Metts v. Almond

363 F.3d 8

Opinion

United States Court of Appeals For the First Circuit

No. 02-2204

HAROLD METTS; JEAN WIGGINS; BRYAN EVANS; STEPHANIE CRUZ;

URBAN LEAGUE; NAACP - PROVIDENCE;

BLACK AMERICAN CITIZENS POLITICAL ACTION COMMITTEE,

Plaintiffs, Appellants,

v.

WILLIAM J. MURPHY, Speaker of the House of Representatives;

ROGER N. BEGIN, in his official capacity as State Board of Elections Chairman; MATTHEW A. BROWN, Secretary of State;

JOSEPH A. MONTALBANO, Senate Majority Leader,

Defendants, Appellees,

DONALD L. CARCIERI, Governor;

CHARLES FOGARTY, Lt. Governor and Presiding Officer of the Senate,

Defendants.

APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF RHODE ISLAND

[Hon. Ernest C. Torres, Chief U.S. District Judge]

Before

Selya, Circuit Judge,

Stahl, Senior Circuit Judge, and Lynch, Circuit Judge.

Anita Hodgkiss, Lawyers' Committee for Civil Rights Under Law, with whom Sunil R. Kulkarni, Morrison & Foerster LLP, Kelli Reynolds, NAACP Legal Department, and Bruce G. Pollock were on brief, for appellants.

Joseph M. Fernandez and Goldenberg & Muri LLP on brief for Rhode Island Affiliate, American Civil Liberties Union, amicus curiae.

Marlene Twaddell on brief for Puerto Rican Political Action Committee and Direct Action for Rights and Equality, amici curiae.

John A. Tarantino, with whom Patricia K. Rocha, Victoria M.

Almeida, and Adler Pollock & Sheehan P.C. were on brief, for appellee Senate Majority Leader.

Richard B. Woolley and Thomas A. Palombo, Assistant Attorneys General, on brief for appellee Secretary of State.

Normand G. Benoit, Eugene G. Bernado, II and Partridge Snow & Hahn LLP on brief for appellee Speaker of the House of Representatives.

Raymond A. Marcaccio on brief for appellee Chairman of the State Board of Elections.

October 28, 2003

LYNCH, Circuit Judge. A group of African-American voters and related organizations brought a challenge under § 2 of the Voting Rights Act, 42 U.S.C. § 1973 (2000), to the Rhode Island state senate redistricting plan adopted in 2002. They allege that although African-Americans did not constitute a numerical majority in any state senate district before redistricting, they have historically had the ability to elect a representative of their choice with the help of crossover votes in one of the former districts. They claim that as a result of the redistricting plan, this opportunity has been adversely affected (indeed, eliminated) by the reduction of the African-American percentage in the relevant district. After the districts were redrawn, their candidate of choice, at that time an incumbent, lost his seat in the Democratic Party primary. Because of the makeup of the newly configured district, the victor in the primary was effectively assured of being the victor in the general election.1 The district court dismissed the claim under Fed. R. Civ.

P. 12(b)(6) because the African-American group could not form a numerical majority in any district and because that group would require crossover votes to elect a candidate of its choice. Under

1 While the loss of the incumbent, Charles D. Walton, in Senate District 9 is not part of the plaintiffs' complaint, we take judicial notice of this electoral outcome. The fact of his loss is undisputed and has been referred to by the parties. It is also an easy inference from the complaint that the African-American voters' candidate of choice would lose after and as a result of the redistricting process.

4

the standard for Rule 12(b)(6) dismissal, which permits dismissal of a complaint "only if it is clear that no relief could be granted under any set of facts that could be proved consistent with the allegations," Hishon v. King & Spalding, 467 U.S. 69, 73 (1984), we reverse the dismissal of the claim.2 I.

On May 2, 2002, a group of individual plaintiffs and advocacy organizations challenged the redistricting plan in the United States District Court for the District of Rhode Island under § 2 of the Voting Rights Act (VRA), 42 U.S.C. § 1973. The plaintiffs named as defendants the Governor, the Lieutenant Governor, the Speaker of the House of Representatives, the State Board of Elections Chairman, the Secretary of State, and the Senate Majority Leader.3 When reviewing the dismissal of a complaint under Fed.

R. Civ. P. 12(b)(6), "[w]e accept as true the well-pleaded factual allegations of the complaint, draw all reasonable inferences therefrom in the plaintiff's favor and determine whether the complaint, so read, sets forth facts sufficient to justify recovery on any cognizable theory." Martin v. Applied Cellular Tech., Inc.,

2 We express our appreciation to amici for their valuable assistance.

3 Many of the defendants originally sued in their official capacities no longer occupy their respective offices. The current incumbents have been substituted as defendants for their predecessors in office. See Fed. R. App. P. 43(c)(2).

5

284 F.3d 1, 6 (1st Cir. 2002). Thus, the following facts are derived from the plaintiffs' amended complaint.

On February 23, 2002, the Rhode Island General Assembly voted to pass a redistricting plan for the state senate. An alteration was necessary for two reasons. First, there was a need to adjust the senate districts to account for shifts in state population. Second, a recent state constitutional amendment reduced the number of senate districts from fifty to thirty-eight, necessitating an entirely new district map with larger districts.

The new senate district plan was highly controversial.

There were concerns from the beginning that the plan might make it more difficult for African-American voters to elect candidates of their choice. Various community groups and individuals testified before the legislature against the plan on the grounds that it would not give African-American voters "an equal opportunity to elect candidates of their choice" to the state senate, and that the plan unnecessarily abridged the voting rights of African-American voters in violation of the VRA. Nonetheless, the senate's judiciary committee "approved the plan[] without taking the time to evaluate the proposals and comments of those opposed to the plan." It is fair to infer, given that there was only one African-American senator, that the plan was approved over the objections of the African-American community and its representative. Governor Lincoln Almond refused to sign the legislation, explicitly

6

questioning its fairness to Rhode Island's minority populations. He did not veto it, however, and the plan became law without his signature on February 23, 2002.

The population of Rhode Island is four percent African-

American,4 over half of whom live in Providence. The state's African-American citizens continue to suffer from past official discrimination in housing, education, health care, and employment. By common measures of socio-economic status, educational attainment, and access to political resources, they continue to lag behind the rest of the state. Only one African-American state senator, the chosen candidate of the African-American community, has ever been elected in Rhode Island; that senator, Charles D. Walton, represented the old Senate District 9 until the redistricting. According to the census data from the year 2000, that district was 25.69% African-American and 41.08% Hispanic. Much of Providence's African-American population is now within the new Senate District 2. The population of this new district is 21.42% African-American and 46.74% Hispanic, and the voting age population is 21.43% African-American and 43.12% Hispanic.5

4 The complaint makes a distinction between Hispanic voters and "Non-Hispanic African-American" voters. We use "African- American" to describe the latter group, as distinguished from Hispanic African-Americans, whom the complaint counts as members of the Hispanic community.

5 The complaint does not specify the voting age population demographics of the old Senate District 9, nor does it specify the percentage of the population that is white in either the old Senate

7

Plaintiffs plead that the African-American voters in Rhode Island are themselves politically cohesive, and that they are not cohesive with Hispanic or white voters. Although no alternative plans were appended to the complaint, plaintiffs claim that it is possible to divide the state into thirty-eight districts such that one senate district would have a population that is at least twenty-six percent African-American, and in such a district it would be possible for "an African-American candidate preferred by African-American voters" to win election due to white and Hispanic crossover support. However, if a district is less than twenty-six percent African-American, "[t]he white and Hispanic communities vote sufficiently in a bloc usually to defeat the candidate of choice of African-American voters."

II.

Without filing a responsive pleading, the defendants quickly moved to dismiss on the basis of Fed. R. Civ. P. 12(b)(6) for failure to state a claim. They argued that the complaint fails to allege that it is possible to create a senate district in which African-Americans are a majority, and that such an allegation is required by Thornburg v. Gingles, 478 U.S. 30 (1986), and its progeny.

On September 9, 2002, the district court granted the defendants' motion. Metts v. Almond, 217 F. Supp. 2d 252 (D.R.I.

District 9 or the new Senate District 2.

8

2002). The district court analyzed the plaintiffs' complaint as both an "ability to influence" claim and an "ability to elect" claim. As to the former, it found that influence claims are not cognizable under § 2. Id. at 257. As to the latter, it held that Gingles requires that a minority group be able to constitute a majority without the help of crossover votes from other groups. Id. at 260. Finally, the district court also dismissed the complaint based upon a failure to demonstrate that the majority in Senate District 9 votes as a bloc, another requirement set out in Gingles. Id. at 260-61. The plaintiffs appeal the dismissal of their claim.

