opinion of STEVENS, J.,follows this page.]

41. Ibid. 42. Id., at 135a.


S 342APPENDIX


541 U.S. 344

Justice SOUTER, with whom JusticeGINSBURG joins, dissenting.

S 343The Constitution guarantees both for-mal and substantial equality among voters.For 40 years, we have recognized thatlines dividing a State into voting districtsmust produce divisions with equal popula-tions: one person, one vote. Reynolds v.Sims, 377 U.S. 533, 568, 84 S.Ct. 1362, 12L.Ed.2d 506 (1964). Otherwise, a vote in aless populous district than others carriesmore clout.

Creating unequally populous districts isnot, however, the only way to skew politi-cal results by setting district lines. Thechoice to draw a district line one way, notanother, always carries some consequencefor politics, save in a mythical State withvoters of every political identity distribut-ed in an absolutely gray uniformity. Thespectrum of opportunity runs from crack-ing a group into impotent fractions, topacking its members into one district forthe sake of marginalizing them in another.However equal districts may be in popula-tion as a formal matter, the consequence ofa vote cast can be minimized or maxim-ized, Karcher v. Daggett, 462 U.S. 725, 734,n. 6, 103 S.Ct. 2653, 77 L.Ed.2d 133 (1983),and if unfairness is sufficiently demonstra-ble, the guarantee of equal protection con-demns it as a denial of substantial equality.Davis v. Bandemer, 478 U.S. 109, 129–134,106 S.Ct. 2797, 92 L.Ed.2d 85 (1986) (plu-rality opinion).

I

The notion of fairness assumed to bedenied in these cases has been describedas ‘‘each political group in a State [having]the same chance to elect representatives ofits choice as any other political group,’’ id.,at 124, 106 S.Ct. 2797, and as a ‘‘right to

‘fair and effective representation,’ ’’ id., at162, 106 S.Ct. 2797 (Powell, J., concurringin part and dissenting in part). Cf. Wellsv. Rockefeller, 394 U.S. 542, 551, 89 S.Ct.1234, 22 L.Ed.2d 535 (1969) (Harlan, J.,dissenting) (describing the need for ‘‘astructure which will in fact as well astheory be responsive to the sentiments ofthe community’’). It is undeniable thatpolitical sophisticates understand suchS 344fairness and how to go about destroyingit, see App. to Juris. Statement 134a,although it cannot possibly be describedwith the hard edge of one person, one vote.The difficulty has been to translate thesenotions of fairness into workable criteria,as distinct from mere opportunities forreviewing courts to make episodic judg-ments that things have gone too far, thesources of difficulty being in the facts thatsome intent to gain political advantage isinescapable whenever political bodies de-vise a district plan, and some effect resultsfrom the intent. Wells, supra, at 554–555,89 S.Ct. 1234 (White, J., dissenting) (‘‘Inreality, of course, districting is itself agerrymandering in the sense that it repre-sents a complex blend of political, econom-ic, regional, and historical considerations’’).Thus, the issue is one of how much is toomuch, and we can be no more exact instating a verbal test for too much partisan-ship than we can be in defining too muchrace consciousness when some is inevitableand legitimate. See Bush v. Vera, 517U.S. 952, 1057–1062, 116 S.Ct. 1941, 135L.Ed.2d 248 (1996) (SOUTER, J., dissent-ing). Instead of coming up with a verbalformula for too much, then, the Court’s jobmust be to identify clues, as objective aswe can make them, indicating that partisancompetition has reached an extremity ofunfairness.

The plurality says, in effect, that courtshave been trying to devise practical crite-


ria for political gerrymandering for nearly20 years, without being any closer to some-thing workable than we were when Daviswas decided. Ante, at 1778.1 While this istrue enough, I do not accept it as soundcounsel of despair. For I take it that theprincipal reason we have not gone fromtheoretical justiciability to practical admin-istrability in political gerrymanderingcases is the Davis plurality’s specificationthat any criterion of forbidden gerryman-dering must require a showing that mem-bers of the plaintiff’s group had ‘‘essential-ly been shut out of the political process,’’478 U.S., at 139, 106 S.Ct. 2797. See, e.g.,Badham v. Eu, 694 S 345F.Supp. 664, 670–671 (N.D.Cal.1988) (three-judge court).That is, in order to avoid a threshold forrelief so low that almost any electoral de-feat (let alone failure to achieve propor-tionate results) would support a gerryman-dering claim, the Davis plurality requireda demonstration of such pervasive devalua-tion over such a period of time as to raisereal doubt that a case could ever be madeout. Davis suggested that plaintiffs mightneed to show even that their efforts todeliberate, register, and vote had been im-peded. 478 U.S., at 133, 106 S.Ct. 2797.This standard, which it is difficult to imag-ine a major party meeting, combined avery demanding burden with significantvagueness; and if appellants have not beenable to propose a practical test for a Davisviolation, the fault belongs less to themthan to our predecessors. As JudgeHigginbotham recently put it, ‘‘[i]t is nowpainfully clear that Justice Powell’s con-cern that [Davis] offered a ‘ ‘‘constitutionalgreen light’’ to would-be gerrymanderers’has been realized.’’ Session v. Perry, 298F.Supp.2d 451, 474 (E.D.Tex.2004) (per cu-riam) (footnote omitted) (quoting Davis,

supra, at 173, 106 S.Ct. 2797 (Powell, J.,concurring in part and dissenting in part)).

IISince this Court has created the prob-

lem no one else has been able to solve, it isup to us to make a fresh start. There area good many voices saying it is high timethat we did, for in the years since Davis,the increasing efficiency of partisan redis-tricting has damaged the democratic pro-cess to a degree that our predecessorsonly began to imagine. E.g., Issacharoff,Gerrymandering and Political Cartels, 116Harv. L.Rev. 593, 624 (2002) (The ‘‘patternof incumbent entrenchment has gottenworse as the computer technology formore exquisite gerrymandering has im-proved’’); Karlan, The Fire Next Time:Reapportionment After the 2000 Census,50 Stan. L.Rev. 731, 736 (1998) (‘‘Finer-grained census data, better predictivemethods, and more powerful computersallow for increasingly sophisticated equipo-pulous gerrymanSders’’);346 Pildes, Princi-pled Limitations on Racial and PartisanRedistricting, 106 Yale L.J. 2505, 2553–2554 (1997) (‘‘Recent cases now documentin microscopic detail the astonishing preci-sion with which redistricters can carve upindividual precincts and distribute thembetween districts with confidence concern-ing the racial and partisan consequences’’).See also Morrill, A Geographer’s Perspec-tive, in Political Gerrymandering and theCourts 213–214 (B. Grofman ed.1990) (not-ing that gerrymandering can produce‘‘high proportions of very safe seats’’);Brief for Bernard Grofman et al. as AmiciCuriae 5–8 (decline of competitive seats).Cf. Wells, 394 U.S., at 551, 89 S.Ct. 1234(Harlan, J., dissenting) (‘‘A computer maygrind out district lines which can totallyfrustrate the popular will on an over-whelming number of critical issues’’).

