holding that non-racial reasons for divergent voting patterns should be considered under the totality-of-circumstances testSummary of this case from Ala. State Conference of N.A. for Advancement of Colored People v. Merrill
Heard August 2, 1995.
Decided December 29, 1995.
Steven P. Perlmutter, with whom Michael D. Lurie, Robinson Cole, and Edward R. Mitnick, Acting City Solicitor, were on brief, for appellant.
Daniel J. Gleason, with whom Nelson G. Apjohn, Nutter, McClennen Fish, Alan J. Rom, Law Office of Sherwin Kantrovitz, P.C., David P. Hoose, and Katz, Sasson, Hoose Turnbull were on brief, for appellees.
APPEAL FROM THE UNITED STATES DISTRICT COURT FOR THE DISTRICT OF MASSACHUSETTS.
Before SELYA, CYR and STAHL, Circuit Judges.
In 1965, Congress enacted the Voting Rights Act (the VRA), Pub.L. No. 89-110, 79 Stat. 437 (codified at 42 U.S.C. §(s) 1973-1973o). Three decades later, the legislation remains a Serbonian bog in which plaintiffs and defendants, pundits and policymakers, judges and justices find themselves bemired.
The case before us opens yet another window on the conceptual complexity that has engulfed the VRA. It arises against the backdrop of the biennial elections that are held for city council in Holyoke, Massachusetts. The plaintiffs, two nonprofit organizations with ties to the Hispanic community and eight voters of Hispanic descent, complain that the electoral structure violates section 2 of the VRA by denying Hispanics equal opportunity to "participate in the political process and to elect representatives of their choice." 42 U.S.C. § 1973(b). The district court found merit in the plaintiffs' complaint with regard to councilmanic elections and granted relief. See Vecinos de Barrio Uno v. City of Holyoke, 880 F. Supp. 911 (D.Mass. 1995). After careful consideration of a bulky record, we are unable to square the lower court's factual findings with its ultimate conclusion of vote dilution. Consequently, we vacate the judgment and remand for further proceedings.
The plaintiffs also challenged the way in which members of the school committee were elected. The district court repulsed that challenge, see Holyoke, 880 F. Supp. at 928, and the plaintiffs do not press the point on appeal.
We sketch the background, reserving a more exegetic treatment of the facts pending our discussion of specific issues. We refer those readers who yearn for an immediate rush of details to the district court's informative opinion. See id. at 917-25.
Since 1963, the Holyoke city council has been composed of fifteen members, eight elected at large and seven elected by ward. Candidates run without party labels for two-year terms. Each voter is entitled to cast a ballot for a candidate in his or her ward, and to vote for up to eight at-large candidates.
The Hispanic community in Holyoke has grown dramatically over the past two decades. By 1990, persons of Hispanic origin accounted for 31.06% of the total population (compared to 13.8% in 1980). Under the current districting scheme — the ward lines were last redrawn in 1992 — Hispanic voters comprise a clear majority in two wards and account for nearly one-third of the population in a third ward. Yet, while Hispanic-preferred city council candidates have prevailed in the two "Hispanic majority" wards, no person of Hispanic descent ever has been elected to an at-large seat. This discrepancy crystallizes into the nub of the plaintiffs' case: their vote dilution claim is that, while Hispanics now constitute 21.89% of Holyoke's voting age population, the electoral structure limits the Hispanic community's ability to elect the candidates its members prefer to only 14% of the available city council seats (two of fifteen).
The district court agreed with the plaintiffs that the Hispanic vote had been impermissibly diluted. See id. at 925-27. To remedy the perceived inequity, the court by separate order left the ward lines and representation intact, but cut back the number of at-large seats from eight to two (thus shrinking the council from fifteen to nine members, and making its electoral structure congruent with that of the school committee). See Vecinos de Barrio Uno v. City of Holyoke, 882 F. Supp. 9, 10 (D.Mass. 1995) ( Holyoke II). The court reasoned that, under the revised format, Hispanics probably would continue to control two of the ward seats, and that decreasing the size of the council would boost Hispanics' percentage representation to a level that would compare favorably with their percentage of the voting age population as a whole. See id. at 12.