III.

A. Standard of Review We review de novo a district court's dismissal of a complaint for failure to state a claim under Rule 12(b)(6), Morales-Villalobos v. Garcia-Llorens, 316 F.3d 51, 52 (1st Cir. 2003), taking well-pleaded facts in the complaint as true and making all reasonable inferences in favor of the plaintiffs. Arruda v. Sears, Roebuck & Co., 310 F.3d 13, 18 (1st Cir. 2002). Rule 12(b)(6) permits dismissal of a complaint for "failure of the pleading to state a claim upon which relief can be granted." For the purposes of Rule 12(b)(6), "it is enough for a plaintiff to sketch a scenario which, if subsequently fleshed out by means of appropriate facts, could support an actionable claim." Garrett v.

9

Tandy Corp., 295 F.3d 94, 105 (1st Cir. 2002). We must reverse if the plaintiffs have included in their complaint well-pleaded facts which, taken as true, "justify recovery on any supportable legal theory." Cruz v. Melecio, 204 F.3d 14, 21 (1st Cir. 2000). B. Legal Background Section 2 of the VRA forbids any "voting qualification or prerequisite to voting or standard, practice, or procedure . . . which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color." 42 U.S.C. § 1973(a). First passed in 1965, it was amended in 1982 to specify that the test is one of adverse impact to minority communities and does not require a showing of discriminatory intent.6 Pub. L. No. 89-110, tit. I, § 2, 79 Stat. 437, 437 (1965), amended by Pub. L. No. 97-205, § 3, 96 Stat. 131, 134 (1982). A violation is established "if, based on the totality of circumstances, it is shown that . . . a class of citizens . . . [has] less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 U.S.C. § 1973(b).

The Supreme Court first construed the post-amendment VRA

6 The amendment was meant, in part, to overturn the Supreme Court's interpretation of the VRA in City of Mobile v. Bolden, 446 U.S. 55, 61 (1980). See P. McCrary, Bringing Equality to Power: How the Federal Courts Transformed the Electoral Structure of Southern Politics, 1960-1990, 5 U. Pa. J. Const. L. 665, 697-699 (2003).

10

in Thornburg v. Gingles, supra. Gingles was a challenge to a redistricting plan that included multimember districts, which are legislative districts from which more than one representative is elected at a time. 478 U.S. at 35. The Gingles court established three "preconditions" for a VRA challenge to multimember districts.

First, the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district. .

. . Second, the minority group must be able to show that it is politically cohesive. . . . Third, the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it . . . usually to defeat the minority's preferred candidate.

Id. at 50-51. The Supreme Court has, in subsequent cases, held that some form of these three preconditions should also apply to challenges to single-member legislative districts. Growe v. Emison, 507 U.S. 25, 40-41 (1993) (applying the second precondition); see also Voinovich v. Quilter, 507 U.S. 146, 157-58 (1993) (applying the third precondition). The Court has summarized the three in shorthand terms as "compactness/numerousness, minority cohesion or bloc voting, and majority bloc voting." Johnson v. De Grandy, 512 U.S. 997, 1011 (1994). As a matter of pleading, the complaint explicitly pleads the second and third of the Gingles preconditions as well as geographic compactness under the first precondition.

This is not a case, as in De Grandy, where minority voters hold a majority in some districts and the issue is whether § 2 of the VRA requires courts to maximize the number of districts

11

in which minority voters may elect their candidates of choice. Here, the only district in which African-American voters could elect a candidate of their choice (with help from crossover voting) was altered significantly; the result, plaintiffs say, is that African-American voters can no longer elect a candidate of their choice in any state senatorial district. Far from complaining that the legislature has failed to maximize their political power, these plaintiffs complain that their opportunity to elect a candidate of their choice has been minimized -- indeed, eliminated. C. First Precondition: Compactness & Numerousness 1. Ability to Influence and Crossover Voting The district court characterized plaintiffs' claim as alternately an "ability to influence" claim and an "ability to elect" claim. The Gingles Court, when fashioning the three preconditions to a redistricting challenge to a multimember district, expressly reserved the question of whether § 2 permitted claims by a minority group "alleging that the use of a multimember district impairs its ability to influence elections," and whether the three preconditions would apply unabated to such a claim. 478 U.S. at 46 n.12 (emphasis in original). The same question of the meaning of an ability to elect as opposed to an ability to influence arises in challenges to single member districts. See De Grandy, 512 U.S. at 1008-09; Voinovich, 507 U.S. at 154.

Since Gingles, there has been much confusion over the

12

definition of an influence claim under the VRA. Most often, influence districts have been defined as ones "in which a minority group has enough political heft to exert significant influence on the choice of candidate though not enough to determine that choice." Barnett v. City of Chicago, 141 F.3d 699, 703 (7th Cir. 1998) (reserving question of whether such a claim is cognizable); see Cousin v. Sundquist, 145 F.3d 818, 828-29 (6th Cir. 1998) (refusing to recognize such a claim under the VRA); McNeil v. Legislative Apportionment Comm'n, 828 A.2d 840, 852-53 (N.J. 2003) (recognizing influence dilution claims under the VRA). This court has also used the "influence district" terminology in this sense. Vecinos de Barrio Uno v. City of Holyoke, 72 F.3d 973, 990-91 (1st Cir. 1995).

The confusion stems from the intersection of this type of influence claim and another type, in which a minority group constituting less than fifty percent of the electorate can elect a candidate of its choice with the help of crossover votes from voters in the majority group. See R.H. Pildes, Is Voting Rights Law Now at War with Itself? Social Science and Voting Rights in the 2000s, 80 N.C. L. Rev. 1517, 1539-40 & n.60 (2002) (referring to this latter type of district as a "coalitional district"); Note, The Future of Majority-Minority Districts in Light of Declining Racially Polarized Voting, 116 Harv. L. Rev. 2208, 2209-10 & n.13 (2003). We will refer to this second type of influence claim as a

13

"crossover district."7 The Supreme Court has not had the opportunity to address this distinction;8 the Court in Voinovich used the term "influence district" to describe a crossover district -- one in which minorities could, despite the inability to form a majority, "elect their candidate of choice nonetheless if they are numerous enough and their candidate attracts sufficient cross-over votes." 507 U.S. at 154.9 Plaintiffs, for their part, forswear any claim under the ability to influence rubric, choosing to stand or fall entirely on an ability to elect claim. However, they do so only as far as the

7 We use "crossover" in a specialized sense with regard to racial blocs. The term is also used in a different sense when members of one political party cross over to vote in the other party's primary. See Easley v. Cromartie, 532 U.S. 234, 245 (2001).

8 The recent Supreme Court decision in Georgia v. Ashcroft, 123 S.Ct. 2498 (2003), considered influence districts and crossover districts in the § 5 context, but did not resolve the relationship between the two.

9 Crossover districts where plaintiffs allege an ability to elect also may be confused with a third type of claim, a "minority coalition" claim, in which two separate minority groups allege that a district could be formed in which they could join forces to elect a representative. See De Grandy, 512 U.S. at 1020 (describing such a VRA claim); Concerned Citizens v. Hardee County Bd., 906 F.2d 524, 526-27 (11th Cir. 1990) (indicating that minority coalition claims meet the first Gingles precondition); Brewer v. Ham, 876 F.2d 448, 453 (5th Cir. 1989) (same). But see Nixon v. Kent County, 76 F.3d 1381, 1392 (6th Cir. 1996) (en banc) (rejecting a minority coalition claim).

We take no position on that issue. Plaintiffs do not allege that they and another minority group form a minority coalition and that such a coalition may qualify as a "class" under § 2. Rather, this suit appears to posit that the interests of African-American voters have been pitted against the interests of Hispanic voters.

14

term "influence district" describes one where a minority group is unable to elect a candidate of its choice even with crossover support. They use the term "influence district" in the complaint to describe what we label a crossover district, in which African- American voters have an ability to elect with crossover support. We consider only this type of influence claim, and not the more nebulous variety described in Barnett and disavowed by plaintiffs.

The Supreme Court has expressly held open the question of whether the Gingles preconditions should apply to influence claims. See De Grandy, 512 U.S. at 1009; Voinovich, 507 U.S. at 154; Gingles, 478 U.S. at 46 n.12. We read the language of these cases, especially Gingles and Voinovich, to profess a willingness to consider a crossover district claim such as the one plaintiffs plead. The Gingles language setting aside the question of an influence claim did not differentiate between crossover district claims and claims in which plaintiffs profess only an ability to affect, not determine, electoral outcomes. But the Court has not flatly refused to consider a crossover district despite the opportunity to do so. See Voinovich, 507 U.S. at 154.