1. And the plurality says the dissenters laborstill in vain today, ante, at 1784; I join in

Justice BREYER’s response, post, at 1829.


541 U.S. 348

I would therefore preserve Davis’s hold-ing that political gerrymandering is a justi-ciable issue, but otherwise start anew. Iwould adopt a political gerrymanderingtest analogous to the summary judgmentstandard crafted in McDonnell DouglasCorp. v. Green, 411 U.S. 792, 93 S.Ct. 1817,36 L.Ed.2d 668 (1973), calling for a plain-tiff to satisfy elements of a prima faciecause of action, at which point the Statewould have the opportunity not only torebut the evidence supporting the plain-tiff’s case, but to offer an affirmative justi-fication for the districting choices, evenassuming the proof of the plaintiff’s allega-tions. My own judgment is that we wouldhave better luck at devising a workableprima facie case if we concentrated asmuch as possible on suspect characteristicsof individual districts instead of statewidepatterns. It is not that a statewide view ofdistricting is somehow less important; theusual point of gerrymandering, after all, isto control the greatest number of seatsoverall. But, as will be seen, we would beable to call more readily on some existinglaw when we defined what is suspect at thedistrict level, and for now I would conceiveof a statewide challenge as itself a functionof claims that individual districts are ille-gitimately drawn. Finally, in the sameinterest of threshold simplicity, I wouldstick to probSlems347 of single-member dis-tricts; if we could not devise a workablescheme for dealing with claims aboutthese, we would have to forget the compli-cations posed by multimember districts.

III

A

For a claim based on a specific single-member district, I would require the plain-tiff to make out a prima facie case withfive elements. First, the resident plaintiffwould identify a cohesive political group towhich he belonged, which would normallybe a major party, as in this case and inDavis. There is no reason in principle,however, to rule out a claimant from aminor political party (which might, if itshowed strength, become the target of vig-orous hostility from one or both majorparties in a State) or from a different butpolitically coherent group whose membersengaged in bloc voting, as a large laborunion might do. The point is that it mustmake sense to speak of a candidate of thegroup’s choice, easy to do in the case of alarge or small political party, though moredifficult when the organization is not de-fined by politics as such.2

Second, a plaintiff would need to showthat the district of his residence, see Unit-ed States v. Hays, 515 U.S. 737, 115 S.Ct.2431, 132 L.Ed.2d 635 (1995) (requiringresidence in a challenged district forstanding), S 348paid little or no heed to thosetraditional districting principles whose dis-regard can be shown straightforwardly:contiguity, compactness, respect for politi-cal subdivisions, and conformity with geo-graphic features like rivers and mountains.Because such considerations are alreadyrelevant to justifying small deviations from

2. The plurality says it would not be easy todefine such a group, because ‘‘a person’s poli-tics is rarely as readily discernible—and neveras permanently discernible—as a person’srace,’’ ante, at 1782. But anytime politicalgerrymandering has been shown to occur,evidence must at least imply that the defen-dants themselves sat down, identified the rele-vant groups, and set out to concentrate thevote of one and dilute that of the others. If aplaintiff has the evidence, a court can figure

out what was going on. In major-party casesI do not see any problem with permitting aplaintiff to allege that he is a registered Re-publican, for example, and that the state legis-lature set out through gerrymandering tominimize the number of Republicans elected.If references to registration will not serve, aplaintiff will need to show the criteria forpartisan affiliation employed by the defen-dants in the challenged districting process.


absolute population equality, Karcher, 462U.S., at 740, 103 S.Ct. 2653, and becausecompactness in particular is relevant todemonstrating possible majority-minoritydistricts under the Voting Rights Act of1965, Johnson v. De Grandy, 512 U.S. 997,1008, 114 S.Ct. 2647, 129 L.Ed.2d 775(1994), there is no doubt that a test relyingon these standards would fall within judi-cial competence.

Indeed, although compactness is at firstblush the least likely of these principles toyield precision, it can be measured quanti-tatively in terms of dispersion, perimeter,and population ratios, and the developmentof standards would thus be possible. Seegenerally Pildes & Niemi, ExpressiveHarms, ‘‘Bizarre Districts,’’ and VotingRights: Evaluating Election–District Ap-pearances After Shaw v. Reno, 92 Mich.L.Rev. 483 (1993); see also Bush v. Vera,517 U.S., at 1057, 116 S.Ct. 1941 (SOUT-ER, J., dissenting) (suggesting that suchmeasuring formulas might have been ap-plied to salvage Shaw v. Reno, 509 U.S.630, 113 S.Ct. 2816, 125 L.Ed.2d 511(1993)).3 It is not necessary now to sayexactly S 349how a district court would bal-ance a good showing on one of these indi-ces against a poor showing on another, forthat sort of detail is best worked out caseby case.

Third, the plaintiff would need to estab-lish specific correlations between the dis-

trict’s deviations from traditional district-ing principles and the distribution of thepopulation of his group. For example, oneof the districts to which appellants objectmost strongly in this case is District 6,which they say ‘‘looms like a dragon de-scending on Philadelphia from the west,splitting up towns and communitiesthroughout Montgomery and Berks Coun-ties.’’ App. to Juris. Statement 136a. Tomake their claim stick, they would need topoint to specific protuberances on the Dra-conian shape that reach out to includeDemocrats, or fissures in it that squirmaway from Republicans. They would needto show that when towns and communitieswere split, Democrats tended to fall on oneside and Republicans on the other. Al-though some counterexamples would nodoubt be present in any complex plan, theplaintiff’s showing as a whole would needto provide reasonable support for, if notcompel, an inference that the district tookthe shape it did because of the distributionof the plaintiff’s group. That would begin,but not complete, the plaintiff’s case thatthe defendant had chosen either to packthe group (drawn a district in order toinclude a uselessly high number of thegroup) or to crack it (drawn it so as toinclude fatally few), the ordinary methodsof vote dilution in single-member districtsystems. Ante, at 1781, n. 7.