The district court, striving to put its remedial order in place in time for the November 1995 municipal election cycle, see id. at 13, entered the order under pressure of time. The city appealed and simultaneously moved for a stay. By an unpublished order, we expedited the appeal and granted the stay. Hence, the November 1995 elections were held under the preexisting scheme.
II. STANDARD OF REVIEW
The bedrock on which the district court's opinion rests is its conclusion that the at-large component of the electoral structure unlawfully dilutes the Hispanic community's voting power. As a general matter, a finding of vote dilution made after a bench trial is a finding of fact subject to review under the "clearly erroneous" rubric. See Thornburg v. Gingles, 478 U.S. 30, 78-79 (1986); Houston v. Lafayette County, 56 F.3d 606, 610 (5th Cir. 1995); Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1116 (3d Cir. 1993), cert. denied, ___ U.S. ___, 114 S.Ct. 2779 (1994); see also Fed.R.Civ.P. 52(a). This means that a reviewing court ought not to disturb such a finding "unless, on the whole of the record, [the court] form[s] a strong, unyielding belief that a mistake has been made." Cumpiano v. Banco Santander P.R., 902 F.2d 148, 152 (1st Cir. 1990).
Though the clear error standard is formidable, it is not a juggernaut that crushes everything in its path. One important qualification is that the jurisprudence of clear error "does not inhibit an appellate court's power to correct errors of law, including those that may infect a so-called mixed finding of law and fact, or a finding of fact that is predicated on a misunderstanding of the governing rule of law." Gingles, 478 U.S. at 106 (quoting Bose Corp. v. Consumers Union, 466 U.S. 485, 501 (1984)); accord LoVuolo v. Gunning, 925 F.2d 22, 25 (1st Cir. 1991). Considering asserted errors of law entails nondeferential review. See In re Extradition of Howard, 996 F.2d 1320, 1327 (1st Cir. 1993).
III. PROVING VOTE DILUTION
In order to sharpen the focus of our inquiry, we first limn the statutory framework and elucidate the requirements that attend a proper showing of vote dilution.
Section 2 of the VRA, as amended in 1982, prohibits any 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). A denial or abridgement of the right to vote is established when,
based on the totality of circumstances, it is shown that the political processes leading to nomination or election . . . are not equally open to participation by all members of a [protected] class of citizens . . . in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice. The extent to which members of a protected class have been elected to office . . . is one circumstance which may be considered: Provided, That nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.
42 U.S.C. § 1973(b). While the statutory scheme does not provide an assurance of success at the polls for minority candidates, see Johnson v. De Grandy, ___ U.S. ___, 114 S.Ct. 2647, 2658 n. 11 (1994), it does provide an assurance of fairness. Thus, when "a certain electoral law, practice, or structure interacts with social and historical conditions to cause an inequality in the opportunities enjoyed by [minority] and white voters to elect their preferred representatives," a section 2 claim lies. Gingles, 478 U.S. at 47.
The platform required to launch a vote dilution claim must contain three interleaved planks. First, the plaintiffs must prove that they are part of a minority group that is "sufficiently large and geographically compact to constitute a majority in a single-member district." Id. at 50. Second, they must show that the group is "politically cohesive." Id. at 51. Third, they must demonstrate significant bloc voting by non-minorities. See id. Each of these showings must be specific to the electoral unit that is under fire.
This precondition will have to be reconfigured to the extent that the courts eventually validate so-called influence dilution claims. See Voinovich v. Quilter, 113 S.Ct. 1149, 1157-58 (1993) (discussing treatment of claims brought on behalf of persons who constitute a potentially influential bloc, but less than the majority, within the relevant electorate, and raising prospect that the first Gingles precondition may have to be "modified or eliminated"). The lower courts are divided on the subject, compare Armour v. Ohio, 775 F. Supp. 1044, 1052 (N.D. Ohio 1991) (three-judge panel) (recognizing influence dilution claim) with McNeil v. Springfield Park Dist., 851 F.2d 937, 947 (7th Cir. 1988) (rejecting influence dilution claim), cert. denied, 490 U.S. 1031 (1989), and the Supreme Court has declined on four occasions to decide whether such claims are cognizable under VRA Section(s) 2. See De Grandy, ___ U.S. ___, 114 S.Ct. at 2656; Voinovich, 113 S.Ct. at 1157-58; Growe v. Emison, 113 S.Ct. 1075, 1084 n. 5 (1993); Gingles, 478 U.S. at 46-47 n. 12. We take no view of the matter today (although we do discuss the potential relevance of evidence from elections in a particular "influence district" on the plaintiffs' claims, see infra Part V).