The Supreme Court's recent opinion in Georgia v.

Ashcroft, 123 S.Ct. 2498 (2003), also supports our conclusion that crossover districts should be considered in the § 2 context.

15

Georgia interpreted § 5 of the VRA.10 The Court has repeatedly warned that § 2 and § 5 "combat different evils and . . . impose very different duties upon the States." Reno v. Bossier Parish Sch. Bd., 520 U.S. 471, 476 (1997). Despite the differences between § 2 and § 5 analysis, the Court's treatment of influence and crossover districts in Georgia is highly instructive.

Georgia held that when assessing retrogression, courts must consider not only majority-minority districts but also the existence of influence districts, including crossover districts. 123 S.Ct. at 2512. The state's plan created two additional districts with a minority population of between thirty and fifty percent, and two districts with a population of between twenty-five and thirty percent. These districts, the Court found, were crucial to determining the overall effect of the new redistricting plan. Id. at 2515. Indeed, the Court was unanimous that crossover districts should be considered in the § 5 analysis; the dissent objected only to the use of those influence districts in which it was not clear that minority voters would have an ability to elect even with crossover support. See id. at 2513, 2514; id. at 2518-19 (Souter, J., dissenting). If crossover districts are important

10 Under § 5, the Attorney General of the United States must preclear a covered jurisdiction's "standard, practice, or procedure." 42 U.S.C. § 1973c. Preclearance depends on whether the change "would lead to a retrogression in the position of racial minorities with respect to their effective exercise of the electoral franchise." Beer v. United States, 425 U.S. 130, 141 (1976).

16

enough to minority voters to be considered when assessing a redistricting plan's retrogression, it would be an odd result if the same voters could not bring a § 2 claim when such a crossover district is eliminated by redistricting.

Given these Supreme Court precedents, we believe that whatever the status of other influence claims, at least crossover district claims are cognizable under § 2 of the VRA. We decline to hold, as a matter of law, that they are not.

This conclusion is consistent with our decision in Vecinos de Barrio Uno, supra, where this court held that an "influence district" that was twenty-eight percent Hispanic should be considered in the determination of whether the minority population's voting strength had been diluted. 72 F.3d at 990-91 ("[T]he voting strength of a minority group is not necessarily limited to districts in which its members constitute a majority of the voting age population, but also extends to every district in which its members are sufficiently numerous to have a significant impact at the ballot box most of the time."). Unlike the present case, the influence district in Vecinos de Barrio Uno was used by the defendant city as evidence that the minority population retained political power. Moreover, the city was not alleging that the minority group could elect its own candidate with crossover support, but only that it was large enough to wield influence over the outcome. Despite these factual differences, this court's

17

recognition that influence districts may be used to show the existence of the political power of minority groups reinforces the decision to recognize, at least in theory, a suit complaining that a crossover district has been unjustly eliminated.

Though Gingles did not apply the preconditions to influence claims, however they are defined, some preconditions must apply in order to link the complained-of voting practice with the harm the plaintiffs allege. Gingles, 478 U.S. at 48-51; see Vecinos, 72 F.3d at 979 n.2 ("[The first] precondition will have to be reconfigured to the extent that the courts eventually validate so-called influence dilution claims."). For the purposes of this discussion, we assume that plaintiffs' claim must satisfy the second and third Gingles preconditions, and that some form of the first precondition will also apply.

2. Majority Requirement The first Gingles precondition requires that "the minority group must be able to demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district." 478 U.S. at 50. Some courts have read this literally to mean that unless plaintiffs can show that they can constitute an absolute majority in a single district -- that is, more than fifty percent -- then there is no possible § 2

18

claim.11 See Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 852-53 (5th Cir. 1999), cert. denied, 528 U.S. 1114 (2000); Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368, 371-73 (5th Cir. 1999), cert. denied, 528 U.S. 1114 (2000); see also Negron v. City of Miami Beach, 113 F.3d 1563, 1571 (11th Cir. 1997); Parker v. Ohio, 263 F. Supp. 2d 1100, 1104-05 (S.D. Ohio 2003). That approach has been criticized as a "talismanic requirement, divorced from any underlying functional reasons." Pildes, supra, at 1555. If that approach were followed here, plaintiffs' complaint would fail to meet the first precondition.

We reject the conclusion that no § 2 cause of action is ever stated, regardless of the nature of the claim, unless plaintiffs can show that a minority group would be a literal majority in a single district. The approach is inconsistent with the Supreme Court's own descriptions of the functions served by the first Gingles precondition. It is also inconsistent with the variety of political realities the VRA was meant to address; a demographic fact of life in some areas of the country is that no single racial group constitutes an absolute majority. And finally, it contravenes the plain text of § 2, which requires courts to consider the "totality of the circumstances."

11 On defendants' theory, a discrete, geographically compact racial group (here, African-Americans) is not entitled to avail itself of § 2 of the VRA until it is large enough to constitute a numerical majority in any given district.

19

Requiring the protected class to show that it is an absolute majority ignores the reality that the class could elect its preferred candidate without such numbers. Thus, a discussion of whether the protected class forms a "majority" is not necessarily helpful in determining whether an "electoral law, practice, or structure interacts with social and historical conditions," Gingles, 478 U.S. at 47, to impair the ability of the class to vote. The plaintiffs here have alleged that African- American voters formed a politically cohesive group that was able, with the assistance of crossover voting, to elect the candidate of its preference in a district that was less than fifty percent African-American and that, in a properly drawn district, they could continue to do so.

In the context of this case, that pleading suffices to satisfy the interests identified by the Supreme Court for the first Gingles precondition. That precondition should not be read without regard to its function: to determine whether "the ability of minority voters to elect representatives of their choice" is impeded. Gingles, 478 U.S. at 48. As the Court has noted, "the Gingles factors cannot be applied mechanically and without regard to the nature of the claim." Voinovich, 507 U.S. at 158.12 Gingles

12 Consonant with its holding that the Gingles preconditions are in some form applicable to single-member districts, the Supreme Court has consistently avoided applying the first precondition to challenges to such districts. See De Grandy, 512 U.S. at 1009 (assuming the first precondition is satisfied); Voinovich, 507 U.S.

20

itself, in reviewing a multi-member district, noted that the function of the first precondition was to assure that there was a causal relationship between the creation of the district lines and the harm to the plaintiffs; if the minority group's candidate could not prevail even in a single district, then "the multimember form cannot be responsible for minority voters' inability to elect its candidates." 478 U.S. at 50 (emphasis removed). Similarly, when discussing majority bloc voting in the context of the third precondition, Gingles defined it as that which is sufficient usually to "defeat the combined strength of minority support plus white crossover votes." Id. at 56 (internal quotation marks omitted); see also Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1123 (3d Cir. 1993) ("[T]he Gingles [third precondition] standard presupposes the existence of crossover voting."). Growe reinforced this functional analysis, noting that "the 'geographically compact majority' and 'minority political cohesion' showings are needed to establish that the minority has the potential to elect a representative of its own choice in some single-member district." 507 U.S. at 40. Such support for a functional approach leaves room to include claims in which an electoral majority is formed only with crossover support.

This functional approach also better accounts for various political realities. In electoral schemes in which representatives

at 158 (same); Growe, 507 U.S. at 41 (same).

21

can usually be elected with less than a majority of the vote, Gingles should not be read to require that the minority group nevertheless be able to form a literal majority in a reconfigured district. Rhode Island law specifically provides that candidates in both primary and general elections for state office may be elected with a plurality of the vote. See R.I. Const. art. IV, § 2 (general elections); R.I. Gen. Laws § 17-15-29 (2002) (primary elections). In such cases, constituting a majority would not be necessary for minorities to "elect a representative of their choice." 42 U.S.C. § 1973(b); see Romero v. Pomona, 883 F.2d 1418, 1424 n.7 (9th Cir. 1989), overruled on other grounds, 929 F.2d 1358 (9th Cir. 1990).13 "To the extent that courts have read Gingles to elevate the ability to create a district with a majority-black electorate into a threshold requirement for establishing liability in all vote dilution litigation, they have improperly applied one particular theory of liability to other distinct types of vote dilution." P.S. Karlan, Maps and Misreadings: The Role of Geographic Compactness in Racial Vote Dilution Litigation, 24 Harv. C.R.-C.L. L. Rev. 173, 202 (1989).