Fourth, a plaintiff would need to presentthe court with a hypothetical district in-

3. Those measures, as defined by ProfessorsPildes and Niemi, include dispersion, the ratioof the area of the district to the area of thesmallest circle that circumscribes the district,92 Mich.L.Rev., at 554–555; perimeter, theratio of the area of the district to the area ofthe circle whose diameter equals the length ofthe area’s perimeter, id., at 555–556; andpopulation, the ratio of the district’s popula-tion to the population contained by the mini-mum convex figure that encloses the district(or ‘‘rubber-band’’ area), id., at 556–557, andn. 206. The population measure can also betaken using the district’s circumscribing cir-

cle in the denominator. Id., at 557. See alsoPolsby & Popper, The Third Criterion: Com-pactness as a Procedural Safeguard AgainstPartisan Gerrymandering, 9 Yale L. & Pol’yRev. 301, 339–351 (1991) (discussing quanti-tative measures of compactness, and favoringthe perimeter measure as superior for antig-errymandering purposes); Schwartzberg, Re-apportionment, Gerrymanders, and the No-tion of ‘‘Compactness,’’ 50 Minn. L.Rev. 443(1966) (discussing proposed legislation thatwould have applied a variant of the perimetermeasure).


541 U.S. 351

cluding his residence, one in which theproportion of the plaintiff’s group was low-er (in a packing claim) or higher (in acracking one) and which at the same timedeviated less from traditional districtingprinciples than the actual district. Cf.Thornburg v. Gingles, 478 U.S. 30, S 35050,106 S.Ct. 2752, 92 L.Ed.2d 25 (1986) (re-quiring a similar showing to demonstratethat a multimember district is ‘‘responsiblefor minority voters’ inability to elect [theirpreferred] candidates’’). This hypotheticaldistrict would allow the plaintiff to claimcredibly that the deviations from tradition-al districting principles were not only cor-related with, but also caused by, the pack-ing or cracking of his group. Drawing thehypothetical district would, of course, nec-essarily involve redrawing at least one con-tiguous district,4 and a plaintiff would haveto show that this could be done subject totraditional districting principles withoutpacking or cracking his group (or another)worse than in the district being challenged.

Fifth, and finally, the plaintiff wouldhave to show that the defendants actedintentionally to manipulate the shape ofthe district in order to pack or crack hisgroup. See Washington v. Davis, 426 U.S.229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976).In substantiating claims of political gerry-mandering under a plan devised by a sin-gle major party, proving intent should notbe hard, once the third and fourth (correla-tion and cause) elements are established,

politicians not being politically disinterest-ed or characteristically naive. Davis v.Bandemer, 478 U.S., at 128, 106 S.Ct. 2797(‘‘[W]e think it most likely that whenever alegislature redistricts, those responsiblefor the legislation will know the likely po-litical composition of the new districts’’). Iwould, however, treat any showing of in-tent in a major-party case as too equivocalto count unless the entire legislature werecontrolled by the governor’s party (or thedominant legislative party were veto-proof).5

S 351If the affected group were not a ma-jor party, proof of intent could, admittedly,be difficult. It would be possible that alegislature might not even have had theplaintiff’s group in mind, and a plaintiffwould naturally have a hard time showingrequisite intent behind a plan produced bya bipartisan commission.

B

A plaintiff who got this far would haveshown that his State intentionally acted todilute his vote, having ignored reasonablealternatives consistent with traditional dis-tricting principles. I would then shift theburden to the defendants to justify theirdecision by reference to objectives otherthan naked partisan advantage. Theymight show by rebuttal evidence that dis-tricting objectives could not be served bythe plaintiff’s hypothetical district betterthan by the district as drawn, or they

4. It would not necessarily involve redrawingother noncontiguous districts, and I wouldnot permit a plaintiff to ask for such a remedyunless he first made out a prima facie case asto multiple districts. See infra, at 1820.

5. Amici JoAnn Erfer et al. suggest that a polit-ical party strong enough to redistrict withoutthe other’s approval is analogous to a firmthat exercises monopolistic control over amarket, and that the ability to exercise suchunilateral control should therefore trigger‘‘heightened constitutional scrutiny.’’ Brief

18–19 (citing Terry v. Adams, 345 U.S. 461, 73S.Ct. 809, 97 L.Ed. 1152 (1953), the TexasJaybird primary case). See also Issacharoff,Gerrymandering and Political Cartels, 116Harv. L.Rev. 593 (2002); Issacharoff &Pildes, Politics as Markets: Partisan Lockupsof the Democratic Process, 50 Stan. L.Rev.643 (1998). The analogy to antitrust is anintriguing one that may prove fruitful, thoughI do not embrace it at this point out of cau-tion about a wholesale conceptual transferfrom economics to politics.


might affirmatively establish legitimate ob-jectives better served by the lines drawnthan by the plaintiff’s hypothetical.

The State might, for example, posit theneed to avoid racial vote dilution. Cf.Bush v. Vera, 517 U.S., at 990, 116 S.Ct.1941 (O’CONNOR, J., concurring) (compli-ance with § 2 of the Voting Rights Act of1965 is a compelling state interest). Itmight plead one person, one vote, a stan-dard compatible with gerrymandering butin some places perhaps unattainable with-out some lopsided proportions. The Statemight adopt the object of proportional rep-resentation among its political partiesthrough its districting process. Gaffney v.Cummings, 412 U.S. 735, 754, 93 S.Ct.2321, 37 L.Ed.2d 298 (1973); 6cf. Johnsonv. De Grandy, 512 U.S., at S 3521024, 114S.Ct. 2647 (totality of the circumstancesdid not support finding of vote dilutionwhere ‘‘minority groups constitute[d] effec-tive voting majorities in a number of state

Senate districts substantially proportionalto their share in the population’’).7

This is not, however, the time or placefor a comprehensive list of legitimate ob-jectives a State might present. The pointhere is simply that the Constitution shouldnot petrify traditional districting objectivesas exclusive, and it is enough to say thatthe State would be required to explainitself, to demonstrate that whatever rea-sons it gave were more than a mere pre-text for an old-fashioned gerrymander.