The first two Gingles preconditions look to whether, putting the challenged practice, procedure, or structure to one side, minority voters within a given constituency have the potential to elect representatives of their choice. See Growe v. Emison, 113 S.Ct. 1075, 1084 (1993); Gingles, 478 U.S. at 50 n. 17. If, for example, minority voters in an at-large system are so widely dispersed that they could not elect preferred candidates under some reasonable alternative scheme, then the "at-large system cannot be responsible for that group's inability to elect its candidates." Solomon v. Liberty County, 899 F.2d 1012, 1018 (11th Cir. 1990), cert. denied, 498 U.S. 1023 (1991). Similarly, unless the minority group is politically cohesive, "it cannot be said that the selection of a [particular] electoral structure thwarts distinctive minority group interests." Gingles, 478 U.S. at 51. The third Gingles precondition — which embodies a showing that the majority votes sufficiently as a bloc to enable it, in the ordinary course, to trounce minority-preferred candidates most of the time, see Voinovich v. Quilter, 113 S.Ct. 1149, 1157 (1993) — addresses 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. See De Grandy, ___ U.S. ___, 114 S.Ct. at 2657; Gingles, 478 U.S. at 51.
Proof of all three preconditions creates an inference that members of the minority are in fact harmed by the challenged electoral practice, procedure, or structure. However, the inference is rebuttable. As a result, establishing the three Gingles preconditions is necessary, but not always in itself sufficient, to ensure success on a section 2 claim. That is to say, because the inference of vote dilution can be rebutted by the force of other evidence, proof of the three preconditions, without more, will not invariably carry the day. See De Grandy, ___ U.S. ___, 114 S.Ct. at 2657. Put another way, the critical question in a vote dilution case is whether minority voters have an equal opportunity to participate in the electoral process. While the threshold elements catalogued by the Gingles Court shed considerable light on this inquiry, they do not comprise the only conceivable source of illumination. Completing the inquiry demands "comprehensive, not limited, canvassing of relevant facts." Id.
Consistent with this approach, courts must be careful not to wear blinders. The judge must sift the evidence produced at trial and gather enough information to paint a true picture of the attendant facts and circumstances. He or she must then make a realistic appraisal of what the picture discloses. See Gingles, 478 U.S. at 45 (advocating achievement of a "practical evaluation of the past and present reality" through a "functional view of the political process"). Some guidance can be found in a list of factors highlighted in the congressional report that accompanied the 1982 amendment to VRA Section(s) 2, see S.Rep. No. 417, 97th Cong., 2d Sess., at 28-29 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 206-07, but the judge should not stop there. Though helpful, the list is not all-encompassing. See Gingles, 478 U.S. at 45; Little Rock Sch. Dist. v. Pulaski County Special Sch. Dist., 56 F.3d 904, 910 (8th Cir. 1995). Since communities differ, and elections play out differently in different venues at different times, the judge must make a case-specific determination, giving due weight to the idiosyncracies that bear upon the particular situation. See Jenkins, 4 F.3d at 1115.
One road that we believe remains open to a court called upon to examine the totality of the circumstances in a vote dilution case is to mull other factors, apart from racial bias, that may have caused the white bloc voting identified in the third Gingles precondition. While the Gingles Court split on this question, compare Gingles, 478 U.S. at 63-64 (opinion of Brennan, J.) (stating for four justices that the etiology of racially polarized voting is irrelevant under VRA Section(s) 2) with id. at 100-02 (O'Connor, J., concurring in the judgment) (stating for four justices that the reasons why white voters reject minority candidates are relevant) and id. at 82-83 (White, J., concurring) (rejecting, without explanation, Justice Brennan's view), and controversy has raged since then, see, e.g., Nipper v. Smith, 39 F.3d 1494, 1513-14 (11th Cir. 1994) (en banc) (holding for two judges, with two judges dissenting, that the existence of racial bias in the community is relevant to a section 2 claim), cert. denied, 115 S.Ct. 1795 (1995); League of United Latin Am. Citizens, Council No, 4434 [LULAC] v. Clements, 999 F.2d 831, 850-63 (5th Cir. 1993) (en banc) (reaching similar conclusion, with three judges dissenting), cert. denied, ___ U.S. ___, 114 S.Ct. 878 (1994), we are of the view that De Grandy has removed much of the doubt.
