Meza v. Galvin

322 F.Supp.2d 52 (2004)

Angel MEZA, et al., Plaintiffs,
v.
William Francis GALVIN, in his Official Capacity as Secretary of the Commonwealth of Massachusetts, Defendant.

No. CIV.A.02-10428.

United States District Court, D. Massachusetts.

June 21, 2004.

Mary L. Cataudella, Robinson & Cole, Boston, MA, for William Francis Galvin, Defendant.

Nadine M. Cohen, Lawyers' Committee for Civil Rights Under Law, Boston, MA, for Chelsea's Commission on Hispanic Affairs, Inc., OISTE, Angel Meza, Francisco Dominguez Gabriel Valerio, Juan Vega, Karen Pazos, Magaly Camacho, Plaintiffs.

Lawrence S. DiCara, Nixon Peabody, LLP, Boston, MA, for William Francis Galvin, Jane Swift, Thomas F. Birmingham, Thomas M. Finneran, Defendants.

Michael D. Lurie, Robinson & Cole, LLP, Boston, MA, for William Francis Galvin, Defendant.

Steven P. Perlmutter, Robinson & Cole LLP., Boston, MA, for William Francis Galvin, Defendant.

Rudolph F. Pierce, Goulston & Storrs, PC, Boston, MA, for Chelsea's Commission on Hispanic Affairs, Inc., OISTE, Angel Meza, Francisco Dominguez, Gabriel Valerio, Juan Vega, Karen Pazos, Magaly Camacho, Plaintiffs.

Elizabeth C. Sackett, Robinson & Cole, Boston, MA, for William Francis Galvin, Defendant.

Peter Sacks, Attorney General's Office, Boston, MA, for William Francis Galvin, Jane Swift, Thomas F. Birmingham, Thomas M. Finneran, Defendants.

Kenneth A. Sansone, Goulston & Storrs, PC, Boston, MA, for Chelsea's Commission on Hispanic Affairs, Inc., OISTE, Angel Meza, Francisco Dominguez, Gabriel Valerio, Juan Vega, Karen Pazos, Magaly Camacho, Plaintiffs.

Brenda Wright, National Voting Rights Institute, Boston, MA, for Chelsea's Commission on Hispanic Affairs, Inc., OISTE, Angel Meza, Francisco Dominguez, Gabriel Valerio, Juan Vega, Karen Pazos, Magaly Camacho, Plaintiffs.

Before SELYA,^[*] Circuit Judge, WOODLOCK and PONSOR, District Judges.

MEMORANDUM AND ORDER

WOODLOCK, District Judge.

This is the companion case to Black Political Task Force v. Galvin, Civil Action No. 02-11190, in which we upheld a challenge by African-American and Hispanic voters to legislation revising the boundaries of the seventeen Massachusetts House of Representatives districts within the city of Boston. Having accepted the revised remedy proposed by the defendant in response to our liability opinion in Black Political Task Force, 300 F.Supp.2d 291 (D.Mass.2004), we have today separately entered final judgment in that case. In the instant case, we initially entered a bare final judgment contemporaneously with issuance of our liability opinion in Black Political Task Force, effectively rejecting a focused challenge to the redistricting of one specific House district within Boston. That judgment was intended to apprise the parties, for purposes of constructing a remedy in the broader Black Political Task Force case, that we did not view the single district at issue here to be problematic as enacted. Given the extended remedial proceedings in Black Political Task Force, however, we later vacated the final judgment in this case in order that the final judgments in both cases and the completed rationale for those judgments would be available simultaneously. This Memorandum will provide an explanation for the renewed final judgment entered in this case today.

The claims in the instant case, brought by three Latino ^[1] voters and two nonprofit organizations, were confined to the redrawing of the lines for the 2d Suffolk House district("2d Suffolk District"),^[2] which under the 2001 Redistricting Act (the "Enacted Plan") included all of the Charlestown neighborhood of the city of Boston and most of the city of Chelsea. Plaintiffs contended that the legislature, in drawing the lines of the 2d Suffolk District for the Enacted Plan, impermissibly diluted the voting strength of the Hispanic community in violation of § 2 of the Voting Rights Act, 42 U.S.C. § 1973(b). Plaintiffs further contended that the legislature intentionally carried out this alleged vote dilution scheme in violation of the Equal Protection Clause of the 14th Amendment. After trial, we found plaintiffs unable to sustain either claim as to the 2d Suffolk District under the Enacted Plan. The remedy we have approved in Black Political Task Force employs the same configuration for the 2d Suffolk District as in the Enacted Plan.

I. BACKGROUND

Although Black Political Task Force and this case were filed separately and the two cases involved distinct legal claims by different collections of plaintiffs, many of the factual and legal issues presented in the cases overlapped. Consequently, we consolidated the cases for trial.

To avoid unnecessary redundancy, we use our Black Political Task Force liability opinion as the template for our analytical approach here. In doing so, we assume familiarity with that opinion and will direct our analysis more narrowly to the aspects of the 2001 redistricting process that concern alleged Hispanic vote dilution in the 2d Suffolk District.

A. The Enacted Plan for the 2d Suffolk District

Prior to the 2001 redistricting, the 2d Suffolk District as drawn by the previous redistricting plan in 1993 comprised all of Charlestown and all but two precincts in Chelsea. The record indicates this district configuration linking Charlestown and Chelsea had been in place for at least thirty years. According to the 2000 Census, the 2d Suffolk District, with the 1993 lines, had a total population of 42,790. This exceeded the target district population for the 2001 redistricting effort,^[3] necessitating a reduction of population in the district. The Enacted Plan, which the legislative redistricting committee (the "Committee") ultimately adopted, did not materially change the configuration of the 2d Suffolk District; it maintained the historic pairing of Charlestown and Chelsea in the district but moved four precincts in the northern part of Chelsea (precincts 3-2, 3-4, 4-2, and 4-3) into the adjoining 16th Suffolk District to bring the 2d Suffolk District within the target population range.^[4]

B. The Plaintiffs' Plan

Plaintiffs' primary contention was that by again configuring the 2d Suffolk District to include Charlestown, the Committee countenanced a split of the rapidly growing Hispanic populations in Chelsea and East Boston and thereby impermissibly diluted the Hispanic vote in those areas.

In support of this contention, plaintiffs pointed to the 2000 Census data which show considerable growth in the Hispanic communities of Chelsea and East Boston from 1990 to 2000: during those ten years, the percentage of Hispanics in the total population increased about 150% (from 31.4% to 48.4%) in Chelsea and slightly more than 200% (from 17.6% to 39.0%) in East Boston. The Hispanic population in Charlestown increased at an even greater rate during that time period, an over 500% change in total population, from 2.1% to 11.6%. The absolute number of Hispanics in Charlestown counted in the 2000 Census (1,764), however, remained significantly below the number of Hispanics in Chelsea (16,984) or in East Boston (14,990).

During the 2001 redistricting process and prior to the Committee's adoption of the Enacted Plan, the Chelsea Commission on Hispanic Affairs ("CCHA"), one of the plaintiff organizations here, submitted two separate plans to the Committee in an attempt to contain the entire city of Chelsea in a single district. One of these plans did so by adding one Charlestown precinct to all the Chelsea precincts, while the other added one East Boston precinct (precinct 1-8) to all the Chelsea precincts.^[5]

Plaintiffs did not offer either of the CCHA pre-enactment plans in this litigation as a means to rectify the alleged dilution. Rather, for purposes of this case, plaintiffs offered a different alternative redistricting plan (the "Plaintiffs' Plan") as a remedial plan.^[6] The Plaintiffs' Plan moved five East Boston precincts (precincts 1-4, 1-6, 1-7, 1-8, and 1-9) from the 1st Suffolk District to the 2d Suffolk District and shifted all of Charlestown into the adjacent 3d Suffolk District. Additionally, the Plaintiffs' Plan moved three Chelsea precincts out of the 2d Suffolk District (precincts 2-1 (to the 28th Middlesex District), 4-1 (to the 16th Suffolk District), and 4-4 (to the 16th Suffolk District)) and brought in two Chelsea precincts from the 16th Suffolk District (precincts 3-2 and 3-4) to replace them.

Plaintiffs contended that the end result of this alternative district realignment would not merely be the unification of a Hispanic population split in the Enacted Plan; it would also comport with the traditional redistricting principle of maintaining "communities of interest" because Chelsea and East Boston form such a community whereas Chelsea and Charlestown are substantially dissimilar.

In short, plaintiffs argued that in configuring the 2d Suffolk District, the Committee disregarded what the Census data and simple facts about Charlestown, East Boston, and Chelsea make patently clear — that East Boston, not Charlestown, now belongs with Chelsea in the 2d Suffolk District. Plaintiffs further contended that the failure to combine East Boston and Chelsea was not due to simple oversight on the part of the Committee; rather, they claimed that the obviousness of the fact that the Enacted Plan divided a substantial Hispanic population, in addition to separating an ethnic community of interest, gives rise to an inference of intentional discrimination on the part of the Committee.