Georgia v. Ashcroft, supra, confirms that influence

13 But see Brewer, 876 F.2d at 454 (requiring a majority even for plurality elections, reasoning that a "plurality feature is of course more responsive to minority voter groups"); McNeil v. Springfield Park Dist., 851 F.2d 937, 943-44 (7th Cir. 1988) (rejecting any showing of less than an absolute majority as unduly speculative).

22

districts, including crossover districts, are important to any practical assessment of minority voting power. In Georgia, the Court emphasized the fact-bound nature of VRA claims, holding that the retrogression inquiry under § 5, like the dilution inquiry under § 2, see 42 U.S.C. § 1973(b), requires an assessment of the "totality of the circumstances." 123 S.Ct. at 2511. One reason for this broad factual inquiry, the Court indicated, is that "[t]he ability of minority voters to elect a candidate of their choice is important but often complex in practice to determine." Id. Accordingly, the Court held, influence and crossover districts must be considered as part of that determination in the retrogression context. Id. at 2512. The Court also cited empirical studies indicating that such districts may maximize minority voting strength. Id. at 2512-13.

We also consider relevant both modern and historical political realities. During the 1970s and 1980s, African-American populations usually could not elect representatives of their choice unless they constituted a majority in an electoral district. See generally Quiet Revolution in the South (C. Davidson & B. Grofman eds., 1994). Indeed, usually a mere majority was not sufficient; many believed that to overcome racial bloc voting patterns, the total minority population needed to be sixty-five percent. See Ketchum v. Byrne, 740 F.2d 1398, 1415-16 (7th Cir. 1984) (collecting sources). But the percentage of minority population

23

necessary to elect a candidate has been steadily declining. By 1990, fifty-five percent was generally considered sufficient. And thereafter, due to increased white crossover voting, the number has slipped below majority level. One study reported that during the 1990s, an African-American candidate could be elected from a congressional district that was between thirty-three and thirty- nine percent African-American. B. Grofman, L. Handley & D. Lublin, Drawing Effective Minority Districts: A Conceptual Framework and Some Empirical Evidence, 79 N.C. L. Rev. 1383, 1407-09 (2001). The percentage of minority voters necessary to elect a candidate depends heavily on the political makeup of the district as a whole, see Pildes, supra, at 1535-36, a matter difficult to determine on a motion to dismiss a complaint.

In sum, it is not an absolute bar to a claim under § 2 of the VRA that some amount of crossover voting is needed for a minority group to elect a candidate of its choice.14 See Armour v.

14 Our dissenting colleague engages in the sort of factual predictions that courts are forbidden to indulge on a motion to dismiss. See Gonzalez-Gonzalez v. United States, 257 F.3d 31, 37 (1st Cir. 2001) (refusing to engage in "speculation" on appeal from a Rule 12(b)(6) dismissal and instead assuming the truth of the averments in the complaint). The dissent assumes that a minimum level of crossover voting of 32% will be required for the African- American minority to elect a candidate of its choice, and it assumes that only two-candidate contests are relevant. It assumes that the African-American voters' inability to elect their preferred candidate "can much more readily be attributed to candidate-specific issues" than to the reduction in the representation of the African-American community in the political process. It assumes that plaintiffs will establish no history of discrimination against black citizens in the political process. It

24

Ohio, 775 F. Supp. 1044, 1059-61 (N.D. Ohio 1991) (three-judge court); see also McNeil, 828 A.2d at 852-53; Powers, 263 F. Supp. 2d at 1109-1113 (three-judge court) (Gwin, J., concurring in judgment); West v. Clinton, 786 F. Supp. 803, 807 & n.2 (W.D. Ark. 1992) (three-judge court).

Though a claim that includes crossover voting may be cognizable under the first precondition for a § 2 cause of action, not every such claim will pass muster. It would be discordant with the Act, for instance, to consider a crossover district claim from a numerically tiny minority population that can only claim a hope to elect a candidate with an overwhelming number of crossover votes. Several limiting principles readily present themselves.15

assumes that there will not be sufficient bloc voting by Hispanic voters after the redistricting to defeat the African-American community's candidate of choice.

Further, the dissent assumes that "whites and Hispanics would have to cast almost half of the votes needed for a successful senatorial candidacy." In a plurality race, as the dissent apparently concedes, that is almost certainly untrue. Even in a two-candidate race, it may be untrue -- given, for example, low overall voter turnout, high African-American turnout, and African- American bloc voting. Similarly, the dissent assumes that a 5% reduction in the African-American population is insignificant. But the former African-American state senator may have lost the election by that 5%.

Each of these assumptions reflects factual inferences that, by law, must be made in plaintiffs' favor on a motion to dismiss. United States v. AVX Corp., 962 F.2d 108, 114 (1st Cir. 1992).

15 We reach a determination only with regard to crossover districts, the sole type of influence claim presented in this appeal. We reach no conclusion concerning other types of influence claims, which, if they are recognized, may require a different application of the Gingles preconditions and different limiting principles.

25

First, this case presents a claim not merely of an abstract hope to elect the African-American voters' preferred candidate through both African-American and crossover voting. The alleged loss is much more concrete. Historically, the African- American community's preferred candidate was consistently elected, even though African-American voters were less than a numerical majority in the district. The redistricting plan, however, significantly reduced the percentage of African-American voters in the district, and the candidate lost his bid for reelection. African-American voters sued, saying they had been denied an equal opportunity to elect the candidate of their choice, and on this motion to dismiss, the redistricting plan must be taken as the cause of the lost election. That is not to say that a history of electoral success is a necessary part of a successful claim, especially if the lack of success is due to historic vote dilution, but the minority group's historical voting success makes this an easier case.

The second is the statutory requirement that a minority population be able to elect, in a potential district, "representatives of their choice." 42 U.S.C. 1973(b) (emphasis supplied). A minority group may require so many crossover votes that it does not truly have the capacity to choose its own candidate, but only to help elect candidates chosen by other groups. If so, plaintiffs cannot make a crossover district claim.

26

Here, however, the plaintiffs clearly plead in their complaint that the African-American community can elect its own candidate with crossover support in a properly drawn district.

The third limitation is expressed in the third Gingles precondition: "the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it . . . usually to defeat the minority's preferred candidate." 478 U.S. at 51. A minority population that is too small, and that therefore requires too high a level of crossover support, will not be able to meet the third precondition. If the majority population is willing to provide crossover support to minority-chosen candidates at very high levels, then it cannot be said to be voting as a bloc against these candidates. For the reasons described below, the plaintiffs' complaint offers enough on the third precondition to render a Rule 12(b)(6) dismissal on that ground inappropriate.16 Finally, we note that this is not a situation, as in Georgia v. Ashcroft, supra, where the leaders of the African- American community developed the redistricting plan at issue. The contrary is true here. With only one state senator out of fifty, the African-American community had precious little political strength in the senate before the redistricting. After the plan

16 The parties agree that the second Gingles precondition is met by the plaintiffs' pleading that "African-American voters in the State of Rhode Island are politically cohesive."

27

was implemented, they lost their only representation. Similarly, this is not a case about the failure to maximize potential African- American voting power. Cf. Abrams v. Johnson, 521 U.S. 74 (1997). Rather, this is a case about the elimination of African-American voters' opportunity to elect the candidate of their choice, an opportunity that they had consistently enjoyed prior to the redistricting. D. Third Precondition: Majority Bloc Voting The third Gingles precondition requires that "the minority must be able to demonstrate that the white majority votes sufficiently as a bloc to enable it . . . usually to defeat the minority's preferred candidate." 478 U.S. at 51. Again, this is part of a functional approach to the Act. See id. ("In establishing this last circumstance, the minority group demonstrates that submergence in a white multimember district impedes its ability to elect its chosen representatives."). This court has described the third condition as addressing "whether the challenged practice, procedure, or structure is the cause of the minority group's inability to mobilize its potential voting power and elect its preferred candidates," Vecinos, 72 F.3d at 980, and referred to the bloc simply as a "majoritarian" bloc, id. at 981, 982.

The district court held that the plaintiffs' complaint failed to satisfy this third condition. First, it read Gingles

28

specifically to require that the majority bloc must be a white numerical majority. While the complaint does not specify the white population of Senate District 2, it can be no more than 31.84% after subtracting the African-American and Hispanic population. Furthermore, the court calculated that because the old district was only twenty-six percent African-American, the crossover vote needed to elect the candidate preferred by African-American voters would have to constitute twenty-four percent of the electorate. If half of that crossover vote were white, the court reasoned, then the white population would be crossing over at a rate of about one- third, which it thought too high to be consistent with "bloc voting" needed to "defeat the minority's preferred candidate." Metts, 217 F. Supp. 2d at 260-61.