S 353C

As for a statewide claim, I would notattempt an ambitious definition withoutthe benefit of experience with individualdistrict claims, and for now I would limitconsideration of a statewide claim to onebuilt upon a number of district-specificones. Each successful district-specificchallenge would necessarily entail redraw-

6. Some commentators have criticized Gaffneyitself for failing to account for the harm ofbipartisan political gerrymandering to the po-litical process. E.g., Issacharoff, PoliticalCartels, supra, at 613 (‘‘Gaffney illustrates theproblem of the use of a discrimination modelunmoored to any positive account of the elec-toral process’’). Gaffney is settled law, andfor today’s purposes I would take as given itsapproval of bipartisan gerrymanders, withtheir associated goal of incumbent protection.The plurality may be correct, ante, at 1787,that the test I propose could catch more ob-jectionable gerrymanders if we rejected in-cumbent protection as an acceptable purposeof districting. But I am wary of lumping allmeasures aimed at incumbent protection to-gether at this point, and I think we wouldgain a better sense of what to do if we waitedupon the experience of the district courts inassessing particular efforts at incumbencyprotection offered by the States in respondingto prima facie cases.

7. It is worth a moment to address the plurali-ty’s charge that any judicial remedy for politi-cal gerrymandering necessarily assumes aright to proportional representation. Ante, at

1782 (‘‘Deny it as appellants may (and do),[their] standard rests upon the principle thatgroups (or at least political-action groups)have a right to proportional representation’’).I agree with this Court’s earlier statementsthat the Constitution guarantees no right toproportional representation. See Davis v.Bandemer, 478 U.S. 109, 130, 106 S.Ct. 2797,92 L.Ed.2d 85 (1986) (plurality opinion) (cit-ing Whitcomb v. Chavis, 403 U.S. 124, 91S.Ct. 1858, 29 L.Ed.2d 363 (1971), and Whitev. Regester, 412 U.S. 755, 93 S.Ct. 2332, 37L.Ed.2d 314 (1973)). It does not follow thatthe Constitution permits every state actionintended to achieve any extreme form of dis-proportionate representation. ‘‘Proportionalrepresentation’’ usually refers to a set of pro-cedural mechanisms used to guarantee, withmore or less precision, that a political party’sseats in the legislature will be proportionateto its share of the vote. See generally S.Issacharoff, P. Karlan, & R. Pildes, The Lawof Democracy 1089–1172 (rev.2d ed.2002)(discussing voting systems other than the sin-gle-member district). The Constitution re-quires a State to adopt neither those mecha-nisms nor their goal of giving a party seatsproportionate to its vote.


541 U.S. 355

ing at least one contiguous district, and themore the successful claims, the more sur-rounding districts to be redefined. At acertain point, the ripples would reach thestate boundary, and it would no longermake any sense for a district court toconsider the problems piecemeal.

D

The plurality says that my proposedstandard would not solve the essentialproblem of unworkability. It says that‘‘[i]t does not solve [the] problem [of deter-mining when gerrymandering has gone toofar] to break down the original unanswera-ble question TTT into four more discretebut equally unanswerable questions.’’Ante, at 1787. It is common sense, howev-er, to break down a large and intractableissue into discrete fragments as a way toget a handle on the larger one, and theelements I propose are not only tractablein theory, but the very subjects thatjudges already deal with in practice. Theplurality asks, for example, ‘‘[w]hat TTT alower court [is] to do when, as will often bethe case, the district adheres to some tra-ditional criteria but not others?’’ Ibid.This question already arises in cases under§ 2 of the Voting Rights Act of 1965, andthe district courts have not had the samesort of difficulty answering it as they havein applying the Davis v. Bandemer plurali-ty. See, e.g., Johnson v. Hamrick, 155F.Supp.2d 1355, 1362–1363 (N.D.Ga.2001)(noncontiguity of a plaintiff’s Gingles dis-tricts was not fatal to a § 2 claim against amunicipal districting scheme because ‘‘thecity’s boundaries are rough and asymme-trical TTT [and] the non-contiguousporStions354 [of the proposed districts] areseparated by unincorporated areas and arerelatively near the districts to which theyare joined’’). The enquiries I am propos-ing are not, to be sure, as hard edged as Iwish they could be, but neither do they

have a degree of subjectivity inconsistentwith the judicial function.

The plurality also says that my standardis destined to fail because I have not givena precise enough account of the extremeunfairness I would prevent. Ante, at1787–1788. But this objection is more thereliable expression of the plurality’s owndiscouragement than the description of anAchilles heel in my suggestion. The harmfrom partisan gerrymandering is (as Ihave said, supra, at 1815, 1818, 1819) aspecies of vote dilution: the point of thegerrymander is to capture seats by manip-ulating district lines to diminish the weightof the other party’s votes in elections. Todevise a judicial remedy for that harm,however, it is not necessary to adopt a full-blown theory of fairness, furnishing a pre-cise measure of harm caused by diver-gence from the ideal in each case. It issufficient instead to agree that gerryman-dering is, indeed, unfair, as the pluralitydoes not dispute; to observe the tradition-al methods of the gerrymanderer, whichthe plurality summarizes, ante, at 1774–1776; and to adopt a test aimed at detect-ing and preventing the use of those meth-ods, which, I think, mine is. If thosemethods are unnecessary to effective ger-rymandering, as the plurality implies, ante,at 1788, it is hard to explain why they havebeen so popular down through the ages ofAmerican politics. My test would nodoubt leave substantial room for a party inpower to seek advantage through its con-trol of the districting process; the onlyway to prevent all opportunism would beto remove districting wholly from legisla-tive control, which I am not prepared tosay the Constitution requires. But thatdoes not make it impossible for courts toidentify at least the worst cases of gerry-mandering, and to provide a remedy. Themost the plurality can show is that myapproach would S 355not catch them all. Cf.Scalia, The Rule of Law as a Law of Rules,


56 U. Chi. L.Rev. 1175, 1178 (1989) (‘‘Toachieve what is, from the standpoint of thesubstantive policies involved, the ‘perfect’answer is nice—but it is just one of anumber of competing values’’).