II. ANALYSIS

We assess plaintiffs' allegations of a Voting Rights Act violation following the approach we set forth for a § 2 claim in our Black Political Task Force liability opinion. In so doing, we first address whether plaintiffs have satisfied the three formal threshold elements of a § 2 violation, as dictated by Thornburg v. Gingles, 478 U.S. 30, 50, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986). Although we conclude plaintiffs have arguably failed to meet the first Gingles requirement and appear to fall short on the third requirement as well, given uncertainty about the precise dimensions of the Gingles threshold elements, we continue beyond the Gingles analysis to undertake a "totality of the circumstances" analysis. Ultimately, our analysis leads us to conclude that the configuration of the 2d Suffolk District in the Enacted Plan does not violate § 2. Because, unlike in the companion Black Political Task Force case, we ultimately find no § 2 violation here, we then continue beyond the § 2 analysis to assess plaintiffs' Equal Protection claims. In the end, we find that the constitutional route is also unsuccessful for plaintiffs, and we conclude that there was no impermissible dilution of Hispanic voters' rights in the redistricting of the 2d Suffolk District.

A. § 2 of the Voting Rights Act

To prevail on a § 2 claim, plaintiffs need not show discriminatory purpose; rather, they must first meet the three threshold Gingles conditions: (1) that they are a part of a minority group that is "sufficiently large and geographically compact to constitute a majority in a single-member district"; (2) that the plaintiff minority group is "politically cohesive"; and (3) that "the white majority votes sufficiently as a bloc to enable it — in the absence of special circumstances, such as the minority candidate running unopposed — usually to defeat the minority's preferred candidate." Gingles, 478 U.S. at 50-51, 106 S.Ct. 2752 (citation omitted). If the Gingles threshold is surmounted, plaintiffs must then demonstrate, given the "totality of the circumstances," that the Enacted Plan deprives Hispanic voters of an equal opportunity to participate in the electoral process and to elect candidates of their choice. See Vecinos De Barrio Uno v. City of Holyoke, 72 F.3d 973, 980 (1st Cir.1995) ("Vecinos III"). We address the Gingles factors and undertake a totality of circumstances analysis in turn.

1. The Gingles Threshold Requirements

Because the Gingles threshold requirements are arguably dispositive in this case, we pause to consider their purpose. As we noted in Black Political Task Force, "[v]oting rights cases are among the most difficult a court must decide. Not only do they implicate the complex relationship between race and politics, but they also plunge courts into the uncomfortable worlds of statistical analysis and legislative policymaking." 300 F.Supp.2d at 315. Given these difficulties, it is appropriate to require certain threshold showings before the courts are invited to enter those worlds. Moreover, for such thresholds to be meaningful, they must have some clarity and prescriptiveness; to the degree that the thresholds are as open-textured and contingent as the "totality of the circumstances" balancing test, they effectively collapse into the balancing test itself. In other words, the Gingles preconditions appear meant to act as a threshold gateway to the totality of the circumstances analysis, rather than as factors within it. See Campos v. City of Houston, 113 F.3d 544, 548 (5th Cir.1997). Thus, in analyzing the Gingles requirements, we note there are policy arguments for observing their plain terms strictly and reasons why the failure to meet any one of them should be found fatal to plaintiffs' case. See generally Rodriguez v. Pataki, 308 F.Supp.2d 346, 383-86 (S.D.N.Y.2004) (three-judge panel) (noting policy arguments for objective rule to discipline courts' discretion under § 2); cf. Vieth v. Jubelirer, ___ U.S. ___, 124 S.Ct. 1769, 158 L.Ed.2d 546 (2004) (concluding that political gerrymandering cases are nonjusticiable because no judicially discernible and manageable standards for adjudicating such claims now exist).

(a) "Majority"

The first Gingles threshold requirement mandates a showing that plaintiffs are members of a minority group "sufficiently large and geographically compact to constitute a majority in a single-member district." Gingles, 478 U.S. at 50, 106 S.Ct. 2752. Plaintiffs claimed their alternative plan did so and is a feasible and adequate remedy for the alleged Hispanic vote dilution resulting from the Enacted Plan. Whether it does, in fact, do so turns on how the term "majority" is defined.^[7]

Plaintiffs' expert, John Harmon,^[8] testified that while under the Enacted Plan the 2d Suffolk District has a Hispanic voting age population ("VAP") of 32.3%, VAP for Hispanics under the Plaintiffs' Plan rose to a majority, albeit slim, of 50.7%. This, plaintiffs claimed, is sufficient to demonstrate the potential to elect as required by Gingles, even under the stringent bright-line rule adopted by most courts requiring proof of a statistical majority. See, e.g., Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 852-53 (5th Cir.1999), cert. denied, 528 U.S. 1114, 120 S.Ct. 931, 145 L.Ed.2d 811 (2000); Cousin v. Sundquist, 145 F.3d 818, 828-29 (6th Cir.1998), cert. denied, 525 U.S. 1138, 119 S.Ct. 1026, 143 L.Ed.2d 37 (1999).

Defendant, however, contended that VAP is not the appropriate measure for determining whether plaintiffs meet the first Gingles precondition. Rather, defendant argued that where data concerning citizen voting age population ("CVAP") are available and differ significantly from VAP data, CVAP is the appropriate measure. Defendant's expert, Kimball Brace,^[9] calculated that the CVAP of Hispanics in the 2d Suffolk District under the Enacted Plan and the Plaintiffs' Plan was 19.3% and 31.85%, respectively. Defendant argued that with such low Hispanic CVAP numbers for the 2d Suffolk District under their alternative plan, plaintiffs could not sufficiently demonstrate an "effective voting majority," Johnson v. De Grandy, 512 U.S. 997, 1024, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994), and thus could not clear the first Gingles hurdle.^[10]

We are therefore faced with two legal questions, yet unresolved in this circuit, that we deferred in our companion Black Political Task Force liability opinion because they were not dispositive there: The first is whether the proper statistical measure for evaluating plaintiffs' ability to meet the first Gingles precondition is VAP or CVAP. The second, if we determine that the latter is the appropriate measuring stick, which we do, is whether plaintiffs can satisfy the first Gingles precondition despite their inability to demonstrate a statistical majority of Hispanic CVAP in the 2d Suffolk District. Because this second question has been carefully avoided by the First Circuit en banc since our Black Political Task Force opinion entered, see Metts v. Murphy, 363 F.3d 8 (1st Cir.2004) (en banc) (per curiam), and because there exists an alternative "totality of the circumstances" ground to dispose of the § 2 claim, we decline to offer a definitive answer to the second question in this case.

(i) CVAP is the Proper Measure When Available

At the outset, we note that neither the Supreme Court nor the First Circuit has defined with precision the kind of "majority" a minority group must be able to demonstrate to satisfy Gingles. Cf. De Grandy, 512 U.S. at 1008, 114 S.Ct. 2647 (noting, without reaching the issue, that "which characteristic of minority populations (e.g., age, citizenship) ought to be the touchstone for proving a dilution claim and devising a sound remedy" is a question of law); Black Political Task Force, 300 F.Supp.2d at 300 ("[T]he First Circuit has not spoken directly to the point."). However, all the circuits, and a number of district courts, that have directly reached the issue have held that citizenship voting age population data, where readily available, should be used in the Gingles analysis. See Chen v. City of Houston, 206 F.3d 502, 515 n. 7 (5th Cir.2000) ("Citizen voting age population is the proper measure under section 2."), cert. denied, 532 U.S. 1046, 121 S.Ct. 2020, 149 L.Ed.2d 1017 (2001); Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368, 372 (5th Cir.1999) (same), cert. denied, 528 U.S. 1114, 120 S.Ct. 930, 145 L.Ed.2d 810 (2000); Negron v. City of Miami Beach, 113 F.3d 1563, 1571 (11th Cir.1997) (first Gingles precondition requires consideration of citizenship information); Cano v. Davis, 211 F.Supp.2d 1208, 1233 (C.D.Cal.2002) (three-judge panel) (CVAP is "the appropriate measure to use in determining whether an additional effective majority-minority district can be created"), aff'd, 537 U.S. 1100, 123 S.Ct. 851, 154 L.Ed.2d 768 (2003); see also Barnett v. City of Chicago, 141 F.3d 699, 704 (7th Cir.1998) (stating, in dictum, that CVAP is proper measure in § 2 suit alleging dilution of Latino voting power), cert. denied, 524 U.S. 954, 118 S.Ct. 2372, 141 L.Ed.2d 740 (1998); cf. Romero v. City of Pomona, 883 F.2d 1418, 1425-26 (9th Cir.1989) (counting only eligible voters), overruled in part on other grounds by Townsend v. Holman Consulting Corp., 914 F.2d 1136, 1141 (9th Cir.1990) (en banc); Aldasoro v. Kennerson, 922 F.Supp. 339, 373 (S.D.Cal.1995) (finding case law stands "for the proposition that at least an eligible voter majority is required"). But see Rodriguez v. Pataki, 308 F.Supp.2d at 405-06 (assuming arguendo"that an ethnic minority group constituting a majority (50.1%) of the VAP and a near majority of the CVAP (45.4%) in a single-member district may under some circumstances satisfy the first Gingles precondition").