We recognize that such widely used terms of art as "white bloc voting" and "racially polarized voting" may not always capture the subtleties of specific problems that arise in the political process. The case at bar, for example, involves the voting patterns of the majority (loosely termed "white") and the specific minority symbolized by the plaintiffs (loosely termed "Hispanics"). Concededly, this taxonomy is imprecise; for example, not all people who are considered "Hispanic" necessarily consider themselves "non-white." To that extent, then, the phrase "white bloc voting," though used repeatedly throughout the decided cases, may be somewhat inaccurate or even slightly misleading. Similarly, VRA Section(s) 2 applies to denials of the right to vote on account of either race or color, yet the opinions harp on the phrase "racially polarized voting." To that extent, the idiom of the case law may neglect potentially important distinctions between the concepts of "race" and "color." While acknowledging these limitations, we can think of no universal solution, and, thus, take refuge in the pat terminology.
Even when the Gingles preconditions coalesce and thereby create an inference of discrimination, lack of equal electoral opportunity remains the central focus of the inquiry. Furthermore, that question "must still be addressed explicitly, and without isolating any other arguably relevant facts from the act of judgment." De Grandy, 114 S.Ct. at 2657. It seems self-evident that the presence or absence of bias is at least "arguably relevant" to the question of whether a minority lacks equal electoral opportunity. After all, a minority group's prospects for electoral success in a community riven along racial lines differ significantly from its prospects in a more unified community. We agree with the Fifth Circuit that "[a] tendency among whites to cast their votes on the basis of race presents a far more durable obstacle to the coalition-building upon which minority electoral success depends than disagreements over ideology." LULAC, 999 F.2d at 858.
By like token, however, sentiments unrelated to race also can be powerful stimuli. When it can be shown that, in a particular community, voters are moved primarily by causes unrelated to race, it is reasonable to assume that a minority-preferred candidate who embodies these values might equally be able to engender majoritarian (white) support. See Gingles, 478 U.S. at 100-01 (O'Connor, J., concurring). Thus:
Evidence that a candidate preferred by the minority group in a particular election was rejected by white voters for reasons other than those which made that candidate the preferred choice of the minority group would seem clearly relevant in answering the question whether bloc voting by white voters will consistently defeat minority candidates.
Id. at 100.
The upshot is that when racial antagonism is not the cause of an electoral defeat suffered by a minority candidate, the defeat does not prove a lack of electoral opportunity but a lack of whatever else it takes to be successful in politics (say, failure to support popular programmatic initiatives, or failure to reflect the majority's ideological viewpoints, or failure to appreciate the popularity of an incumbent). Section 2 does not bridge that gap — nor should it. See De Grandy, ___ U.S. ___, 114 S.Ct. at 2658 n. 11; see also Baird v. Consolidated City of Indianapolis, 976 F.2d 357, 361 (7th Cir. 1992) (explaining that section 2 "is a balm for racial minorities, not political ones — even though the two often coincide"), cert. denied, 113 S.Ct. 2334 (1993). We believe it follows that, after De Grandy, plaintiffs cannot prevail on a VRA Section(s) 2 claim if there is significantly probative evidence that whites voted as a bloc for reasons wholly unrelated to racial animus. We so hold.
This holding draws sustenance from the language of section 2 itself, particularly the statute's prohibition of electoral structures that result in a denial or abridgement of the right to vote "on account of race or color." 42 U.S.C. § 1973(a). Other courts have found this language determinative of the question, see, e.g., Nipper, 39 F.3d at 1515-17; LULAC, 999 F.2d at 850, especially when coupled with legislative history indicating that an electoral scheme violates VRA Section(s) 2 only when it "interacts with racial bias in the community and allows that bias to dilute the voting strength of the minority group." Nipper, 39 F.2d at 1520 (commenting upon legislative history).