This reasoning has several flaws. Inherent in the court's analysis is the assumption that the electoral contest would have only two candidates. That is not necessarily so, particularly in primaries. In many jurisdictions, the winner of a particular party's primary is de facto the winner of the general election; it may be inferred that this was historically the case in the old Senate District 9 and it remains true in the reconfigured Senate District 2. The court also allocated the crossover vote half to whites and half to Hispanics, but there is simply no evidence of the racial composition of the crossover votes, either historically or in a proposed alternative district.

29

Further, we reject the district court's impermissible focus on only the white voters in Senate District 2 for purposes of the third Gingles precondition. The plaintiffs, in their complaint, claim that "[t]he white and Hispanic communities vote sufficiently as a bloc usually to defeat the candidate of choice of African-American voters when that candidate is African-American and the district is less than twenty-six percent black in total population." Under the standards of Rule 12(b)(6), such claims in the complaint may be rejected only if they are "bald assertions" or "unsupportable conclusions." Chongris v. Bd. of Appeals, 811 F.2d 36, 37 (1st Cir. 1987). Neither characterization can be said to be true here.

The VRA does not, by its terms, afford protection to or against any particular racial or ethnic group; if it did, it might well be suspect under the Equal Protection Clause. U.S. Const. amend. XIV, § 1. The language of Gingles referred to a "white majority" only because that happened to be the composition of the majority on the facts before the Court. See 478 U.S. at 51.

Nor must the majority bloc be comprised of only one race.

While the "protected class" being discriminated against must be constituted of a particular "race or color," see 42 U.S.C. § 1973(a), there is no requirement in the VRA that a contrary voting bloc be of just one race. Coalitions of certain races that characteristically vote against the preferred candidate of a

30

different racial group may well constitute bloc voting for purposes of the third Gingles precondition. In De Grandy, the Supreme Court considered such a challenge to Florida's state legislative districts. In one county, there were three large voter groups: African-Americans, Hispanics, and whites. The trial court found, based on expert testimony, that during elections pitting a minority candidate against a white one, the white voters would vote as a bloc along with the other minority group's voters to elect the white candidate. De Grandy v. Wetherell, 815 F. Supp. 1550, 1572 (N.D. Fla. 1992) (three-judge court). The Supreme Court did not find this fact pattern problematic as a means to satisfy the third Gingles precondition. De Grandy, 512 U.S. at 1007. In a similar case, also reviewing a challenge to a Florida redistricting plan, the Eleventh Circuit found that "a coalition of Hispanics and Non Latin Whites could form the relevant majority voting bloc for the purpose of the third Gingles factor." Meek v. Metro. Dade County, 908 F.2d 1540, 1545-46 (11th Cir. 1990). We agree.

The district court's rationale does, however, highlight a potential difficulty with the plaintiffs' complaint. In the plaintiffs' proposed remedial district, the African-American population would be at least twenty-six percent. Depending on how the facts are developed, that number may raise issues related to the third precondition. If it is true that a majority (rather than only a plurality) is needed to elect a candidate, if the racial

31

makeup of the voters is proportional to the racial composition of the district, and if, as the plaintiffs assert, the African- American voters are politically cohesive, then crossover voting would need to reach twenty-four percent. In that scenario, thirty- two percent of the non-African-American voters would have to support the African-American community's chosen candidate in order to reach the majority needed. The district court believed that such a high rate of crossover voting would be inconsistent with a finding of bloc voting.

At the Rule 12(b)(6) stage, that conclusion is premature.

A series of factual assumptions would be required to judge the average level of crossover support the plaintiffs are alleging, assumptions that cannot be confirmed or repudiated at this stage of the proceedings. While the complaint's description of a reconfigured district includes total population figures, there is no demographic information concerning the voting age population, the number of registered voters, or the expected voters in any given election, much less the typical voting patterns of various groups. There are also no facts about the number of candidates that typically run in the primary or general elections. Without such information, it is impossible to know the percentage of crossover support necessary to elect the candidate of the African- American community's choice.

More importantly, even if the facts show that crossover

32

voting of thirty-two percent would be required, that number, without more, does not warrant a Rule 12(b)(6) dismissal for failure to state a claim in light of the third Gingles precondition. The statute commands an examination of "the totality of circumstances." 42 U.S.C. § 1973(b). An inquiry into the third precondition is thus an inherently factual enterprise. "[T]he degree of racial bloc voting that is cognizable as an element of a § 2 vote dilution claim will vary according to a variety of factual circumstances." Gingles, 478 U.S. at 57-58; see Vecinos, 72 F.3d at 989. The Supreme Court has been chary of per se rules in this area, whether the claim is that a device is a per se violation of § 2, Voinovich, 507 U.S. 154, or whether the claim is that a single factor is a safe harbor for defendants, De Grandy, 512 U.S. at 1017-18.

One important factor about which the record is undeveloped is the pattern of voting behavior over time. Gingles stressed the importance of determining whether racial bloc voting is a pattern extending over time or merely a phenomenon in a single election. 478 U.S. at 57. The reverse is also true: the success of a minority candidate, or the absence of bloc voting in a few elections, cannot be taken to mean that the district does not experience racial bloc voting overall. Id.

Furthermore, a crossover rate of thirty-two percent is within the range of fact patterns in which courts have found

33

majority bloc voting. Gingles itself found majority bloc voting where the majority group supported African-American candidates in the general election at a rate between twenty-eight and forty-nine percent, with an average support of one-third. Id. at 59; see Campos v. Baytown, 840 F.2d 1240, 1249 (5th Cir. 1988) (finding majority bloc voting when the crossover vote was thirty-seven percent). Of course, that does not mean that a crossover rate less than one-third would always disprove majority bloc voting: in other circumstances, the Supreme Court has found a crossover rate averaging between twenty-two and thirty-eight percent sufficient to suggest "a general willingness of white voters to vote for black candidates," especially when minority candidates have a record of success.17 Abrams v. Johnson, 521 U.S. 74, 93 (1997) (internal quotation marks omitted).

At this stage of the litigation there is no evidence of the degree or effect, if any, of racially polarized voting, or whether a voting district could have been constituted to protect the ability of both African-American and Hispanic voters to elect candidates of their choice.

17 The dissent's citation to Abrams v. Johnson, 521 U.S. 74, 92-93 (1997), does not assist it. Abrams was decided after trial on a full record. The record revealed that there was, over time, an increased general willingness of white voters to vote for black candidates, and a corresponding decrease in racial polarization. The Abrams Court did not purport to establish a mathematical litmus test for screening cases under the third Gingles precondition on a motion to dismiss.

34

The dissent misses the point when it objects that the VRA is not meant "to ensure the success of candidates favored by minority groups." In this case, it is undisputed that a minority group's preferred candidate, an incumbent, failed to win reelection in the first election after the state legislature adopted a redistricting plan that decreased the percentage representation of that minority in the candidate's home electoral district. No court has ever held -- and it would be clear error for a court to hold -- that such a defeat is irrelevant to the question whether members of that minority group "have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice." 42 U.S.C. § 1973(b). Plaintiffs seek equality of opportunity, not a guarantee of electoral success.

The district court's dismissal of the plaintiffs'

complaint under Rule 12(b)(6) for failure to meet the third Gingles precondition was inappropriate.

IV.

Congress, in enacting the Voting Rights Act, was cognizant of this country's long and shameful history of excluding African-Americans from our political processes. In light of the purposes of the Fourteenth Amendment and its guarantee of equal protection of the laws, it is no answer to say, as the dissent does, that the federal courts should close their doors to possibly meritorious complaints under the Voting Rights Act out of deference

35

to majoritarian will and "difficult" legislative choices. That is particularly true at this early stage of the litigation. Plaintiffs must still prove their case; they must establish the Gingles preconditions, as well as a substantive violation of the VRA, with evidence. There is no frustration of majoritarian will in requiring the defendants to respond to allegations that state a claim under the Act. The dissent describes such a claim as a quest for "unfair advantage." To the contrary, the complaint states a claim of unfair disadvantage to African-American voters in the exercise of the most important right in our American democracy. If plaintiffs ultimately prevail, it will be because they have proven that the Rhode Island legislature, acting for the majority, has violated the Voting Rights Act by impermissibly denying members of the African-American community in Providence an equal opportunity to elect a state senator of their choice. As the Supreme Court stated in Georgia v. Ashcroft, "[t]he purpose of the Voting Rights Act is to prevent discrimination in the exercise of the electoral franchise and to foster our transformation to a society that is no longer fixated on race." 123 S.Ct. at 2517.