As we noted in our Black Political Task Force liability opinion, using citizen information fits with the underlying purpose of the first Gingles precondition, which is to determine voter ability to elect preferred candidates. 300 F.Supp.2d at 300. Indeed, the Voting Rights Act is specifically addressed to the mechanics of voting. See Gingles, 478 U.S. at 88, 106 S.Ct. 2752 (O'Connor, J., concurring); see also Holder v. Hall, 512 U.S. 874, 887, 114 S.Ct. 2581, 129 L.Ed.2d 687 (1994) (O'Connor, J., concurring) ("In order for an electoral system to dilute a minority group's voting power, there must be an alternative system that would provide greater electoral opportunity to minority voters." (emphasis added)); Voinovich v. Quilter, 507 U.S. 146, 152, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993) ("Congress enacted § 2 of the Voting Rights Act of 1965 ... to help effectuate the Fifteenth Amendment's guarantee that no citizen's right to vote shall `be denied or abridged ... on account of race, color, or previous condition of servitude.'" (emphasis added)).

Because non-citizens by definition cannot vote, it makes little sense to consider them for the purposes of determining whether the particular remedial scheme proffered by plaintiffs would adequately remedy the alleged vote dilution. This is especially so where, as here, the discrepancy between VAP and CVAP is substantial. As Judge Posner explained in Barnett v. City of Chicago:

Neither the census nor any other policy or practice suggests that Congress wants noncitizens to participate in the electoral system as fully as the concept of virtual representation would allow.... The right to vote is one of the badges of citizenship. The dignity and very concept of citizenship are diluted if noncitizens are allowed to vote either directly or by the conferral of additional voting power on citizens believed to have a community of interest with the noncitizens.

141 F.3d at 704.

It is appropriate for analytical purposes to distinguish the goal of ensuring electoral equality through the maintenance of fair voting practices and procedures, on the one hand, from a more general concern with representational equality, on the other. The latter, which ensures that constituents have equal access to their representatives and to governmental services, has been the conventional consideration in devising an apportionment system to meet the one-person, one-vote obligation. Indeed, the Supreme Court has indicated that apportionment systems based on total population, without regard to voter eligibility, are constitutional. See Burns v. Richardson, 384 U.S. 73, 92, 86 S.Ct. 1286, 16 L.Ed.2d 376 (1966) ("The decision to include or exclude [those ineligible to vote from the apportionment base] involves choices about the nature of representation with which we have been shown no constitutionally founded reason to interfere."); see also Garza v. County of Los Angeles, 918 F.2d 763 (9th Cir.1990) (upholding remedial reapportionment plan based on total population and rejecting appellant's argument that apportionment should rely on CVAP), cert. denied, 498 U.S. 1028, 111 S.Ct. 681, 112 L.Ed.2d 673 (1991).

The Voting Rights Act, in contrast, is not aimed at ensuring equality of representation. Rather, its aim is to safeguard voters' opportunities to elect their preferred representatives and as such it promotes electoral, not representational, equality.^[11] This underlying purpose of the Voting Rights Act translates here into the use of CVAP rather than VAP because the former more accurately captures the group of individuals whose electoral rights are at stake and allows for a more appropriate remedial recalibration of improperly-balanced voting strength. Thus, we conclude that CVAP is the appropriate measure here in assessing plaintiffs' ability to meet the first Gingles precondition.

We must consequently turn to the second issue: whether Gingles requires a showing that under the Plaintiffs' Plan, Hispanic CVAP constitutes a statistical majority in the 2d Suffolk District. But before we explore that legal issue, we digress briefly to resolve a factual dispute concerning the appropriate CVAP for Hispanics in the 2d Suffolk District.

(ii) CVAP Under the Plaintiffs' Plan is Less than a Statistical Majority

As noted above, defendant's expert, Brace, testified that the Hispanic CVAP under the Plaintiffs' Plan is 31.85%. Brace, however, did not simply take this figure directly from the Census data; rather, the figure is an estimation based on a series of calculations that break down and rearrange citizenship data he received from the Census Bureau. Moreover, the raw citizenship data itself came from the Bureau's "Long Form," which unlike the "Short Form," is not distributed to every household but rather is collected from an average of one in six households. Thus, as an initial critique, plaintiffs' expert, Harmon, questioned the validity of Brace's Hispanic CVAP estimate given that the raw citizenship numbers on which Brace performed his calculations are subject to sampling error.

While Brace's best estimate for Hispanic CVAP in the Plaintiffs' Plan 2d Suffolk District was 31.85%, he also provided a range of estimates that incorporated the Census Bureau's standard error calculations. For instance, in his testimony, Brace reported three "confidence intervals" to the data: the 90% confidence interval, the 95% confidence interval, and the 95% confidence interval with the design factor. The design factor confidence interval specifically took into account the Census Bureau's calculation of the likelihood of any error caused by the Long Form sampling method. We conclude that this interval, which Brace reported to be between 29.30% and 34.40%, sufficiently accounts for deviations caused by sampling error and thus is a reliable range of estimates for Hispanic CVAP in the 2d Suffolk District under the Plaintiffs' Plan.^[12]

Plaintiffs' criticism of Brace's Hispanic CVAP estimates on methodological and substantive grounds is not compelling. Using an alternate method for calculating Hispanic CVAP in the 2d Suffolk District in the Plaintiffs' Plan, Harmon undertook to expand the range of Hispanic CVAP for the Plaintiffs' Plan 2d Suffolk District from 25.93% to 45.23% with a midpoint of 35.58%. We do not find Harmon's methodology persuasive. Harmon did not present any reasoned basis for his alternative. His inadequately-explained methodological preferences appear to be simply an attempt to stretch Brace's CVAP figures statistically, and we therefore conclude that the range Harmon reported for Hispanic CVAP in the 2d Suffolk District is artificially wide. Brace's range of 29.30% to 34.40% is a more credible estimate of Hispanic CVAP for the 2d Suffolk District under the Plaintiffs' Plan. It bears emphasizing as a substantive matter that even under Harmon's strained analysis, the highest point in the range for the Hispanic CVAP for the 2d Suffolk District falls nearly five percentage points below a statistical majority and the midpoint is nearly fifteen percentage points below a statistical majority.

(iii) Plaintiffs' Failure to Show a Statistical Majority Hispanic CVAP in Their Proposed 2d Suffolk District is Arguably Fatal to Their § 2 Claim

Turning to the specific legal issue concerning the prescriptiveness of a statistical majority, we again confront the question we were able to avoid in Black Political Task Force: whether a bright-line mathematical rule should be used to determine what specific level of CVAP is needed to satisfy the first precondition of Gingles. As we noted in Black Political Task Force, the majority of courts that have specifically reached the question have adopted an absolute rule requiring plaintiffs to demonstrate that under an alternative plan the population of the minority group bringing the litigation comprises a statistical majority of the total relevant population in the affected district. ^[13] 300 F.Supp.2d at 299; see, e.g., Valdespino, 168 F.3d at 852-53; Cousin, 145 F.3d at 828-29; McNeil v. Springfield Park Dist., 851 F.2d 937, 943-45 (7th Cir.1988), cert. denied, 490 U.S. 1031, 109 S.Ct. 1769, 104 L.Ed.2d 204 (1989); Parker v. Ohio, 263 F.Supp.2d 1100, 1104-05 (S.D.Ohio 2003) (three-judge panel), aff'd, ___ U.S. ___, 124 S.Ct. 574, 157 L.Ed.2d 426 (2003). Such a rule, applied here, would necessarily end the § 2 inquiry because, even were we to accept Harmon's inflated CVAP numbers, plaintiffs are unable to demonstrate a statistical majority for Hispanics in the 2d Suffolk District under their alternative plan.

We recognize that the Supreme Court has never endorsed a bright-line statistical majority rule and has explicitly left open the issue of whether so-called "crossover districts," in which minority groups constitute under 50% of the relevant population in the proposed district but with the help of non-minority crossover votes have the ability to elect preferred officials, are viable alternatives under the first prong of Gingles.^[14]See, e.g., De Grandy, 512 U.S. at 1008-09, 114 S.Ct. 2647; Voinovich, 507 U.S. at 154, 113 S.Ct. 1149; Growe v. Emison, 507 U.S. 25, 41 n. 5, 113 S.Ct. 1075, 122 L.Ed.2d 388 (1993).