IV

In drafting the complaint for this case,appellants’ counsel naturally proceeded onthe assumption that they had to satisfy theDavis v. Bandemer plurality, or some revi-sion in light of Shaw, but not the primafacie case I have in mind. Richard andNorma Jean Vieth make only statewideclaims, for which the single district claimbrought by Susan Furey provides insuffi-cient grounding. As for Furey’s ownclaim, her allegations fall short, for exam-ple, on the feasibility of an alternativedistrict superior to her own, as I wouldrequire. But she might well be able toallege what I would require, if given leaveto amend. I would grant her that leave,and therefore would vacate the judgmentof the District Court and remand for fur-ther proceedings. From the Court’s judg-ment denying her that opportunity, I re-spectfully dissent.

Justice BREYER, dissenting.

The use of purely political considerationsin drawing district boundaries is not a‘‘necessary evil’’ that, for lack of judiciallymanageable standards, the Constitution in-evitably must tolerate. Rather, pure poli-tics often helps to secure constitutionallyimportant democratic objectives. Butsometimes it does not. Sometimes purelypolitical ‘‘gerrymandering’’ will fail to ad-vance any plausible democratic objectivewhile simultaneously threatening seriousdemocratic harm. And sometimes whenthat is so, courts can identify an equalprotection violation and provide a remedy.Because the plaintiffs could claim (but

have not yet proved) that such circum-stances exist here, I would reverse theDistrict Court’s dismissal of their com-plaint.

S 356The plurality focuses directly on themost difficult issue before us. It says,‘‘[n]o test—yea, not even a five-part test—can possibly be successful unless oneknows what he is testing for.’’ Ante, at1787 (emphasis in original). That is true.Thus, I shall describe a set of circum-stances in which the use of purely politicaldistricting criteria could conflict with con-stitutionally mandated democratic require-ments—circumstances that the courtsshould ‘‘test for.’’ I shall then explain whyI believe it possible to find applicable judi-cially manageable standards. And I shallillustrate those standards.

I

I start with a fundamental principle.‘‘We the People,’’ who ‘‘ordain[ed] and es-tablish[ed]’’ the American Constitution,sought to create and to protect a workableform of government that is in its ‘‘ ‘princi-ples, structure, and whole mass,’ ’’ basical-ly democratic. G. Wood, The Creation ofthe American Republic, 1776–1787, p. 595(1969) (quoting W. Murray, PoliticalSketches, Inscribed to His ExcellencyJohn Adams 5 (1787)). See also, e.g., A.Meiklejohn, Free Speech and Its Relationto Self–Government 14–15 (1948). In amodern Nation of close to 300 million peo-ple, the workable democracy that the Con-stitution foresees must mean more than aguaranteed opportunity to elect legislatorsrepresenting equally populous electoraldistricts. Reynolds v. Sims, 377 U.S. 533,568, 84 S.Ct. 1362, 12 L.Ed.2d 506 (1964);Kirkpatrick v. Preisler, 394 U.S. 526, 530–531, 89 S.Ct. 1225, 22 L.Ed.2d 519 (1969);Karcher v. Daggett, 462 U.S. 725, 730, 103S.Ct. 2653, 77 L.Ed.2d 133 (1983). Theremust also be a method for transforming


541 U.S. 358

the will of the majority into effective gov-ernment.

This Court has explained that politicalparties play a necessary role in that trans-formation. At a minimum, they help vot-ers assign responsibility for current cir-cumstances, thereby enabling those voters,through their votes for individual candi-dates, to express satisfaction or dissatisfac-tion with the political status quo. Thosevoters can either vote to support that sta-tus quo or vote to ‘‘throw the rascals out.’’S 357See generally McConnell v. FederalElection Comm’n, 540 U.S. 93, 188, 124S.Ct. 619, 686, 157 L.Ed.2d 491 (2003);California Democratic Party v. Jones, 530U.S. 567, 574, 120 S.Ct. 2402, 147 L.Ed.2d502 (2000); Colorado Republican FederalCampaign Comm. v. Federal ElectionComm’n, 518 U.S. 604, 615–616, 116 S.Ct.2309, 135 L.Ed.2d 795 (1996). A party-based political system that satisfies thisminimal condition encourages democraticresponsibility. It facilitates the transfor-mation of the voters’ will into a govern-ment that reflects that will.

Why do I refer to these elementaryconstitutional principles? Because I be-lieve they can help courts identify at leastone abuse at issue in this case. To under-stand how that is so, one should begin byasking why single-member electoral dis-tricts are the norm, why the Constitutiondoes not insist that the membership oflegislatures better reflect different politi-cal views held by different groups of vot-ers. History, of course, is part of theanswer, but it does not tell the entirestory. The answer also lies in the factthat a single-member-district system helpsto ensure certain democratic objectivesbetter than many ‘‘more representative’’(i.e., proportional) electoral systems. Ofcourse, single-member districts mean thatonly parties with candidates who finish‘‘first past the post’’ will elect legislators.

That fact means in turn that a party witha bare majority of votes or even a pluralityof votes will often obtain a large legislativemajority, perhaps freezing out smallerparties. But single-member districtsthereby diminish the need for coalitiongovernments. And that fact makes it easi-er for voters to identify which party is re-sponsible for government decisionmaking(and which rascals to throw out), while si-multaneously providing greater legislativestability. Cf. C. Mershon, The Costs ofCoalition: Coalition Theories and ItalianGovernments, 90 Am. Pol. Sci. Rev. 534(1996) (noting that from 1946 to 1992, un-der proportional systems ‘‘almost no [Ital-ian] government stayed in office more thana few years, and many governments col-lapsed after only a few months’’); Her-mens, Representation and ProportionalRepresentation, S 358in Choosing an Elector-al System: Issues and Alternatives 15, 24(A. Lijphart & B. Grofman eds.1984) (de-scribing the ‘‘political paralysis which hadbecome the hallmark of the Fourth Repub-lic’’ under proportional representation).See also Duverger, Which is the BestElectoral System? in Choosing an Elector-al System, supra, at 31, 32 (arguing thatproportional systems ‘‘preven[t] the citi-zens from expressing a clear choice for agovernmental team,’’ and that nonpropor-tional systems allow voters to ‘‘choose gov-ernments with the capacity to make deci-sions’’). This is not to say that single-member districts are preferable; it is sim-ply to say that single-member-district sys-tems and more-directly-representationalsystems reflect different conclusions aboutthe proper balance of different elements ofa workable democratic government.