Those — including the present plaintiffs — who favor a more single-minded interpretation of section 2 marshal a regiment of counterarguments. Their most serious objection questions the compatibility of our holding with Congress's action in amending section 2 to scrap the "intent" test imposed by City of Mobile v. Bolden, 446 U.S. 55, 62 (1980), and to insert in its place the "results" test earlier adumbrated in White v. Regester, 412 U.S. 755, 765-66 (1973), and Whitcomb v. Chavis, 403 U.S. 124, 143 (1971). This substitution permits plaintiffs to show vote dilution by proving that electoral structures "result 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. §(s) 1973(a), and, concomitantly, relieves them of the burden of proving that the structures were set in place to advance a racially discrimination purpose. Against this mise-en-scene, some have equated Congress's adoption of the "results" test with an intention to foreclose any inquiry whatever into the reasons why minority groups lack opportunities for electoral participation.
We do not believe that the 1982 amendment lends itself to this restrictive conclusion. The now-discarded "intent" test specifically required plaintiffs to prove that government created or maintained the challenged electoral structure with a discriminatory purpose, actually intending that a structure would disadvantage minority voters. See Mobile, 446 U.S. at 62-63. In enacting the amendment, Congress shifted the law's focus: plaintiffs no longer have to prove discriminatory intent but instead have to carry the burden of proving that the challenged electoral structure results in a denial of equal opportunity on account of race.
Properly conceived, the results test protects racial minorities against a stacked deck but does not guarantee that they will be dealt a winning hand. Whitcomb — an opinion purportedly codified in the 1982 amendment — illustrates the point. There, the Court discerned no denial of equal opportunity when a minority group's failure to elect its preferred candidates "emerges more as a function of losing elections than of built-in bias" directed by the establishment majority against the minority group. Whitcomb, 403 U.S. at 153. The lesson to be learned is that, even when election returns in effect short-circuit a minority group's voting power, the electoral structure is not illegal if the defeat represents nothing more than the routine operation of political factors. See id. In other words, even under the 1982 amendment, a lack of electoral success unrelated to race is not a proxy for a lack of opportunity to succeed. Hence, VRA Section(s) 2, as amended, despite its focus on results, does not require courts to ignore evidence that factors other than race are the real obstacles to the political success of a minority group. See Gingles, 478 U.S. at 101 (O'Connor, J., concurring) ("The overall vote dilution inquiry neither requires nor permits an arbitrary rule against consideration of all evidence concerning voting preferences other than statistical evidence of racial voting patterns.").
We recognize, of course, that permitting inquiry into the causes of white bloc voting potentially jeopardizes the remedial purposes of the VRA in the sense that it may make proof of vote dilution more difficult. Courts have expressed concern on this score, see, e.g., id. at 72 (opinion of Brennan, J.); LULAC, 999 F.2d at 860, and these concerns are not without foundation. Yet, two responses spring to mind. First, the VRA is designed to ensure that the electoral process is fair and the opportunities for access to it are equal. Forcing courts to turn a blind eye to other causes of majoritarian bloc voting serves neither of these ends, but, rather, facilitates a back-door approach to proportional representation. That is not a door through which Congress desired courts to pass. See 42 U.S.C. § 1973(b) (stating that nothing in the VRA "establishes a right to have members of a protected class elected in numbers equal to their proportion in the population").
Second, we suspect that at bottom the skeptics misapprehend the nature of the showing needed to support a section 2 claim. As amended, the statute allows plaintiffs to establish a prima facie case of vote dilution by proving the three Gingles preconditions.
The second and third preconditions are designed to assay whether racial cleavages in voting patterns exist and, if so, whether those cleavages are deep enough to defeat minority-preferred candidates time and again. If proven, these preconditions give rise to an inference that racial bias is operating through the medium of the targeted electoral structure to impair minority political opportunities. See De Grandy, ___ U.S. ___, 114 S.Ct. at 2657 (noting that a "lack of equal electoral opportunity may be readily imagined and unsurprising when demonstrated under circumstances that include the three essential Gingles factors"); Nipper, 39 F.3d at 1525 (stating that "proof of the second and third Gingles factors will ordinarily create a sufficient inference that racial bias is at work"); United States v. Marengo County Comm'n, 731 F.2d 1546, 1567 (11th Cir.) (stating that the second and third Gingles preconditions remain the "surest indication of race-conscious politics"), cert. denied, 469 U.S. 976 (1984).