The plaintiffs must be given the opportunity to prove their case. We express no view as to the outcome. The dismissal of the plaintiffs' complaint is reversed and the case is remanded for proceedings consistent with this opinion.

36

Dissenting opinion follows.

37

SELYA, Circuit Judge (dissenting). Although it is regrettable that redistricting may make it more difficult for a candidate preferred by African-American voters to win election to the Rhode Island state senate, the Voting Rights Act is not intended as a means of ensuring that every minority group has exactly the district lines that it deems most advantageous. In the circumstances of this case, I can discern no valid legal basis for us to superimpose the appellants' will on that of the Rhode Island General Assembly. I would therefore affirm the district court's dismissal of the amended complaint.

My reasoning is rooted in precedent. The Supreme Court has made it reasonably clear that, as a threshold requirement for the maintenance of a vote dilution claim in a single-member district, plaintiffs must show (1) that they are part of a minority group that is sufficiently large and geographically compact to constitute a majority in some plausible iteration of the affected district; (2) that the group is politically cohesive; and (3) that other racial groups engage in bloc voting significant enough to defeat the minority group's preferred candidate. Voinovich v. Quilter, 507 U.S. 146, 157-158 (1993) (applying Thornburg v. Gingles, 478 U.S. 30, 50-51 (1986), to single-member districts); Growe v. Emison, 507 U.S. 25, 40 (1993) (same). Given the nature of their claim, I think it is fairly clear that the appellants cannot satisfy the first precondition. It is absolutely clear,

38

however, that they cannot simultaneously satisfy both the first and third preconditions.

I start with the first Gingles precondition. The appellants concede that they are unable to show that African- Americans can constitute a literal majority in any plausible iteration of Senate District 2. Rather, their complaint rests on the novel premise that a minority group whose members cannot conceivably comprise a numerical majority, even in what is from their point of view an ideally configured single-member district, nonetheless can mount a viable vote dilution claim by demonstrating that the district's lines could have been drawn in such a way as to give the minority group the ability to elect the candidate its members prefer. Whether or not this type of claim ever can fall within the purview of section 2 of the Voting Rights Act (VRA), 42 U.S.C. § 1973, the instant claim does not. The identified minority group is so small and its need to rely on crossover voting so great that the appellants' section 2 claim necessarily fails.

It is common ground that courts must apply the Gingles preconditions to the ideal district proposed by those who challenge a redistricting plan. See Holder v. Hall, 512 U.S. 874, 880 (1994) (opinion of Kennedy, J.); Negrón v. City of Miami Beach, 113 F.3d 1563, 1571 (11th Cir. 1997). In this case, the appellants concede that the adoption of a constitutional amendment downsizing the General Assembly required redistricting of the state senate. Thus,

39

Senate District 9 could not remain intact and had to be reconfigured. With this in mind, the appellants describe their ideal version of Senate District 2 as a district in which African- Americans comprise roughly 26% of the population. This configuration differentiates the appellants' case from the mine- run. Typically, vote dilution claims address redistricting schemes that take a racial minority group whose members have the potential to comprise a numerical majority in a geographically compact district and disperse the group across two or more districts (with the result that its members constitute a majority in none). See Voinovich, 507 U.S. at 153.

My colleagues suggest that we can change the paradigm because the appellants may have been deprived of a "crossover district" — a kind of "influence district" in which a numerical minority is so positioned that it has an ability to swing elections. See generally id. at 154 (describing an influence district as one in which minority group members "could not dictate electoral outcomes independently [but] could elect their candidate of choice nonetheless if they are numerous enough and their candidate attracts sufficient cross-over votes from white voters"). The Supreme Court has repeatedly refrained from deciding the cognizability of claims based on legislative dismantling of crossover districts, e.g., Johnson v. De Grandy, 512 U.S. 997, 1008-09 (1994); Voinovich, 507 U.S. at 154, and, until today, this

40

court has exhibited the same restraint, e.g., Vecinos de Barrio Uno v. City of Holyoke, 72 F.3d 973, 979 n.2 (1st Cir. 1995).

To be sure, the preservation of influence or crossover districts may constitute a relevant factor in defending against a vote dilution claim. See, e.g., De Grandy, 512 U.S. at 1020; Vecinos, 72 F.3d at 990-91 & n.13; Latino Political Action Comm. v. City of Boston, 784 F.2d 409, 414-15 (1st Cir. 1986). But to say that a court may consider crossover districts in deferring to a state's redistricting plan is very different from saying that a minority has the legal right, under section 2 of the VRA, to demand that the legislature establish such a district. The ultimate goal of the VRA is "transition to a society where race no longer matters." Georgia v. Ashcroft, 123 S. Ct. 2498, 2517 (2003). Thus, minorities still bear the burden "to pull, haul, and trade to find common political ground, the virtue of which is not to be slighted in applying a statute meant to hasten the waning of racism in American politics." Id. at 2512 (quoting De Grandy, 512 U.S. at 1020) (internal quotations marks omitted). My colleagues' freshly minted "functional approach" would significantly lighten this burden.

In all events, deciding this case does not require us to go so far as to rule out all section 2 claims based on a legislature's failure either to assemble or to preserve a crossover district. Here, the raw numbers are inimical to such a claim. The

41

pertinent demographic for analysis of the first Gingles precondition is the voting age population. See Growe, 507 U.S. at 38 n.4; Ketchum v. Byrne, 740 F.2d 1398, 1412-13 (7th Cir. 1984). In this case, the appellants maintain that African-Americans represented approximately 26% of the voting age population in former Senate District 9 yet represent only 21% of the voting age population in the new district (Senate District 2). They claim that this 5% differential is a political kiss of death.

Stripped of rhetorical flourishes, the appellants' thesis proceeds along the following lines. Whenever a candidate preferred by African-Americans runs for the state senate in the new district, he or she will receive all the African-American votes plus no less than 32% but no more than 37% of the combined white and Hispanic votes (these being the percentages of all white and Hispanic voters necessary to form a majority in conjunction with African-American voters when African-Americans constitute 26% and 21% of the population, respectively).18 In the appellants' view, those crossover voters will favor the African-Americans' preferred candidate regardless of the race or politics of his or her opponent(s). Consequently, the redistricting plan is vulnerable under section 2 of the VRA because the electorate's polarization is

18 This estimate is conservative. To the extent that voter registration or voter turnout differs, or that African-Americans are not completely monolithic in their voting preferences, the needed thresholds become harder to achieve.

42

so deeply entrenched that candidate-specific variations will operate only within a 5% margin.

Whether viewed as a matter of logic, political science, or human behavior, this prediction strikes me as utterly conjectural. For good reason, a difference of a few percentage points in the minority population of a single-member district generally has been thought unlikely to affect election outcomes. See S. Christian Leadership Conf. v. Sessions, 56 F.3d 1281, 1296 (11th Cir. 1995) (en banc). The appellant's claim flies in the teeth of this conventional wisdom — and the mere fact that one very popular candidate, running uphill, had a series of successes in the "old" district does not validate the appellants' claim.

Even if I must indulge the claim because the district court chose to act at the Rule 12(b)(6) stage, the most that can be said is that the appellants had forged a sort of functional majority in former Senate District 9. By that I mean that African- Americans, though neither numerous nor concentrated enough to comprise a majority in the district, exhibited an ability to elect a particularly appealing candidate with the aid of a large and predictable non-African-American crossover vote. But whatever may be said for functional majority claims in general, the appellants' functional majority claim is a non-starter. Where, as here, a minority group comprises only a relatively small fraction of the total population of an electoral district before redistricting, the

43

inability of group members to elect the candidate of their choice after redistricting can much more readily be attributed to candidate-specific issues than to a slight reduction in their numbers. In all events, a minority group of that modest size must rely so heavily on crossover votes, both before and after redistricting, that section 2 of the VRA provides no safe harbor.