If single-member districts are the norm,however, then political considerations willlikely play an important, and proper, rolein the drawing of district boundaries. Inpart, that is because politicians, unlikenonpartisan observers, normally under-


stand how ‘‘the location and shape of dis-tricts’’ determine ‘‘the political complexionof the area.’’ Gaffney v. Cummings, 412U.S. 735, 753, 93 S.Ct. 2321, 37 L.Ed.2d298 (1973). It is precisely because politi-cians are best able to predict the effects ofboundary changes that the districts theydesign usually make some political sense.See, e.g., Persily, In Defense of FoxesGuarding Henhouses: The Case for Judi-cial Acquiescence to Incumbent–ProtectingGerrymanders, 116 Harv. L.Rev. 649, 678,and nn. 94–95 (2002) (recounting the au-thor’s experience as a neutral court-ap-pointed boundary drawer, in which theplan he helped draw moved an uninhabitedswamp from one district to another, there-by inadvertently disrupting environmentalprojects that were important to the politi-cian representing the swamp’s former dis-trict).

More important for present purposes,the role of political considerations reflectsa surprising mathematical fact. Given afairly large state population with a fairlylarge conSgressional359 delegation, districtsassigned so as to be perfectly random inrespect to politics would translate a smallshift in political sentiment, say a shift from51% Republican to 49% Republican, into aseismic shift in the makeup of the legisla-tive delegation, say from 100% Republicanto 100% Democrat. See M. Altman, Mod-eling the Effect of Mandatory DistrictCompactness on Partisan Gerrymanders,17 Pol. Geography 989, 1002 (1998) (sug-gesting that, where the state population islarge enough, even randomly selected com-pact districts will generally elect no politi-cians from the party that wins fewer votesstatewide). Any such exaggeration of tinyelectoral changes—virtually wiping outlegislative representation of the minorityparty—would itself seem highly undemoc-ratic.

Given the resulting need for single-member districts with nonrandom bound-aries, it is not surprising that ‘‘traditional’’districting principles have rarely, if ever,been politically neutral. Rather, because,in recent political memory, Democratshave often been concentrated in citieswhile Republicans have often been concen-trated in suburbs and sometimes rural ar-eas, geographically drawn boundaries havetended to ‘‘pac[k]’’ the former. See ante,at 1783 (plurality opinion) (citing Davis v.Bandemer, 478 U.S. 109, 159, 106 S.Ct.2797, 92 L.Ed.2d 85 (1986) (O’CONNOR,J., concurring in judgment)); Lowenstein& Steinberg, The Quest for LegislativeDistricting in the Public Interest: Elusiveor Illusory? 33 UCLA L.Rev. 1, 9 (1985)(explaining that the ‘‘ ‘formal’ criteria TTT

do not live up to their advance billing as‘fair’ or ‘neutral’ ’’). Neighborhood orcommunity-based boundaries, seeking togroup Irish, Jewish, or African–Americanvoters, often did the same. All this is wellknown to politicians, who use their knowl-edge about the effects of the ‘‘neutral’’criteria to partisan advantage when draw-ing electoral maps. And were it not so,the iron laws of mathematics would haveworked their extraordinary volatility-en-hancing will.

S 360This is to say that traditional or his-torically based boundaries are not, andshould not be, ‘‘politics free.’’ Rather,those boundaries represent a series ofcompromises of principle—among the vir-tues of, for example, close representationof voter views, ease of identifying ‘‘govern-ment’’ and ‘‘opposition’’ parties, and stabili-ty in government. They also represent anuneasy truce, sanctioned by tradition,among different parties seeking politicaladvantage.

As I have said, reference back to theseunderlying considerations helps to explainwhy the legislature’s use of political bound-


541 U.S. 362

ary-drawing considerations ordinarily doesnot violate the Constitution’s Equal Pro-tection Clause. The reason lies not simplyin the difficulty of identifying abuse orfinding an appropriate judicial remedy.The reason is more fundamental: Ordi-narily, there simply is no abuse. The useof purely political boundary-drawing fac-tors, even where harmful to the membersof one party, will often nonetheless findjustification in other desirable democraticends, such as maintaining relatively stablelegislatures in which a minority party re-tains significant representation.

II

At the same time, these considerationscan help identify at least one circumstancewhere use of purely political boundary-drawing factors can amount to a serious,and remediable, abuse, namely, the unjust-ified use of political factors to entrench aminority in power. By entrenchment Imean a situation in which a party thatenjoys only minority support among thepopulace has nonetheless contrived to take,and hold, legislative power. By unjusti-fied entrenchment I mean that the minori-ty’s hold on power is purely the result ofpartisan manipulation and not other fac-tors. These ‘‘other’’ factors that could leadto ‘‘justified’’ (albeit temporary) minorityentrenchment include sheer happenstance,the existence of more than two major par-ties, the unique constitutional require-ments of certain representational bodSies361

such as the Senate, or reliance on tradi-tional (geographic, communities of interest,etc.) districting criteria.

The democratic harm of unjustified en-trenchment is obvious. As this Court haswritten in respect to popularly based elec-toral districts:

‘‘Logically, in a society ostensiblygrounded on representative government,it would seem reasonable that a majority

of the people of a State could elect amajority of that State’s legislators. Toconclude differently, and to sanction mi-nority control of state legislative bodies,would appear to deny majority rights ina way that far surpasses any possibledenial of minority rights that might oth-erwise be thought to result. Since legis-latures are responsible for enacting lawsby which all citizens are to be governed,they should be bodies which are collec-tively responsive to the popular will.’’Reynolds, 377 U.S., at 565, 84 S.Ct.1362.

Where unjustified entrenchment takesplace, voters find it far more difficult toremove those responsible for a govern-ment they do not want; and these demo-cratic values are dishonored.

The need for legislative stability cannotjustify entrenchment, for stability is com-patible with a system in which the loss ofmajority support implies a loss of power.The need to secure minority representa-tion in the legislature cannot justify en-trenchment, for minority party representa-tion is also compatible with a system inwhich the loss of minority support impliesa loss of representation. Constitutionallyspecified principles of representation, suchas that of two Senators per State, cannotjustify entrenchment where the House ofRepresentatives or similar state legislativebody is at issue. Unless some other justi-fication can be found in particular circum-stances, political gerrymandering that soentrenches a minority party in power vio-lates basic democratic norms and lackscountervailing justification. For thisS 362reason, whether political gerrymander-ing does, or does not, violate the Constitu-tion in other instances, gerrymanderingthat leads to entrenchment amounts to anabuse that violates the Constitution’sEqual Protection Clause.