The resultant inference is not immutable, but it is strong; it will endure unless and until the defendant adduces credible evidence tending to prove that detected voting patterns can most logically be explained by factors unconnected to the intersection of race with the electoral system. See Nipper, 39 F.3d at 1524. It is only when such evidence possesses convictive force that the inference of racial animus will be called into serious question. See De Grandy, 114 S.Ct. at 2658.
Such factors might include, for example, organizational disarray, lack of funds, want of campaign experience, the unattractiveness of particular candidates, or the universal popularity of an opponent.
Even if such proof is forthcoming, the defendant does not automatically triumph. Instead, the court must determine whether, based on the totality of the circumstances (including the original inference and the factual predicate that undergirds it), the plaintiffs have proven that the minority group was denied meaningful access to the political system on account of race. The burden of proof at all times remains with the plaintiffs; defendant's burden is an entry-level burden of production. Thus, once the defendant proffers enough evidence to raise a legitimate question in regard to whether nonracial factors adequately explain racial voting patterns, the ultimate burden of persuading the factfinder that the voting patterns were engendered by race rests with the plaintiffs.
Despite the allocation of the burden of proof, this framework imposes a high hurdle for those who seek to defend the existing system despite meaningful statistical evidence that suggests bloc voting along racial lines. See Jenkins, 4 F.3d at 1135. We predict that cases will be rare in which plaintiffs establish the Gingles preconditions yet fail on a section 2 claim because other facts undermine the original inference. In this regard, we emphasize that establishing vote dilution does not require the plaintiffs affirmatively to disprove every other possible explanation for racially polarized voting. Rather, plaintiffs must simply prove that the three threshold preconditions (alone or in combination with the totality of the circumstances) are strong enough in a given case that, notwithstanding the countervailing evidence of other causative agents mustered by the defendant, the record sustains a claim that racial politics — specifically, the interaction of race and the electoral system — have resulted in significantly diminished opportunities for minority participation in elective government.
The proceedings below illustrate this point. The district court acknowledged — but did not accept — the City's attribution of the Hispanic community's lack of electoral success to "factors other than the at-large system itself, including voter apathy, unattractive candidates, poor campaign organizations and the like — all evidenced by low voter turnout." Holyoke, 880 F. Supp. at 926. The court seemingly rejected the City's alternative explanations as a matter of fact. See id. (concluding that, based on the overall evidence, an Hispanic candidate, "no matter how attractive and no matter how effective at bringing out the Hispanic vote, [would not have had] a fair opportunity to win any at-large election in Holyoke during this period"). Having used this illustration, we hasten to add that, on remand, the court is free to revisit the evidence and reconsider all its findings, including those that touch upon other possible causes of racially polarized voting.
IV. THE ASSIGNMENTS OF ERROR
Having cemented into place the general framework for evaluating vote dilution claims, we shine the light of our gleaned understanding on the City's objections to the decision below. We divide our discussion into four segments.
A. The Analytic Model .
In this case, the district court analyzed fifteen different races in six different election years spanning a ten-year period from 1983 through 1993. Taking this evidence as an undifferentiated whole, the court found a pattern of racially polarized voting sufficient to support the plaintiffs' prayer for relief. The City assigns error, positing that racially polarized voting cannot be deduced from an overview which blends data from a series of separate elections, some suspect and some unexceptionable. The City's point is that only evidence from "legally significant" elections can be relevant to, or can underbrace, a finding that VRA Section(s) 2 has been abridged. Warming to this theme, the City asserts that each of the three Gingles preconditions must be shown to exist vis-a-vis a particular election before a court may mull what transpired at that election as a link in the evidentiary chain that leads to a determination of vote dilution. If this approach were adopted, the court below, in considering whether the plaintiffs had established a pattern of racially polarized voting over the years, could not have relied upon evidence drawn from any discrete election unless the plaintiffs first proved a violation of the VRA in regard to that election. We reject the City's approach.
To give concrete examples, Holyoke contends that in 1983 Hispanics did not constitute a sufficiently compact group to satisfy the first Gingles precondition, thus rendering any evidence of white bloc voting in that year legally irrelevant. In the same vein, the City insists that the district court should have ignored evidence of racially polarized voting in any elections won by minority candidates or in which Hispanics did not sufficiently cohere.