The First Circuit, noting these Supreme Court reservations, has recently declined to foreclose entirely the possibility that districts with less than a 50% majority can meet the first Gingles hurdle. See Metts, 363 F.3d at 11-12. Metts provides some superficial support for the proposition that plaintiffs — with their 29.30% to 34.40% CVAP figures — could actually satisfy the first Gingles requirement. A closer look at Metts, however, reveals important differences which distinguish that case from this. The majority opinion in Metts makes clear that the decision to allow the plaintiffs to proceed rested heavily on two factors not present here. For one thing, Metts was at an embryonic stage (it had been decided in the district court on a motion to dismiss, and no discovery had occurred). Id. at 9. For another thing, the legislative district at issue had "a relatively unusual history," that is, a history of African Americans being able to elect their preferred candidate even though they numbered only 26% of the relevant population. Id. at 11-12. In contradistinction to Metts, discovery and trial have run their course in the case at hand and, therefore, the parties have had a full and fair opportunity to present the complete picture of voting in Chelsea, Charlestown, and East Boston. That evidence does not suggest that the affected communities have a comparable history of sending Hispanic-preferred candidates to the House of Representatives.

We were not required to resolve the majority definition issue in Black Political Task Force because it was not outcome-determinative there. 300 F.Supp.2d at 299. It could be outcome-determinative here, but given the existence of alternative grounds for resolution, we decline to decide whether plaintiffs' proposed district fails to meet the first Gingles precondition because it does not show a statistical CVAP majority for Hispanics.^[15] Having in mind the lack of prescriptive precedent for this court on this issue and recognizing the myriad possible facts that may underlie a voting rights case, we hesitate to ground our decision on the conclusion that the Plaintiffs' Plan does not meet a required bright-line CVAP threshold. Accordingly, while we find the lack of a statistical Hispanic CVAP majority under the Plaintiffs' Plan is arguably fatal to plaintiffs' § 2 claim because we conclude it could constitute an inability to meet the first Gingles requirement, we nevertheless continue with the Gingles inquiry and will take up a full totality of the circumstances analysis to provide a resolution of the case on the ultimate grounds raised.

(b) Ethnic Voting Polarization ^[16]: Cohesiveness and Bloc Voting

As we did in Black Political Task Force, we will treat the second Gingles precondition (whether Hispanics in the 2d Suffolk District under the Plaintiffs' Plan are likely to be politically cohesive) and the third Gingles precondition (whether whites vote sufficiently as a bloc to defeat Hispanic preferred candidates most of the time) together in examining the evidence. 300 F.Supp.2d at 302-10. We do so while keeping in mind the distinctive purposes of the two preconditions.

Of the elections fully analyzed by the parties' experts to determine the extent of ethnically-polarized voting in the 2d Suffolk District, only one, the 1996 Democratic primary for the 2d Suffolk District House seat, was endogenous. ^[17] We find its usefulness for our analysis to be limited at best. This is in part because it was not a multi-ethnic race and in part because neither expert could reliably determine Hispanic voters' candidate of choice in the election.

The mainstay of evidence concerning the ethnic voting polarization (or lack thereof) in this case therefore comes from the results of seven exogenous elections. Four of the elections were in the city of Chelsea, two were city council elections in Boston, and one was the Democratic primary for the Middlesex, Suffolk, and Essex state Senate district. Generally speaking, we glean from the evidence in the record an increasing political experience and sophistication by Hispanic voters and a slow movement over the past decade away from reflexive non-Hispanic bloc voting within the 2d Suffolk District. In order to consider the second and third Gingles requirements in light of such trends, we will address the identified elections in chronological order. Of these, we find the 2002 Senate primary to be most significant because it is the most inclusive; it is the only election in which all voters within the Enacted Plan 2d Suffolk District were presented with a choice between Hispanic and non-Hispanic candidates. We also find the recent 2001 and 2003 ^[18] Boston City Council elections especially significant for the insights they provide regarding voting in Charlestown and East Boston, which contain the precincts plaintiffs seek respectively to shed and to gain with their proposed 2d Suffolk District.

(i) Chelsea School Committee Elections: 1997, 1999, 2001

Latino-identified candidates (Valentin in 1997, Lopez ^[19] in 1999, and Valerio in 2001) received the most Hispanic votes in all three elections, each of which was an at-large election for seven seats. Additionally, the other Latino candidates in the elections placed high among Hispanic voters, finishing second (Lopez in the 1997 election) and third (Pazos in the 2001 election). While the Latino candidates were the Hispanic voters' preferred candidates, they were last or next-to-last in the non-Hispanic vote and they were unsuccessful. These elections on their face provide evidence of ethnic voting polarization by both Hispanic and non-Hispanic voters in Chelsea.

We note that the force of this evidence is diminished to some extent because the election results reveal low turnout rates for Hispanic voters in these elections. The reason for this low Hispanic turnout appears to relate to the demographics of the Hispanic community in Chelsea. Defendant's demographic expert, Peter A. Morrison, however, concluded that once differences in age composition and citizenship are accounted for — as they are by the use of CVAP figures — any disparities in turnover rates between Hispanic and non-Hispanic voters in Chelsea completely disappear.

In any event, low Hispanic turnout emphasizes that in at least two of the elections, there was significant support for the Latino candidates by non-Hispanic voters.^[20] We conclude that this was so even accepting plaintiffs' suggestion that the Latino candidates were preferred candidates of Hispanic voters, but not of non-Hispanic voters. Moreover, in the votes for all but one election,^[21] precinct-by-precinct analyses show that in three of the four precincts that were over 50% Hispanic in VAP during the elections, the Latino candidates did poorly, placing only as high as fourth and more typically sixth through tenth.^[22]

We think that these aspects of the evidence, however, speak more to the general issue of opportunity to elect, discussed in the totality of circumstances analysis below, and less to the issue of voter polarization. We acknowledge that low turnout could suggest, at least to some degree, lack of cohesion on the part of Hispanic voters, cf. Vecinos III, 72 F.3d at 987 (evidence of low turnout "may — or may not — be probative on the issue of minority cohesion"), and we do not deny that the percentages hardly paint a picture of Latino candidates supported exclusively by Hispanic voters. Low turnout rates, however, do not alter the conclusion which lies at the base of the second and third Gingles prongs: that Hispanic voters who turned out strongly supported the Latino candidates ^[23] and non-Hispanic voters, at least percentage-wise, did not. See id. at 986 ("[E]ven with a modest turnout, the actual votes cast may be probative of minority cohesion if a sufficiently strong pattern emerges.").

Additionally, we find defendant's precinct analysis indeterminate. While the poor showings by Latino candidates in precincts with high Hispanic VAP are indeed initially striking, even defendant admitted that whether they resulted from lack of Hispanic voter support or are merely another indication of low Hispanic turnout cannot be determined. We decline to speculate as to which is the better explanation, and thus the figures carry little probative weight. We conclude that the school committee election results provide evidence, though not without caveats, of ethnically-polarized voting by both Hispanic and non-Hispanics in Chelsea.

(ii) Chelsea City Council Election: 2001

The Chelsea city council election in 2001 was an at-large election for three seats, in which voters could cast up to three votes. One of the successful candidates in the election, Avellaneda, was Latino. Avellaneda was Hispanic voters' candidate of choice, receiving 97.5% of the Hispanic votes according to regression analysis. There was also a statistically significant positive correlation between the ethnic composition of the precincts and the votes for Avellaneda. Unlike in the school committee elections, however, evidence of ethnically-polarized voting by non-Hispanics is not especially strong.^[24] Even by the calculations of plaintiffs' expert, Richard Engstrom, Avellaneda received a significant number of non-Hispanic votes (28.1% by regression but as high as 42.6% under homogenous precinct analysis), which placed him third or fourth among non-Hispanic voters, depending on the analytical method used. These election results indicate cohesiveness on the part of Hispanic voters but are equivocal on the question of bloc voting by the majority.

(iii) Boston City Council At-Large Election: 2001

The 2001 at-large Boston City council election sheds some additional light on the voting preferences in Charlestown. A Latino, Arroyo — who may be considered the Hispanic preferred candidate — was one of seven candidates seeking four city council seats, and voters could cast up to four votes in the election. The parties' experts did not have much to say about this election. Engstrom reported that only 1.8% of total ballots went to Hispanic voters and that Arroyo received 14.9% of the total Charlestown votes, which placed him fifth among the seven candidates failing to win a seat on the City Council. ^[25] Defendant's expert, Harold Stanley, largely discounted the election results, stating only that given the significant support Barrios received in Charlestown in the 2002 Democratic primary, see infra § II.A.1.b.iv, Arroyo's low placement among voters should be taken merely as a reflection of the strength of his campaign and not generally as a reflection of Charlestown voters' nonpreference for Hispanic candidates. While we do not make so dismissive an assumption, we think it is appropriate to consider Arroyo's 2001 finish in light of Barrios's relative success in the area the following year as well as Arroyo's ultimate success in the 2003 City Council election, both of which weigh against taking the 2001 election results as unqualified evidence of non-Hispanic bloc voting in Charlestown.

More striking than Arroyo's performance among Charlestown voters, however, is his relative performance among voters in the five East Boston precincts plaintiffs seek to substitute for Charlestown in their proposed plan. Despite higher Hispanic populations in those East Boston precincts, Arroyo's performance there was poorer than in Charlestown. The percentage of votes he received in Charlestown was 14.9%; the percentage of votes he received in East Boston ranged from 5.4% to 7.6% in precincts 1-4 (5.4%), 1-6 (7.0%), 1-7 (7.6%), 1-8 (6.2%) and 1-9 (5.5%). While undoubtedly affected by low turnout resulting from low citizenship rates, these figures suggest the possibility that the electability of a Hispanic candidate may be negatively affected if East Boston is substituted for Charlestown in the 2d Suffolk District as plaintiffs proposed.