The figures tell the tale. In the appellants' idealized district, whites and Hispanics would have to cast almost half of the votes needed for a successful senatorial candidacy. Those votes would not correlate with the individual voter's race, but, rather, with the race of the candidate, or, alternatively, with the race of the minority group members with whom the crossover voters identify. This fact has two important implications. In the first place, it confirms that, regardless of how the district's lines are drawn, African-Americans by themselves do not have anything close to an ability to elect the candidate of their choice. In the second place, it demonstrates that the appellants' claim puts the emphasis not on assuring equal opportunity for minority voters but on assuring a victory by the African-Americans' preferred candidate. That is the wrong emphasis. See 42 U.S.C. § 1973(b) (identifying "members of a class of citizens," not candidates, as the operative unit of statutory protection); see also De Grandy, 512 U.S. at 1014 n.11 (explaining that "the ultimate right of § 2 is equality of opportunity, not a guarantee of electoral success");

44

Smith v. Brunswick County Bd. of Supervisors, 984 F.2d 1393, 1400 (4th Cir. 1993) (abjuring classification of protected groups by the way they vote rather than by their race; to do otherwise would impermissibly "resolv[e] discrimination issues on the basis of whether members of the protected group are elected"); cf. Gingles, 478 U.S. at 99-100 (O'Connor, J., concurring) (agreeing that, for purposes of a section 2 claim, voting must correlate with the race of the voter).

The existence of this misdirected emphasis is borne out by the fact that the appellants' proposed reconfiguration of Senate District 2 would strengthen the electoral power not only of the African-American community but also of the sizable white and Hispanic crossover vote. That increased political clout would come at the expense of the remaining two-thirds of the white and Hispanic voters. The VRA empowers courts to protect the rights of a minority group to participate in the electoral process so that such a group, if treated fairly, can become a majority. It does not give courts the raw power to privilege the interests of the few over the interests of the many, much less the power to override the normal functioning of the majoritarian process. See Vecinos, 72 F.3d at 982; Smith, 984 F.2d at 1400-02.

The democratic system remains the best and fairest electoral system ever devised. Even so, the realities of democracy are sometimes harsh. The appellants seek to avoid these realities

45

by tempting us to treat crossover voters as if they constitute part of a protected minority within the purview of section 2. Fidelity to core democratic values demands that we resist this temptation. While the Gingles preconditions contemplate a certain degree of crossover voting, see Gingles, 478 U.S. at 56; Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1123 (3d Cir. 1993), there is a point at which crossover voting becomes so large a part of the picture as to crowd out the possibility of a legally cognizable vote dilution claim.

That is the picture painted by the amended complaint.

Where, as in this case, the bricolage comprises a roughly equal mix of minority and crossover voters, allowing a vote dilution claim to go forward would make sense only if the end game were to ensure the success of candidates favored by minority groups. As I already have explained, however, that is not the objective of the VRA. Nor should it be; my colleagues' "functional approach" would create a topsy-turvy world in which legislatures would have to base redistricting plans not on the need to preserve legitimate majority/minority districts, but, rather, on guesswork about the way in which each constituent was likely to vote.19

19 My colleagues write that "[i]n this case, it is undisputed that a minority group's preferred candidate, an incumbent, failed to win reelection in the first election after the state legislature adopted a redistricting plan that decreased the percentage representation of that minority in the candidate's home electoral district." Maj. Op. at 34. That is not a relevant consideration. When the appellants served their amended complaint (May 14, 2002)

46

In short, I do not believe that section 2 of the VRA authorizes vote dilution claims that are wholly dependent upon massive crossover voting. There is a critical distinction between minority-preferred candidates who lose because redistricting excludes too much of the minority electorate from a particular district (illegal vote dilution) and minority-preferred candidates who lose because they do not attract enough votes from other folks within the district (legal majoritarian rule). The amended complaint, even when taken at face value, blurs this distinction.

My colleagues attempt to blunt the force of this reasoning in two ways. First, they posit that vote dilution claims must be decided based on the totality of the circumstances. That is so — but the statutory provision they cite, 42 U.S.C. § 1973(b), does not inoculate all such claims against Rule 12(b)(6) challenges. A plaintiff class must do more than cry "vote dilution" to engage the gears of the VRA. The Gingles preconditions act as a sentry at the gates — a bright-line rule that must be satisfied before the totality of the circumstances comes into play. See Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 852 (5th Cir. 1999) (collecting cases); City of Carrollton Branch of N.A.A.C.P. v. Stallings, 829 F.2d 1547, 1550-51 (11th Cir. 1987). This framework helps ensure the

and when the district court dismissed the case (September 9, 2002), no elections had yet been held under the redistricting plan.

47

effectiveness of the remedy created by the VRA without distorting either its scope or intent.

My colleagues' second gambit is to stress that Rhode Island elects its state senators under what amounts to a plurality system. This seems to me to be a bit of a red herring. On the one hand, cases holding plaintiffs to the requirements of the first Gingles precondition despite the existence of a plurality election system are ubiquitous.20 See, e.g., Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368, 370-71 (5th Cir. 1999); Stabler v. County of Thurston, 129 F.3d 1015, 1025 (8th Cir. 1997); Cane v. Worcester County, 35 F.3d 921, 924 n.4, 925 (4th Cir. 1994); McNeil v. Springfield Park Dist., 851 F.2d 937, 943-44 (7th Cir. 1988). On the other hand, cases in which courts have recognized a section 2 claim by members of a small minority group simply because they reside in a jurisdiction that employs a plurality election system are nonexistent. Moreover, plurality election rules are, as my colleagues apparently concede, more responsive to minority voters than simple majority election rules. It would be ironic to relax the first Gingles precondition for vote dilution claims arising

20 Given the widespread popularity of plurality election systems, the Supreme Court must certainly have taken their existence into account in formulating the Gingles preconditions. Cf. Voinovich, 507 U.S. at 157 (applying the Gingles preconditions without making any allowance for Ohio's use of a plurality voting system). I therefore see no reason why we should not adhere to the Supreme Court's rendition of the first Gingles precondition without engaging in rank speculation about the possibility of multiple candidacies.

48

under an electoral structure already more favorable to minorities. It also would make little sense, politically or mathematically, to proclaim that a plurality rule supports a minority group's ability to elect when its members number 26% of the electorate yet utterly forecloses that ability when they number 21% of the electorate.

Last — but far from least — my colleagues' reliance on the existence of a plurality election system ignores the vicissitudes of such systems. For example, in elections in which only two candidates are on the ballot or in which one of several candidates enjoys great popularity, minorities will have to muster a clear majority of all votes cast in order to elect the candidate of their choice. The permutations are endless. To my mind, this means that the putative effects of a plurality voting system are simply too speculative to provide a basis for a convincing vote dilution claim. See Brewer v. Ham, 876 F.2d 448, 455-56 (5th Cir. 1989); McNeil, 851 F.2d at 944. These problems may explain why the appellants never made reference to Rhode Island's plurality election laws in their amended complaint or their appellate briefs.

That ends this aspect of the matter. While I am willing to leave open the possibility that a racial minority group constituting less than 50% of the electorate in a particular single-member district may in special circumstances satisfy the

49

first Gingles precondition,21 the facts alleged in this case reflect no such special circumstances. I conclude, therefore, that the appellants' claim does not and cannot satisfy the first Gingles precondition.

If more were needed — and I doubt that it is — the appellants' claim also fails to satisfy the third Gingles precondition. That precondition requires a showing of nonminority bloc voting (which, for purposes of this case, encompasses the combined voting power of whites and Hispanics). Here, such a showing is inconsistent with the theme around which the appellants' case is constructed.

The appellants showcase Senator Walton's past electoral successes as proof of the cogency of their ability to elect claim — but this is a two-edged sword. Consistent electoral success on the part of a racial or ethnic minority group that comprises considerably less than a numerical majority of the electorate is indicative of the absence of nonminority bloc voting and, thus, is presumptively inconsistent with the third Gingles precondition.

21 Such a situation may occur, for example, where evidence of intentional vote dilution exists, e.g., Garza v. County of Los Angeles, 918 F.2d 763, 770-72 (9th Cir. 1990); Armour v. Ohio, 775 F. Supp. 1044, 1060-62 (N.D. Ohio 1991), or a minority group comprises nearly 50% of the population of a particular district, e.g., Martinez v. Bush, 234 F. Supp. 2d 1275, 1299 (S.D. Fla. 2002) (three-judge court) (per curiam), or the demographic trend lines are such that the affected minority group reasonably can be expected to attain majority status in the near future, e.g., Solomon v. Liberty County, 899 F.2d 1012, 1018 n.7 (11th Cir. 1990) (en banc) (Kravitch, J., specially concurring).