In this enlightened day and age, bigots rarely advertise an intention to engage in race-conscious politics. Not surprisingly, therefore, racially polarized voting tends to be a silent, shadowy thief of the minority's rights. Where such activity is detected at all, the process of detection typically involves resort to a multifaceted array of evidence including demographics, election results, voting patterns, campaign conduct, and the like. Usually, such evidence is not neatly packaged but must be pieced together bit by bit from data accumulated in a series of elections. See Gingles, 478 U.S. at 57; Jenkins, 4 F.3d at 1119; Gomez v. City of Watsonville, 863 F.2d 1407, 1417 (9th Cir. 1988), cert. denied, 489 U.S. 1080 (1989); City of Carrollton Branch of NAACP v. Stallings, 829 F.2d 1547, 1557 (11th Cir. 1987), cert. denied, 485 U.S. 936 (1988). Thus, the question whether a given electoral district experiences racially polarized voting to a legally significant extent demands a series of discrete inquiries not only into election results but also into minority and white voting practices over time.
We think that this analysis exposes the principal flaw in Holyoke's thesis. The requirement of "legal significance" is not a barometer for deciding what evidence of racially polarized voting may be considered; rather, it is the benchmark against which all the evidence, taken in sum, must be measured. And although weaknesses in plot lines siphoned from individual elections may well imperil an overall conclusion of legally significant racially polarized voting — the whole is frequently not greater than the sum of the parts — such weaknesses do not render evidence from those elections inadmissible. It follows that reliance on such evidence does not necessarily invalidate an overall conclusion that unlawful vote dilution exists. See Gingles, 478 U.S. at 57 (explaining that "in a district where elections are shown usually to be polarized, the fact that racially polarized voting is not present in one or a few individual elections does not necessarily negate the conclusion that the district experiences legally significant bloc voting").
This paradigm is fully consistent with the reality of events. One swallow does not a summer make, and the results of a single election are unlikely, without more, to prove the existence or nonexistence of embedded racial cleavages. Thus, race-conscious politics (or its absence, for that matter) can more readily be seen by producing a documentary that spans a series of elections than by taking an isolated snapshot of a single election. After all, to be legally significant, racially polarized voting in a specific community must be such that, over a period of years, whites vote sufficiently as a bloc to defeat minority candidates most of the time. See id. at 56. In order reliably to tell whether racial groups do (or do not) band together behind particular candidates with regularity, all elections in the relevant time frame (or, at least, a representative sampling of them) must be studied — not just those elections that, taken in isolation, reveal the cicatrices of racially polarized voting.
On this basis, we reject the City's contention that the failure to prove any one Gingles precondition in any one election eliminates that election from judicial consideration. The preconditions are necessary to prove an overall conclusion of vote dilution, not to demonstrate the relevance vel non of particular morsels of evidence. Hence, the court below had every right to analyze all the elections (suspect and non-suspect) in its effort to ascertain both whether (1) the Hispanic community usually coheres behind particular candidates, and (2) Holyoke's white citizens usually vote against Hispanic-preferred candidates in sufficient numbers to prevent their election.
Holyoke's city council model — seven ward seats and eight at-large seats — dates back more than three decades (to an era when few persons of Hispanic descent dwelt in the municipality). Currently, the Hispanic community effectively controls two of the fifteen city council seats (Wards 1 and 2). In addition, Hispanics constitute about 28% of the voting age population in Ward 4. Based on these population statistics, the City strives to persuade us that the plaintiffs cannot satisfy the Gingles preconditions because Hispanics, as a group, are insufficiently compact to constitute the majority in a third ward. We are unconvinced for two reasons.
First, the City failed to make this claim in its brief, asserting it for the first time at oral argument. Thinking on one's feet is a useful tool of appellate advocacy only if the thinker has a suitable foothold in the record. Here, the thought was too little too late. See United States v. Gertner, 65 F.3d 963, 971 n. 7 (1st Cir. 1995) (refusing to entertain an argument not raised in the government's appellate brief); see also Fed.R.App.P. 28(a).