(iv) Middlesex, Suffolk and Essex Senate Primary for the Senate District Representing, inter alia, all of the 2d Suffolk District under the Enacted Plan and Most of the 2d Suffolk District Under the Plaintiffs' Plan.

Three candidates competed in this 2002 Senate primary, which included all of the 2d Suffolk District under the Enacted Plan and was a one-seat, one-vote election. One of the candidates was Latino (Barrios), and the other two were white (DeMaria and Galluccio). Barrios won the election by a slim 77-vote plurality. Barrios was the overwhelming candidate of choice of Hispanic voters in the district, receiving 95.1% of the Hispanic vote as estimated by regression analyses. The white voters' candidate of choice, however, was Galluccio (51% by regression), and thus, Barrios's win resulted at least partially from a split in the white vote between the white candidates. Together, the two white candidates received 54.8% of the total vote compared to the 43.7% Barrios received, and in Charlestown, Galluccio received more votes than Barrios (51.8% to 38.9%).

While Barrios was not the non-Hispanic voters' candidate of choice, he nevertheless received a significant amount of support from non-Hispanic voters in the election. Stanley calculated that non-Hispanic voters cast 1,550 of the 2,426 votes Barrios received in the election, and by Engstrom's regression estimates, Barrios received 39% of the non-Hispanic vote. Moreover, while Galluccio received more votes than Barrios in Charlestown, it appears that support for Barrios there was just as strong as in the district as a whole. Among the Chelsea precincts included in the Enacted Plan's 2d Suffolk District, Barrios was in fact the candidate of choice of white voters. Stanley concluded that this support of Barrios in Chelsea, "combined with his sizable non-Hispanic support in Charlestown, demonstrates that a Hispanic candidate has an opportunity to win election in the area covered by the 2d Suffolk House district."

We credit Stanley's testimony that the election results demonstrate a "sizeable" non-Hispanic support for Barrios, and we conclude that while the 2002 primary results demonstrate cohesiveness on the part of Hispanic voters in Chelsea, they seriously undercut plaintiffs' contention that the white majority will consistently vote as a bloc to defeat Hispanic voters' candidate of choice. As an initial point, Barrios, the Hispanic voters' candidate of choice, actually won the race. Even if we look beneath this overall result to recognize that Barrios was not the preferred candidate of non-Hispanic voters and give due consideration to the assumption that had Barrios faced only one white candidate, not two, he may have lost the primary, we are left with the fact that a substantial percentage of non-Hispanic voters supported Barrios in an election in which they had two non-Hispanic alternatives. Indeed, Barrios was white voters' preferred candidate in the Chelsea precincts that form the core of the 2d Suffolk District in the Plaintiffs' Plan. Thus, we conclude that the 2002 Democratic primary for the 2d Suffolk Senate district significantly undermines the conclusion that plaintiffs have satisfied the third Gingles precondition.

(v) Boston City Council At-Large Election: 2003

The 2003 Boston City Council election occurred too close to the time of trial to receive meaningful expert analysis. Nevertheless, the pattern evident in the 2001 City Council election remained essentially in place, except that Arroyo was elected city-wide to one of the four at-large seats. Arroyo received 10.5% of the ballots cast in Charlestown; this is roughly the average of the East Boston precincts (1-4 (9.72%), 1-6 (11.37%), 1-7 (13.55%), 1-8(10%) and 1-9 (10.11%)) which undercuts further plaintiffs' contention that a Hispanic-preferred candidate would gain an advantage by substituting East Boston for Charlestown in the 2d Suffolk District. Thus, our analysis of the 2003 election returns — admittedly preliminary given the parties' incompletely-developed evaluations of the election — suggests that while there is reason to believe there is some degree of ethnically-polarized voting among Hispanics and non-Hispanics in both Charlestown and East Boston, the evidence is not sufficiently substantial to provide much weight to a majority polarization calibration.

(vi) Conclusion

In our overall analysis of the seven elections considered here, we find plaintiffs have met the second Gingles condition by demonstrating that Hispanic voters typically vote in a cohesive fashion. By contrast, we find that the election results, viewed as a whole and as they have evolved in the past decade, do not demonstrate the level of ethnically-polarized non-Hispanic voting preferences needed for plaintiffs to satisfy the third Gingles precondition. The most recent available evidence showing significant support by non-Hispanic voters for Barrios in the Senate primary, and to a lesser degree for Avellaneda and Arroyo in their respective city council races, leads us to question the degree of non-Hispanic bloc voting and the extent to which it pervades the 2d Suffolk District either as enacted or in the Plaintiffs' Plan. In the end, we think that this support seriously belies the conclusion, required by Gingles, that the "majority votes sufficiently as a bloc to enable it — in the absence of special circumstances ... usually to defeat the minority's preferred candidate." Gingles, 478 U.S. at 51, 106 S.Ct. 2752. Thus, keeping in mind that "determining whether [ ] bloc voting exists is not merely an arithmetic exercise that consists of toting up columns of numbers, and nothing more," Vecinos III, 72 F.3d at 989, but rather must be a "commonsense assay of all the evidence," id., we find that plaintiffs have not sufficiently demonstrated bloc voting on the part of the majority.

Again, recognizing that our determination that plaintiffs have not met the third Gingles requirement is based on relatively scant electoral information about voting patterns across the entire 2d Suffolk District, we decline to rely simply on that conclusion to enter judgment for defendant. Accordingly, we turn to the totality of the circumstances analysis.^[26]

(2) Totality of the Circumstances

At the outset of our totality analysis, it is appropriate to step back and reflect on the gravamen of plaintiffs' § 2 claim. Plaintiffs contended that the 2d Suffolk District's historic association between the city of Chelsea and the Charlestown precincts of the city of Boston must be broken because of changing demographics. These demographics, they argued, demand a new association between Chelsea precincts and certain precincts in the East Boston section of Boston. However, efforts to obtain judicial intervention designed to form an entirely new electoral district by collecting minority communities from several different pre-existing districts have been viewed skeptically by the courts. See Goosby v. Town Bd. of Hempstead, 180 F.3d 476, 501 n. 3 (2d Cir.1999) (Leval, J., concurring) (terming argument that judicial invalidation and revision of long-established district lines "without any showing of discriminatory intent" as "alarming and far beyond the probable contemplation of Congress"), cert. denied, 528 U.S. 1138, 120 S.Ct. 982, 145 L.Ed.2d 932 (2000). And courts have long held that maximization is not the proper office of § 2. Rodriguez v. Pataki, 308 F.Supp.2d at 386 (observing disapprovingly that claim indicated "not the use of the Voting Rights Act to remedy a prohibited practice of vote dilution but rather a proposed use of section 2 to maximize minority voting strength" (citing De Grandy, 512 U.S. at 1017, 114 S.Ct. 2647)); see also Latino Political Action Comm. v. City of Boston, 784 F.2d 409, 412 (1st Cir.1986) (rejecting definition of vote dilution as "minimization, cancellation or submergence of minority voting strength below what might otherwise have been" (emphasis in original)). With these broad principles in mind, we turn to the particulars of a totality analysis.

We begin with defendant's assertions that the Enacted Plan and the Plaintiffs' Plan, when compared precinct by precinct, are not significantly different in ethnic composition. This is in many ways analogous to the proportionality assessment we undertook in the companion case, see Black Political Task Force, 300 F.Supp.2d at 311-12, except here, because the focus is only on one district, the 2d Suffolk District, the analysis is at the precinct, not district, level. Defendant pointed out that both the Enacted Plan and the Plaintiffs' Plan kept the most Hispanic precincts in the 2d Suffolk District and place four relatively non-Hispanic Chelsea precincts in the 16th Suffolk District. Defendant further argued that the four Chelsea precincts included in the 16th Suffolk District under the Enacted Plan were some of the least Hispanic precincts in Chelsea (ranking tenth, eleventh, fourteenth, and fifteenth in Hispanic VAP out of sixteen) and, perhaps more importantly, are not significantly more Hispanic than the four precincts included in the 16th Suffolk District under the Plaintiffs' Plan.

These observations concerning plaintiffs' tinkering with the Chelsea precincts add little to the information provided by the more general CVAP and VAP figures. Moreover, they gloss over the fundamental feature of the Plaintiffs' Plan, which is the replacement of Charlestown with precincts from East Boston. Much more germane to our purposes here, therefore, is defendant's contention that the Plaintiffs' Plan, by replacing Charlestown with East Boston, does not significantly increase Hispanic voters' ability to elect preferred candidates.