IIICourts need not intervene often to pre-

vent the kind of abuse I have described,because those harmed constitute a politicalmajority, and a majority normally canwork its political will. Where a State hasimproperly gerrymandered legislative orcongressional districts to the majority’sdisadvantage, the majority should be ableto elect officials in statewide races—partic-ularly the Governor—who may help toundo the harm that districting has causedthe majority’s party, in the next round ofdistricting if not sooner. And where aState has improperly gerrymandered con-gressional districts, Congress retains thepower to revise the State’s districting de-terminations. See U.S. Const., Art. I, § 4;ante, at 1775–1776 (plurality opinion) (dis-cussing the history of Congress’ ‘‘power tocheck partisan manipulation of the electionprocess by the States’’).

Moreover, voters in some States, per-haps tiring of the political boundary-draw-ing rivalry, have found a procedural solu-tion, confiding the task to a commissionthat is limited in the extent to which itmay base districts on partisan concerns.According to the National Conference ofState Legislatures, 12 States currentlygive ‘‘first and final authority for [state]legislative redistricting to a group otherthan the legislature.’’ National Confer-ence of State Legislatures, RedistrictingCommissions and Alternatives to the Leg-islature Conducting Redistricting (2004),available at http://www.ncsl.org/programs/legman/Redistrict/Com & alter.htm (all In-ternet materials as visited Mar. 29, 2004,and available in Clerk of Court’s case file).A number of States use a commission forcongressional redistricting: Arizona, Ha-waii, Idaho, Montana, New Jersey, andWashington, with Indiana S 363using a com-mission if the legislature cannot pass aplan and Iowa requiring the district-draw-ing body not to consider political data.

Ibid.; Iowa General Assembly, LegislativeService Bureau, Legislative Guide to Re-districting (Dec. 2000), availableat http://www.legis.state.ia.us/Central/LSB/Guides/redist.htm. Indeed, wherestate governments have been unwilling orunable to act, ‘‘an informed, civically mili-tant electorate,’’ Baker v. Carr, 369 U.S.186, 270, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962)(Frankfurter, J., dissenting), has occasion-ally taken matters into its own hands,through ballot initiatives or referendums.Arizona voters, for example, passed Propo-sition 106, which amended the State’s Con-stitution and created an independent redis-tricting commission to draw legislative andcongressional districts. Ariz. Const., Art.4, pt. 2, § 1 (West 2001). Such reformsborrow from the systems used by othercountries utilizing single-member districts.See, e.g., Administration and Cost of Elec-tions Project, Boundary Delimitation(hereinafter ACE Project), Representationin the Canadian Parliament, available athttp://www.aceproject.org/main/english/bd/bdy ca.htm (describing Canada’s indepen-dent boundary commissions, which draftmaps based on equality of population, com-munities of interest, and geographic fac-tors); ACE Project, The United KingdomRedistribution Process, available at http://www.aceproject.org/main/english/bd/bdygb.htm (describing the United Kingdom’sindependent boundary commissions, whichmake recommendations to Parliament af-ter consultation with the public); G. Gud-gin & P. Taylor, Seats, Votes, and theSpatial Organisation of Elections 8 (1979)(noting that the United Kingdom’s bound-ary commissions are ‘‘explicitly neutral ina party political sense’’).

But we cannot always count on a severe-ly gerrymandered legislature itself to findand implement a remedy. See Bandemer,478 U.S., at 126, 106 S.Ct. 2797. The


541 U.S. 365

party that controls the process has noincentive to change it. And the politicaladvantages of a gerrymander may becomeever greater in the future. S 364The avail-ability of enhanced computer technologyallows the parties to redraw boundaries inways that target individual neighborhoodsand homes, carving out safe but slim victo-ry margins in the maximum number ofdistricts, with little risk of cutting theirmargins too thin. See generally Handley,A Guide to 2000 Redistricting Tools andTechnology, in The Real Y2K Problem:Census 2000 Data and Redistricting Tech-nology (N. Persily ed.2000); Karlan, TheFire Next Time: Reapportionment Afterthe 2000 Census, 50 Stan. L.Rev. 731, 736(1998); ante, at 1816 (SOUTER, J., dis-senting). By redrawing districts every 2years, rather than every 10 years, a partymight preserve its political advantages not-withstanding population shifts in the State.The combination of increasingly precisemap-drawing technology and increasinglyfrequent map drawing means that a partymay be able to bring about a gerrymanderthat is not only precise, but virtually im-possible to dislodge. Thus, court actionmay prove necessary.

When it is necessary, a court shouldprove capable of finding an appropriateremedy. Courts have developed district-ing remedies in other cases. See, e.g.,Branch v. Smith, 538 U.S. 254, 123 S.Ct.1429, 155 L.Ed.2d 407 (2003) (affirmingthe District Court’s injunction of use ofstate court’s redistricting plan and orderthat its own plan be used until a state plancould be precleared under the VotingRights Act of 1965); Karcher, 462 U.S.725, 103 S.Ct. 2653, 77 L.Ed.2d 133 (up-holding the District Court’s holding that acongressional reapportionment plan wasunconstitutional); Reynolds, 377 U.S., at586–587, 84 S.Ct. 1362 (upholding the Dis-trict Court’s actions in ordering into effecta reapportionment of both houses of the

state legislature). See also Issacharoff,Judging Politics: The Elusive Quest forJudicial Review of Political Fairness, 71Texas L.Rev. 1643, 1688–1690, and nn.227–233 (1993) (reporting that, in the wakeof the 1980 census, there were 13 court-ordered plans for congressional redistrict-ing, 5 plans that the courts rejected andreturned to state legislatures for redraft-ing, 7 court-ordered state senate plans, 8state senate S 365plans rejected and sentback to the state legislatures, 6 court-ordered state house plans, and 9 statehouse plans sent back for further legisla-tive action—all of which meant that, leav-ing aside the preclearance provisions of§ 5 of the Voting Rights Act of 1965,about one-third of all redistricting wasdone either directly by the federal courtsor under courts’ injunctive authority (cit-ing cases)). Moreover, if the dangers ofinadvertent political favoritism prove toogreat, a procedural solution, such as theuse of a politically balanced boundary-drawing commission, may prove possible.