50

See Gingles, 478 U.S. at 102 (O'Connor, J., concurring); see also S. Christian Leadership Conf., 56 F.3d at 1291-94 (finding no white bloc voting where African-Americans, though less than a numerical majority, had been largely successful in electing their preferred candidates); Overton v. City of Austin, 871 F.2d 529, 540 (5th Cir. 1989) (per curiam) (similar); see also Brooks v. Miller, 158 F.3d 1230, 1241 (11th Cir. 1998) (noting that claims dependent on substantial white crossover voting are inherently inconsistent with fulfillment of the third Gingles precondition); Turner v. Arkansas, 784 F. Supp. 553, 570-71 (E.D. Ark. 1991) (three-judge court) (similar). This line of cases reflects a common-sense proposition: that the ability of a racial minority group actually to elect its preferred candidate may depend upon such a high degree of crossover voting that the third Gingles precondition inevitably fails of satisfaction. So it is here: the appellants' reliance on a high level of crossover voting, ranging upward from a minimum of 32% and nearly equalling the whole of the African-American vote, defenestrates their claim of illegal vote dilution.22

22 The appellants' argument necessarily presupposes that this crossover voting peaks at a point below 37%. That is a purely arbitrary figure and, as such, need not be credited (even for purposes of a motion to dismiss). See, e.g., Dartmouth Review v. Dartmouth Coll., 889 F.2d 13, 16 (1st Cir. 1989) (warning that courts should be wary of "unsupported conclusions, subjective characterizations, and problematic suppositions" when reviewing dismissal orders under Rule 12(b)(6)).

51

The appellants — and my colleagues — cite a few cases suggesting (or so they say) that a high rate of crossover voting does not necessarily preclude a finding of racially polarized voting. See, e.g., Gingles, 478 U.S. at 59-61 (upholding lower court's finding of white bloc voting despite white crossover voting ranging from 8% to 50%); Campos v. City of Baytown, 840 F.2d 1240, 1249 (5th Cir. 1988) (upholding lower court's finding of white bloc voting despite the fact that 3% to 37% of whites crossed over). But all of these cases addressed multi-member or at-large districts — situations that pose a much more subtle threat to minority electoral strength precisely because they require higher levels of crossover voting for minorities to prevail.23 See Growe, 507 U.S. at 40; Cane, 35 F.3d at 926; see also S. Rep. No. 97-417, at 29 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206. In the realm of challenges to single-member redistricting plans, no less an authority than the Supreme Court has held average majority crossover voting of 22% to 38% sufficient to demonstrate the "general willingness of [majority] voters to vote for [minority]

23 The appellants do cite one case, Old Person v. Cooney, 230 F.3d 1113 (9th Cir. 2000), that involves single-member districts. There, the Ninth Circuit found white bloc voting because white voting in excess of 60% defeated minority candidates in most elections. Id. at 1124-27. That case was not decided based on the rate of crossover voting, but, rather, on the regularity with which the white majority had banded together to defeat minority candidates. See id. at 1127-28 (distinguishing Abrams v. Johnson, 521 U.S. 74, 92-93 (1997), on that very ground). For that reason, the Ninth Circuit never specified what rate of crossover voting actually existed.

52

candidates," particularly in conjunction with a record of significant success by minority candidates. Abrams v. Johnson, 521 U.S. 74, 92-93 (1997) (citation and internal quotation marks omitted); cf. Voinovich, 507 U.S. at 151-52, 158 (approving lower court's finding of no majority bloc voting where "black candidates have been repeatedly elected from [single-member] districts with only a 35% black population").

To be sure, the appellants asseverate that their past victories occurred only because the "old" district (in which African-Americans comprised approximately 26% of the population) was different than the "new" district (in which African-Americans comprise approximately 21% of the population). But this modest change in the level of African-American penetration does not render the voters' track record irrelevant. Whatever the precise numbers, African-Americans were and are a numerical minority in the district — and the appellants are in effect arguing that whenever crossover voting is large enough to secure the success of a minority- preferred candidate, that crossover voting cannot be used to disprove nonminority bloc voting. Such a rule would conflict with both the realities of modern politics and the objectives of section 2. The better rule is that when African-Americans constitute a relatively small numerical minority yet repeatedly attract a crossover vote sizable enough to elect their preferred candidate,

53

that fact is highly relevant to (and, as here, may be conclusive in) an analysis of the third Gingles precondition.

In a final effort to salvage the vote dilution claim, my colleagues posit that we cannot make any determinations as to the third Gingles precondition until we have evidence of voter registration, turnout, and voting patterns. That might ordinarily be true — but the appellants have not presented us with an ordinary vote dilution claim. Rather, they make a very specific and highly idiosyncratic claim premised on the notion that at least 32% of the white and Hispanic population can be expected regularly to cross over in order to form the majority required by the first Gingles precondition. This approach inextricably intertwines the first and third Gingles preconditions, so that allowances given as to one necessarily have repercussions as to the other. See Sanchez v. Colorado, 97 F.3d 1303, 1315 (10th Cir. 1996) (remarking the interrelatedness of these preconditions); Jenkins, 4 F.3d at 1133 n.32 (same). That is one reason why this case cannot survive a motion to dismiss.

Although "the degree of racial bloc voting that is cognizable as an element of a § 2 vote dilution claim will vary according to a variety of factual circumstances," Gingles, 478 U.S. at 57-58, the touchstone of the third Gingles precondition is whether the majority votes sufficiently as a bloc to enable it to defeat the minority's preferred candidate most of the time. See

54

id. at 56; Sanchez, 97 F.3d at 1319. Crossover voting in South Providence enabled African-American voters regularly to elect the candidate of their choice despite the relatively small African- American constituency in the predecessor district. This is a telling bit of political history. See Gingles, 478 U.S. at 56 (noting that the amount of nonminority bloc voting that is legally significant varies in part with the size of the minority group within the district); Rangel v. Morales, 8 F.3d 242, 245 (5th Cir. 1993) (same). It demonstrates to my satisfaction that no legally cognizable anti-minority bloc voting exists here (and that, therefore, the appellants have failed to meet the third Gingles precondition).

I give the majority its due. In the ordinary course, district courts should allow colorable vote dilution claims to proceed beyond the Rule 12(b)(6) stage. And, moreover, if one is willing to split an infinite number of hairs, it always will be possible to conjure up remote scenarios that might be disinterred during discovery (and, thus, prevent the entry of a motion to dismiss). But Rule 12(b)(6) does not invite courts to engage in such endless conjecture. See Garrett v. Tandy Corp., 295 F.3d 94, 105 (1st Cir. 2002) ("The method of Rule 12(b)(6) requires courts . . . to resolve all realistic possibilities in the pleader's favor." (emphasis supplied)). Some cases are sufficiently clear that, on any rational view of the facts alleged, a vote dilution

55

claim is insupportable. See Mixson v. Ohio, 193 F.3d 389, 399-400, 406-08 (6th Cir. 1999) (affirming dismissal of section 2 claim under Rule 12(b)(6)); Mirrione v. Anderson, 717 F.2d 743, 746 (2d Cir. 1983) (similar); Martinez v. Bush, 234 F. Supp. 2d 1275, 1280 n.7 (S.D. Fla. 2002) (three-judge court) (per curiam) (granting defendants' Rule 12(b)(6) motion with respect to a section 2 claim). This is such a case: the allegations are unapologetic, the key facts are essentially undisputed, and the amended complaint stands or falls on the cogency of the appellants' avant-garde legal theory. Like the district court, I find that theory unacceptable.

I add a coda. Reapportionment and redistricting are thorny matters — and matters in which state legislatures are best suited to lead. Within wide limits, courts ought to respect legislative choices. See Voinovich, 507 U.S. at 156-57 (collecting cases). I understand that respect is not equivalent to blind allegiance, and if there were signs that the Rhode Island General Assembly had acted in derogation of the Constitution or federal law, I would not hesitate to support judicial intervention. But such signs are lacking here, so respect counsels restraint.

Given the mixed racial and ethnic composition of South Providence, the Rhode Island General Assembly was caught between a rock and a hard place. It made a series of difficult choices, not perfectly, but within the compass of its legal and constitutional authority. Whether or not I would have drawn the lines of the

56

affected district in the same manner is beside the point. What matters is that the General Assembly's line-drawing is a product of legitimate legislative choices made within allowable limits. Accepting the appellants' vote dilution claim would nullify these choices and give an unfair advantage to a particular subset of voters — an advantage beyond any that Congress contemplated in drafting the VRA. In the bargain, accepting the claim would shrink the district-wide Hispanic population, thereby disadvantaging another group of minority voters.

I have said my piece. Because the appellants fail to allege the kind of impermissibly race-based distortion of electoral opportunity that would sustain a claim under section 2 of the VRA, I respectfully dissent.