On this score, plaintiffs have provided little evidence either on the voting preferences and patterns of the East Boston precincts included in the 2d Suffolk District under the Plaintiffs' Plan or on the anticipated impact of replacing Charlestown precincts with East Boston precincts.^[27] Moreover, as we discussed in our analyses of the Boston City Council elections, evidence from two recent elections indicates that Arroyo received no greater support in the East Boston precincts included in the Plaintiffs' Plan 2d Suffolk District than he did in Charlestown. We are diffident about placing too much weight on the success (or lack thereof) of a single candidate in an isolated election, since it might reflect no more than the strength of the particular candidate in a particular campaign year. Nevertheless, we find the Arroyo evidence useful here because it shows a lack of meaningfully greater support for Arroyo in East Boston as compared to Charlestown and the relevant comparison can be made within a single election for two separate years.

We also find probative the evidence offered by defendant that the voter registration rate in East Boston is a low 11.3% and that the Hispanic CVAP in East Boston is only 14.2% of East Boston's CVAP, apparently the result of a large number of new Hispanic immigrants. This voter registration evidence is related to evidence concerning voter turnout to which we alluded in our polarization discussion, and we now take up the two types of evidence together.

While the turnout issue was of contextual importance to our voter polarization analysis, it has a substantial impact on the question of ability to elect. The fact that Hispanic voters register at a low rate in East Boston and turn out at low rates in Chelsea, at least for certain elections, significantly undercuts plaintiffs' contention that implementation of their proposed plan would substantially increase Hispanic voters' electoral opportunities. The low rates emphasize the concerns raised by plaintiffs' difficulties in satisfying the first Gingles precondition. Given that the Hispanic CVAP numbers for the Plaintiffs' Plan were submajority to begin with, the fact that only a portion of this CVAP population can be expected regularly to go to the polls further undermines plaintiffs' case. This is demonstrated in the Chelsea school committee elections in which the strong preferences of the Hispanic voters were, through low turnout and registration rates, diminished to the point that the Latino candidates could not rely on the Hispanic vote to carry them.

We recognize that low turnout rates can themselves be the result of official discrimination and that we must therefore be cautious in treating that circumstance as evidencing lack of force in plaintiffs' case. See Vecinos III, 72 F.3d at 986 ("[L]ow voter turnout in the minority community sometimes may result from the interaction of the electoral system with the effects of past discrimination, which together operate to discourage meaningful electoral participation."). But plaintiffs provided insufficient evidence that low turnout, at least recently, can be tied to discriminatory election practices. In fact, the record shows that Hispanics turned out in high rates to support Barrios in 2002 and Avellaneda in 2001. Thus, the pattern of low turnout in the Chelsea school committee elections does not seem to result from any current hindrance caused by official discrimination but rather from the interaction of demographic factors and the relatively lacking salience of school committee elections.

Plaintiffs sought to divert attention from the low registration in East Boston and turn it to a comparison of the ethnic demographics of Chelsea and Charlestown. They argued that the low registration in East Boston means only that the central effect of their proposed plan, at least in the short run, would be to cut out the influence of the heavily non-Hispanic Charlestown precincts and thereby maximize the influence of the heavily Hispanic Chelsea precincts. The East Boston precincts would, they seem to have conceded, act much like place holders until eventually the Hispanic community is able to mobilize Hispanic voters in East Boston.

This response fails in at least two respects. First, as noted above, plaintiffs have not provided any supportive electoral evidence as to the East Boston precincts, and thus their contention that inclusion of them in the 2d Suffolk District would minimize the negative presence of Charlestown and allow the Hispanic community in Chelsea to take center stage is based entirely on speculation; indeed, it is contradicted by the only evidence on this point, that Arroyo has not received more support in East Boston than he has in Charlestown.

More importantly, plaintiffs' response falters because it rests on an assumption about which we are skeptical: that there is dispositively significant ethnic polarization of voting preferences by non-Hispanics in Charlestown and Chelsea. As noted in the previous section, the support for Barrios and Arroyo in Charlestown indicates a lack of significant non-Hispanic bloc voting there, and the non-Hispanic support for both Avellaneda and Barrios in elections involving Chelsea precincts indicates the same in Chelsea.

With respect to evidence of "lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group," S.Rep. No. 97-417, at 29 (1982), reprinted in 1982 U.S.C.C.A.N. at 207, while plaintiffs offered evidence of specific, if anecdotal, instances of unresponsiveness by certain elected officials in past years, several of plaintiffs readily admitted that Representative O'Flaherty, the House member from the 2d Suffolk District since 1996 and a Chelsea resident since 1983, has, in fact, been responsive to the needs of the Hispanic community in the district. Indeed, the evidence is quite strong that O'Flaherty has been attuned to the needs of his Hispanic constituency.

We do not find evidence pertaining to the "extent to which members of the minority group have been elected to public office in the jurisdiction," id., to be compelling. Plaintiffs point to a number of unsuccessful campaigns by Latino candidates in Chelsea, including the school committee elections considered in our voter polarization analysis. Yet the success of the Barrios 2002 senate campaign in the relevant portions of the 2d Suffolk District suggests that attractive Hispanic candidates with well-run campaigns are currently quite competitive within the Enacted Plan's configuration of the 2d Suffolk District.

As in the Black Political Task Force case, plaintiffs here raised a number of concerns about the redistricting process itself, arguing that the procedures used by the Committee unfairly excluded participation by minorities generally and the Hispanic community in particular. We reiterate our conclusion in the companion opinion that these contentions are not persuasive as evidence of improper racial considerations. The record shows that the Committee held five public hearings and, more importantly, that the refusal to allow further public participation resulted from the Committee's policy to limit contact with the public to the hearings, which it applied without regard to race or ethnicity. Thus, the Committee's apparent disinclination to consider the two alternative plans presented by plaintiff CCHA during the redistricting process was consistent with its general policy and does not indicate that any of the Committee's procedures were discriminatory.

The issue of incumbency, on which we focused in Black Political Task Force, 300 F.Supp.2d at 313-15, takes on a different perspective in this case. In Black Political Task Force we found that the Committee, in textbook fashion, manipulated district lines to pack the 6th Suffolk District with the area's black population — thereby stripping that population from the 11th, 12th, and 15th Suffolk Districts — and, moreover, that the Committee performed this manipulation with an eye toward using race as a proxy in an effort to protect incumbents. Crucial to this finding was evidence that the Committee at the last minute adopted the so-called Fitzgerald Amendment in which it scrapped a plan to create a majority-black district without an incumbent in order to protect Representative Fitzgerald's reelection bid. Id. at 315. No such ethnically-related incumbency protection evidence can be found in this case. While the protection of incumbents no doubt played a central role in the redistricting of the 2d Suffolk District, we find no evidence that the Committee used ethnicity as a proxy for such protection as it did with race in redrawing the 6th Suffolk District (and those surrounding it).

Plaintiffs urged upon us the inference that the Committee impermissibly considered ethnicity in drawing the 2d Suffolk District because the Committee ignored an obvious opportunity to join communities of interest in Chelsea and East Boston and to unite a growing Hispanic community.^[28] The evidence, however, is not as obvious as plaintiffs would have it. Indeed, neither of the two alternative plans submitted to the Committee by plaintiff organization CCHA contemplated the inclusion of substantial parts of East Boston in the 2d Suffolk District. The idea of uniting the Hispanic communities of East Boston and Chelsea therefore seems to have been a post-enactment development devised by plaintiffs' experts for purposes of this litigation. This evolution in plaintiffs' contentions undercuts the argument that the Committee should have considered and implemented such an idea on its own and that its failure to do so gives rise to an inference of discriminatory animus.^[29]

Plaintiffs alternatively contended, setting aside East Boston, that we should infer discriminatory intent from the Committee's decision to exclude Chelsea precincts rather than Charlestown precincts from the 2d Suffolk District under the Enacted Plan. They claimed that there is no explanation for this decision other than discriminatory animus. We disagree. History sufficiently explains the Committee's decision. Prior to the current redistricting, certain of the northern precincts of Chelsea were in the 16th, not the 2d, Suffolk District while Charlestown had for many years been maintained intact in the 2d Suffolk District. The ultimate effect of the current redistricting process was to rearrange somewhat the northern Chelsea precincts that would be placed in the 16th Suffolk District in order to meet the reconstituted target populations. Thus, the evidence shows that a historical pattern, not ethnically-related considerations, explains the decision to keep Charlestown intact.

Up to this point, we have touched on the issue of discriminatory animus without directly addressing it. As we noted earlier, proof of discriminatory intent is not required for a § 2 claim. Moreover, if plaintiffs can sufficiently establish that redistricting impermissibly dilutes a minority group's voting strength, claims by defendant that the practice was adopted for neutral, nondiscriminatory reasons are unavailing. See S.Rep. No. 97-417, at 29 (1982), reprinted in 1982 U.S.C.C.A.N. at 207 n. 117; Solomon v. Liberty County Sch. Bd., 899 F.2d 1012, 1016 (11th Cir.1990), cert. denied, 498 U.S. 1023, 111 S.Ct. 670, 112 L.Ed.2d 663 (1991). This is not to say, however, that proof of animus has no place in assessing the totality of the circumstances; if plaintiffs can demonstrate discriminatory animus on the part of the Committee, that certainly would play into the totality of the circumstances. We take up the intent issue more directly in considering plaintiffs' Equal Protection claim and observe here only that we find no evidence of discriminatory intent on the part of the Committee.