The bottom line is that courts should beable to identify the presence of one impor-tant gerrymandering evil, the unjustifiedentrenching in power of a political partythat the voters have rejected. Theyshould be able to separate the unjustifiedabuse of partisan boundary-drawing con-siderations to achieve that end from theirmore ordinary and justified use. And theyshould be able to design a remedy forextreme cases.

IV

I do not claim that the problem of iden-tification and separation is easily solved,even in extreme instances. But courts canidentify a number of strong indicia ofabuse. The presence of actual entrench-ment, while not always unjustified (beingperhaps a chance occurrence), is such asign, particularly when accompanied by


the use of partisan boundary-drawing cri-teria in the way that Justice STEVENSdescribes, i.e., a use that both departsfrom traditional criteria and cannot be ex-plained other than by efforts to achievepartisan advantage. Below, I set forthseveral sets of circumstances that lay outthe indicia of abuse I have in mind. Thescenarios fall along a continuum: Themore permanently entrenched the minori-ty’s hold on power becomes, the less evi-dence courts will need that the minorityengaged in gerrymandering to achieve thedesired result.

Consider, for example, the following setsof circumstances. First, suppose that thelegislature has proceeded to redrawS 366boundaries in what seem to be ordinaryways, but the entrenchment harm has be-come obvious. E.g., (a) the legislature hasnot redrawn district boundaries more thanonce within the traditional 10–year period;and (b) no radical departure from tradi-tional districting criteria is alleged; but (c)a majority party (as measured by the votesactually cast for all candidates who identifythemselves as members of that party inthe relevant set of elections; i.e., in con-gressional elections if a congressional mapis being challenged) has twice failed toobtain a majority of the relevant legislativeseats in elections; and (d) the failure can-not be explained by the existence of multi-ple parties or in other neutral ways. Inmy view, these circumstances would besufficient to support a claim of unconstitu-tional entrenchment.

Second, suppose that plaintiffs couldpoint to more serious departures from re-districting norms. E.g., (a) the legislaturehas not redrawn district boundaries morethan once within the traditional 10–yearperiod; but (b) the boundary-drawing cri-teria depart radically from previous or tra-ditional criteria; (c) the departure cannotbe justified or explained other than by

reference to an effort to obtain partisanpolitical advantage; and (d) a majorityparty (as defined above) has once failed toobtain a majority of the relevant seats inelection using the challenged map (whichfact cannot be explained by the existenceof multiple parties or in other neutralways). These circumstances could alsoadd up to unconstitutional gerrymander-ing.

Third, suppose that the legislature clear-ly departs from ordinary districting norms,but the entrenchment harm, while serious-ly threatened, has not yet occurred. E.g.,(a) the legislature has redrawn districtboundaries more than once within the tra-ditional 10–year census-related period—ei-ther, as here, at the behest of a court thatstruck down an initial plan as unlawful, seeVieth v. Pennsylvania, 195 F.Supp.2d 672(M.D.Pa.2002) (per curiam) (finding thatPennsylvania’s first redistricting plan vio-lated the one-Sperson,367 one-vote mandate),or of its own accord; (b) the boundary-drawing criteria depart radically from pre-vious traditional boundary-drawing crite-ria; (c) strong, objective, unrefuted statis-tical evidence demonstrates that a partywith a minority of the popular vote withinthe State in all likelihood will obtain amajority of the seats in the relevant repre-sentative delegation; and (d) the jettison-ing of traditional districting criteria cannotbe justified or explained other than byreference to an effort to obtain partisanpolitical advantage. To my mind, suchcircumstances could also support a claim,because the presence of midcycle redis-tricting, for any reason, raises a fair infer-ence that partisan machinations played amajor role in the map-drawing process.Where such an inference is accompaniedby statistical evidence that entrenchmentwill be the likely result, a court may con-clude that the map crosses the constitu-tional line we are describing.


541 U.S. 368

The presence of these, or similar, cir-cumstances—where the risk of entrench-ment is demonstrated, where partisan con-siderations render the traditional district-drawing compromises irrelevant, where nojustification other than party advantagecan be found—seem to me extreme enoughto set off a constitutional alarm. The riskof harm to basic democratic principle isserious; identification is possible; andremedies can be found.

V

The plurality sets forth several criti-cisms of my approach. Some of thosecriticisms are overstated. Compare ante,at 1789 (‘‘[O]f course there always is aneutral explanation [of gerrymandering]—if only the time-honored criterion of incum-bent protection’’), with Brief for Appellants13 (pointing to examples of efforts to ger-rymander an incumbent of the oppositionparty out of office and elect a new memberof the controlling party); compare ante, at1789 (complaining of ‘‘the difficulties ofassessing partisan strength statewide’’),with supra, at 1828 (identifying the ‘‘ma-jority party’’ simply by S 368adding up ‘‘thevotes actually cast for all candidates whoidentify themselves as members of thatparty in the relevant set of elections’’).

Other criticisms involve differing judg-ments. Compare ante, at 1789 (complain-ing about the vagueness of unjustified po-litical machination, ‘‘whatever that means,’’and of unjustified entrenchment), with su-pra, at 1825 (detailed discussion of ‘‘justi-fied’’ and Reynolds v. Sims); compareante, at 1789 (finding costs of judicial in-tervention too high), with supra, at 1827(finding costs warranted to ensure majori-ty rule).

But the plurality makes one criticismthat warrants a more elaborate response.It observes ‘‘that the mere fact that thesefour dissenters come up with three differ-

ent standards—all of them different fromthe two proposed in Bandemer and theone proposed here by appellants—goes along way to establishing that there is noconstitutionally discernible standard.’’Ante, at 1784.

Does it? The dissenting opinions rec-ommend sets of standards that differ incertain respects. Members of a majoritymight well seek to reconcile such differ-ences. But dissenters might instead be-lieve that the more thorough, specificreasoning that accompanies separatestatements will stimulate further discus-sion. And that discussion could lead tochange in the law, where, as here, onemember of the majority, disagreeing withthe plurality as to justiciability, remainsin search of appropriate standards. Seeante, at 1796 (KENNEDY, J., concurringin judgment).

VI

In the case before us, there is a stronglikelihood that the plaintiffs’ complaintcould be amended readily to assert circum-stances consistent with those I have setforth as appropriate for judicial interven-tion. For that reason, I would authorizethe plaintiffs to proceed; and I dissentfrom the majority’s contrary determina-tion.

,

Documents

VIETH v. JUBELIRER 1769 - Brennan Center