We do not discount the history of racial polarization, particularly in Charlestown. We recognize the impact of socio-economic disparities upon the Hispanic community. And we do not minimize the significance of anecdotal evidence offered by plaintiffs of incidents of racism encountered by Hispanic candidates. But we do not find significantly probative evidence of any current official discrimination that touches upon Hispanic participation in the democratic process nor do we find any structural dimensions to the machinery of elections which hinders minority participation. We find that ethically-based prejudice is less a factor in the political life of the 2d Suffolk District as now enacted than it has been in the past. "Viewing the evidence as a whole, this court cannot say that the interaction of [ethnicity] and the electoral system [as it pertains to the 2d Suffolk District] now results in `significantly diminished opportunities for minority participation in elective government.'" Vecinos De Barrio Uno v. City of Holyoke, 960 F.Supp. 515, 527 (D.Mass.1997) ("Vecinos V") (quoting Vecinos III, 72 F.3d at 983-84). Our evaluation of the totality of the circumstances satisfies us that plaintiffs have not successfully made out a § 2 claim.

B. Constitutional Claims

We observe, at the outset of the constitutional branch of our analysis, that failure to establish a § 2 claim is generally considered mutatis mutandis fatal to constitutional claims because the latter, unlike the former, require proof of discriminatory intent. See Latino Political Action, 784 F.2d at 415 ("[T]o prevail on [their constitutional claims], appellants must satisfy a burden of proof equal to or greater than that already discussed."); see also Barnett v. City of Chicago, 141 F.3d at 701 ("[Section 2], unlike the constitutional provision, does not require any showing of intentional discrimination, and by the same token is not limited to impairments of minority voting power so egregious as to compel an inference of such discrimination." (citations omitted)). However, we hesitate to dispose mechanically of plaintiffs' constitutional claims because we think it may be possible, as an analytical matter, for them to prove the requisite intent for that claim despite their inability to make out a § 2 claim.^[30] But in the final analysis, we find that plaintiffs have failed to demonstrate the Committee intentionally discriminated against Hispanics in drawing the 2d Suffolk District.

We have already effectively addressed the bulk of plaintiffs' contentions concerning discriminatory intent. Thus, as noted above, plaintiffs' numerous submissions concerning the manner in which the Committee deflected an effort by Hispanic individuals and groups to participate in the redistricting process, including CCHA's effort to submit its two alternative plans (which ironically were designed not so much to add East Boston to the district as to keep Chelsea whole), do not suggest discriminatory animus on the part of the Committee but rather are a reflection of its policy to limit public participation to the five public hearings.

As also noted above, we find that an inference of discriminatory intent does not properly arise out of the Committee's failure to combine East Boston and Chelsea or to fragment Charlestown instead of Chelsea. Doing so might have further benefitted the Hispanic community by maximizing an ethnically-based community of interest in a fashion which would have future, if not immediate, electoral implications. But the failure to do so, absent some further showing of intent, does not here demonstrate a constitutional violation. Rather, the choice made in the Enacted Plan as to the 2d Suffolk District falls within the wide discretion of legislatures to decide how best to ensure equal voting opportunities. See Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 2511, 156 L.Ed.2d 428 (2003).

Finally, we offer some additional comments about the role of incumbency protection in this case. The law in this area is not a model of clarity. But even assuming that a cause of action exists for plaintiffs who can prove that a legislature used ethnicity as a tool to achieve the otherwise legitimate redistricting goal of incumbency protection, see generally, Clark v. Putnam County, 293 F.3d 1261, 1271-72 (11th Cir.2002); Barnett v. Daley, 32 F.3d 1196, 1199 (7th Cir.1994); Garza, 918 F.2d at 771; Ketchum v. Byrne, 740 F.2d 1398, 1408 (7th Cir.1984); Rybicki v. State Bd. of Elections, 574 F.Supp. 1082, 1109 (N.D.Ill.1982) (three-judge court), the plaintiffs here have failed to provide even a scintilla of evidence to support a finding that the Committee manipulated the Hispanic population data in an effort to secure victory for the incumbent in the 2d Suffolk District, Representative O'Flaherty. In the absence of credible evidence, an ipse dixit — that Representative O'Flaherty's success depended on the fracturing of the Hispanic population — is not enough to carry the day. This is especially true given the circumstances of this case, where the decision to pair Chelsea with Charlestown instead of East Boston can logically be attributed to a linkage that has existed for at least thirty years.

We therefore conclude that plaintiffs have failed sufficiently to demonstrate discriminatory intent on the part of the Committee in drawing the 2d Suffolk District lines, and we accordingly find for the defendant on the constitutional claims.

III. CONCLUSION

For the reasons set forth more fully above, we direct that a renewed final judgment be entered for the defendant.

NOTES

[*] Of the United States Court of Appeals for the First Circuit, sitting by designation.

[1] While the individual plaintiffs here are Latino and the majority of plaintiffs' briefs use the term "Latino," plaintiffs use the terms "Latino" and "Hispanic" interchangeably in many portions of their filings, especially when referencing demographic information. Any distinctions between the terms are of no significance in this case. For consistency and ease of discussion, we follow the Census Bureau's use of "Hispanic" (and "non-Hispanic") in its demographic statistics because those statistics form the crux of plaintiffs' claims. However, where, as is the case for the individual plaintiffs, the record specifically identifies an individual or individuals as "Latino" or "Latina" we will use those terms accordingly.

[2] In general, we use "Suffolk District" to refer to Suffolk House districts. Where we refer to other types of districts — such as Senate districts — we will so specify.

[3] The House redistricting committee determined that to comport with the one-person, one-voter requirement of the United States Constitution, see Georgia v. Ashcroft, 539 U.S. 461, 123 S.Ct. 2498, 2516 n. 2, 156 L.Ed.2d 428 (2003), the population for each district would have to be between 37,698 and 41,666. See Black Political Task Force, 300 F.Supp.2d at 294-95.

[4] Two Chelsea precincts, 3-1 and 3-2, were outside the 2d Suffolk District prior to the 2001 redistricting. However, the effect of the Enacted Plan was not precisely to move precinct 3-1 back into the 2d Suffolk District and precincts 3-4, 4-2, and 4-3 out. This is because just before the redistricting process began in 2001, local communities redrew precinct lines, and thus, the lines for precinct 3-2 under the Enacted Plan were different than they were under the 1993 redistricting plan.

[5] Neither of CCHA's pre-enactment plans would have created a district with a majority Hispanic voting age or citizen voting age population.

[6] Two different plaintiffs'"plans," in addition to the Enacted Plan, were the subject of analysis at trial. The first was referred to as the "Harmon Plan" and the second was a revised version of the Harmon Plan, which we considered in our Black Political Task Force liability opinion and to which we referred in that opinion as "Plan No. 2." See 300 F.Supp.2d at 302. We need not, however, concern ourselves with the differences between the plaintiffs' two alternative plans because they are identical with respect to the 2d Suffolk District. Thus, we need not be more specific than to refer to the "Plaintiffs' Plan" in this case.

[7] The parties framed this issue prior to trial. Defendant moved for summary judgment on the basis that plaintiffs could not satisfy the first Gingles condition, but we denied the motion without prejudice, deferring disposition of the issue until after trial to allow the parties to develop a complete record. Cf. Metts v. Murphy, 363 F.3d 8 (1st Cir.2004) (en banc) (per curiam) (declining to resolve Gingles majority issue at complaint stage).

[8] Plaintiffs offered the testimony of two experts, Harmon and Richard Engstrom. Harmon's testimony concerned demographic statistics while Engstrom's testimony focused on electoral analyses.

[9] Brace was defendant's expert as to demographic statistics. Defendant's second expert, Harold Stanley, testified as to electoral analyses.

[10] We note that as an alternative to reliance on VAP simpliciter as the statistical measure for compliance with the first Gingles precondition, certain courts have indicated that requiring a demonstration of a supermajority of VAP is warranted where there is evidence of low citizenship rates. In Puerto Rican Legal Defense & Education Fund, Inc. v. Gantt, 796 F.Supp. 681 (E.D.N.Y.1992), for example, the court rejected any bright-line test for determining the VAP necessary to produce an effective voting majority, noting that "[e]ven higher VAP percentages [than 65%] may be required in Hispanic districts to account for their even lower citizenship ratio, lower voter turnout, and lower voter registration." Id. at 689. Because we find CVAP to be an appropriate measure, we need not address whether a supermajority might be required when using VAP as the measure.

[11] We recognize that as a consequence, differences in the demography of districts may yield different effective weights for votes when, as is customarily the case, representational equality for purposes of the one-person, one-vote obligation is measured by total population and electoral equality for purposes of examining vote dilution is measured by some standard relating to voter qualification. Thus, for example, even when the total population is evenly-distributed among districts, if the population of a particular district is disproportionately below voting age as compared to other districts, the voting strength of individual voters in that district is comparatively strong — that is, fewer voters in the district would be needed to elect a representative than would be needed in the other districts where the voting age populations is proportionally higher. Similarly, when a citizenship qualification is considered, the effect is that even when total population is evenly-distributed among districts, fewer citizens of voting age would be needed to elect a representative in districts with disproportionately low CVAP numbers than in districts with higher CVAP numbers. Such consequences of using total population to measure representational equality and a different standard in the evaluation of the fairness of voting practices and procedures, however, have not yet prompted judicial disapproval.

[12] The 90% and 95% confidence intervals were 30.47% to 33.22% and 30.25% to 33.44%, respectively.

[13] We note that a Ninth Circuit panel has held "that, to the extent that Gingles does require a majority showing, it does so only in a case where there has been no proof of intentional dilution of minority voting strength." Garza, 918 F.2d at 769. This holding is a bit puzzling. The ability to prove a § 2 case without proof of intent, a circumstance which Gingles involved, has generally been viewed as a substantial relaxation of the proof of a statutory violation. Indeed, the principal difference in proof between a § 2 vote dilution case and one brought under the 14th Amendment resides in the necessity to prove intent in the latter but not the former. Garza actually involved both Voting Rights Act and Equal Protection claims, but the panel chose to offer its unique observation concerning the "extent" of the first Gingles condition despite the fact that the liability portion of the opinion, where the observation was located, could as easily have been directed to the Equal Protection claim without offering this construction regarding the Voting Rights Act. In any event, Garza's construction of the first Gingles precondition, even if applied here, would have no effect because, as we determine below, plaintiffs ultimately fail to demonstrate discriminatory intent on the part of the Committee.

[14] The Supreme Court has also not resolved whether a claim involving an "influence" district — one in which a minority group, while unable to elect preferred candidates even with crossover support, can exercise substantial influence in an election — is viable under § 2. See Gingles, 478 U.S. at 46 n. 12, 106 S.Ct. 2752; see generally Rodriguez v. Pataki, 308 F.Supp.2d at 375-76 (distinguishing crossover and influence districts). Plaintiffs here have explicitly disavowed an influence district claim. Defendant argued that plaintiffs similarly have failed to raise, or have explicitly waived, any potential crossover district claims. However, the summary judgment hearing transcript defendant pointed to in support of this contention is not definitive on this point. Plaintiffs' counsel, on questioning from the court, merely confirmed that plaintiffs were "not claiming a coalition district." But plaintiffs have been steadfast in maintaining that they were pursuing a claim based upon an "effective" majority. We decline to foreclose any crossover claim on the basis of inartful labeling. This is an evolving area where precision in characterization has not yet been fully achieved. Moreover, insofar as plaintiffs can be said to have denied pursuing a "crossover" district claim, they did so only in arguing that such a claim was unnecessary because they satisfied the first prong of Gingles given the Hispanic VAP numbers. Thus, we do not view their somewhat ambiguous statements to be a broad waiver as to all potential "crossover" claims.

[15] We note that a three-judge court in the Southern District of New York, while accepting arguendo that a district in which a minority with a CVAP of 45.4% "may under certain circumstances satisfy the first Gingles precondition," Rodriguez v. Pataki, 308 F.Supp.2d at 405-06, also found as a matter of fact that a different district with a CVAP of 36.9% could not satisfy the several Gingles preconditions. Id. at 399-400; see also id. at 407-09 (comparing questions presented by two districts).

[16] We use the term "ethnically-polarized" and the phrase "ethnic voting polarization" to accord with our general use of the term "Hispanic" (and "non-Hispanic"). See supra note 1. We recognize the potential non-fit of these terms with references to white voters, but for clarity of discussion we omit inclusion of terms related to racial polarization.

[17] Endogenous elections are those in the affected electoral district; exogenous elections are those for other electoral offices in the general geographic area of the affected electoral district. See generally, Black Political Task Force, 300 F.Supp.2d at 304-10.

[18] The 2003 Boston City Council election took place on the eve of trial and consequently was not the subject of meaningful expert analysis. That election, however, was addressed in the parties' submissions and provides significant insights, even in the absence of meaningful expert analysis, regarding the Plaintiffs' Plan.

[19] We recognize that the record indicates Lopez is not Hispanic but is Cape Verdean. In this connection, we note that the plaintiffs' expert, Engstrom, who based his analysis on surnames, misclassified Lopez. Because we do not give any particular individual weight to the election Lopez won, and because the results for that election fit within the same general pattern of the other two school committee elections, we think the misclassification, which may (or may not) have been shared by voters, is potentially of little import. We report it to provide a full context for the School Committee elections and note that the misidentification underscores the often ambiguous quality of ethnic categories.

[20] Stanley reported that in the 1997 election over 230 of Valentin's 825 votes came from non-Hispanics and Lopez received more non-Hispanic votes than Hispanic votes (435 to 359). Stanley calculated that Lopez also received more non-Hispanic votes than Hispanic votes in the 1999 election (548 to 342). For the 2001 election, Stanley calculated that Valerio received a substantial portion of his votes from non-Hispanic voters (452 out of 998) and that Pazos received more votes from non-Hispanics than Hispanics (528 to 268).

[21] The correlation for Pazos in the 2001 election was positive (.545) but not statistically significant.

[22] The candidates, however, did well in the precinct with the highest Hispanic VAP (precinct 2-2), where the top Latino vote-getter placed first in the precinct in each of the three elections.

[23] We note that Pazos's receipt of only 36.8% of the Hispanic vote in the 2001 election may belie this conclusion to some extent, but we are concerned more with the general trends and results than any particular result, which may have resulted from any number of factors unrelated to voter polarization.

[24] Defendant's expert, Stanley, also pointed out that according to the raw election return numbers, almost half of Avellaneda's votes came from non-Hispanic voters (733 out of 1,471). Unlike with the school committee elections, this does not seem to be a result of low voter turnout. Indeed, Engstrom reported that 22.5% of those receiving ballots were Hispanic and Chelsea's voter registration for Hispanics at the time was 25.3%. However, we take the same view about actual election returns here as we did with the school committee elections, which is to say that we think they are not particularly probative on the issue of voter polarization but rather indicate the import of considering bottom-line numbers in assessing the relative voting strength of groups to be considered as part of the totality of circumstances analysis.

[25] In a supplemental affidavit, Stanley reported in a table a tally of the votes received by the seven candidates, and according to these numbers, Arroyo finished fifth, not sixth, as Engstrom asserted. This slight discrepancy (based on a difference of just over 100 votes) makes no difference in our evaluation of these election results.

[26] We note, however, that one of the factors to be considered as part of the totality of the circumstances inquiry is the "magnitude of racially polarized voting within the relevant jurisdiction." S.Rep. No. 97-417, at 28-29, reprinted in 1982 U.S.C.C.A.N. 177, 206-07. Consequently, we will revisit the question of polarized voting in our totality of the circumstances analysis.

[27] In this regard, we find Engstrom's "reaggregation" analysis of the 2002 Senate primary, in which he estimated election results under the Plaintiffs' Plan, to be of limited probative value. Engstrom reported that in the Chelsea precincts included in the 2d Suffolk District as drawn in the Plaintiffs' Plan, Barrios won 53.2% of the votes. However, Engstrom's analysis did not include any of the East Boston precincts included in the 2d Suffolk District under the Plaintiffs' Plan. He merely engaged in speculative extrapolation to reach the conclusion that "[g]iven that the Latino percentage of the VAP in the five East Boston precincts in the district is 44.6% percent, it would appear that Mr. Barrios is likely to have won a majority of the votes in the illustrative district." He presented no electoral data to support this claim.

[28] Plaintiffs noted that the Committee was aware of — albeit somewhat surprised by — the growing Hispanic population in East Boston, as it stated in its report regarding the Redistricting Act:

A decade ago, no one could have predicted the largest Hispanic population in the city of Boston would not be in Jamaica Plain or Roxbury or Dorchester or the South End, but in East Boston.

[29] We recognize that there are geographic and socio-economic dimensions as well to plaintiffs' community of interest argument. Two bridges allow for substantial foot and car traffic over Chelsea Creek, which separates Chelsea and East Boston. In contrast, the principal direct connection over the Mystic River, which separates Chelsea and Charlestown, is an elevated toll bridge with no sidewalks. Moreover, plaintiffs submitted substantial evidence demonstrating that Chelsea and East Boston are similar, and Chelsea and Charlestown are dissimilar, on a number of socio-economic indicators. This evidence, however, does not establish that ethnic considerations, as opposed to an effort to provide minimal disruption to the historic configuration of the 2d Suffolk District, played any role in the redistricting process.

[30] We note, however, that such a possibility approaches territory unexplored, as far as we can tell, in the case law. It is unclear what remedial outcome a successful constitutional claim would entail, given plaintiffs' failure to demonstrate that their alternative plan is superior to the Enacted Plan, at least as far as § 2 criteria are concerned. See Garza, 918 F.2d at 771 ("[S]ome showing of injury must be made to assure that the district court can impose a meaningful remedy." (emphasis in original)).