In all events, the City's spur-of-the-moment retort is founded on a faulty premise. It assumes that the relevant benchmarks for matching the Hispanic population and its opportunity for access are the seven single-member wards. This assumption is faulty because the litigation challenges Holyoke's electoral system as a whole, and, to the extent the challenge is scissile, its cynosure is not the wards but the system's at-large component — a component that allegedly dilutes the plaintiffs' opportunity for full political participation in municipal affairs. Refined to bare essence, the plaintiffs' theory is that, because of the combined impact of the at-large electoral structure and an incipient pattern of racially polarized voting, Hispanics can only hope to elect candidates to two of the fifteen city council seats (i.e., about 14% of the seats) even though they comprise at least 22% of Holyoke's total population. Thus, the City's emphasis on the seven wards misses the point.
This does not mean that the wards are an irrelevancy. A successful vote dilution challenge "must find a reasonable alternative practice as a benchmark against which to measure the existing voting practice." Holder v. Hall, ___ U.S. ___, 114 S.Ct. 2581, 2585 (1994). In that sense, the single-member districts are relevant to an assessment of the system's at-large component. While it may be theoretically possible to analogize the plaintiffs' claim to a challenge addressed to a multimember at-large district — in which case the court would have to compare Hispanic opportunities to elect candidates to one of the eight undifferentiated at-large seats to the potential opportunities that might exist if the multimember district were divided into eight contiguous single-member districts, see, e.g., id. at 2589 (O'Connor, J., concurring) ("In a challenge to a multimember at-large system . . . a court may compare it to a system of multiple single-member districts.") — the analogy cannot be carried past its logical limits. Here, the analogy would be imperfect because the plaintiffs' challenge is addressed to Holyoke's electoral system as a whole. Accordingly, the district court had an obligation to consider whether that system — not just its at-large component — provides minorities with an equal opportunity to elect candidates of their choice. See Baird, 976 F.2d at 360; NAACP v. City of Columbia, 850 F. Supp. 404, 429 (D.S.C. 1993), aff'd, 33 F.3d 52 (4th Cir. 1994), cert. denied, ___ U.S. ___, 115 S.Ct. 1095 (1995).
Bearing this in mind, we think that the lower court constructed a reasonable benchmark by comparing current Hispanic opportunities to potential opportunities that would exist if the municipality were divided into fifteen single-member councilmanic districts. And since we find no clear error in the court's conclusion that, under its projected set of circumstances, Hispanics would constitute a majority of the votes in at least three of fifteen reconstituted wards, we cannot set aside its finding that Hispanics are a sufficiently compact group.
C. Low Voter Turnout .
The City also asserts that, given the consistently low turnout among Holyoke's Hispanic voters, see Holyoke, 880 F. Supp. at 925 ("Hispanic turnout rates in Holyoke have varied from 22% to as low as 2% over a ten-year period, . . . differing considerably from election to election and from precinct to precinct."), the district court erred as a matter of law in declaring the Hispanic community to be politically cohesive. In the City's view, low turnout betokens voter apathy and precludes a finding that particular candidates received significant minority support (as required to show minority political cohesion under Gingles, 478 U.S. at 56). The plaintiffs concede the anemic turnout but argue that it is irrelevant to the political cohesion inquiry. They take the position that courts should frame answers to such inquiries after considering the behavior of those minority voters who actually opt to participate in the electoral process, and not gaze beyond that group (whatever its size) to count the number that sit on the sidelines.
We walk a middle path. A principal objective of the VRA is to provide a level playing field on which minority candidates — like all candidates — will be exposed only to the routine vicissitudes of the electoral process, not to special impediments arising out of the intersection of race and the electoral system. So, if a defeat at the polls (or even a string of defeats) is caused by, say, a candidate's lack of merit or a campaign's lack of focus, the Voting Rights Act is not infringed. See Whitfield v. Democratic Party of State of Ark., 890 F.2d 1423, 1430 (8th Cir. 1989) (explaining that a "causal connection between the challenged practice . . . and the diluted voting power of the minority must be established"). By like token, if the defeat of minority candidates occurs because the votes of the members of the minority community are scattered due to their diverse interests, then the requisite causal connection is lacking. See Gingles, 478 U.S. at 51. Under such circumstances, the interaction of race with the electoral system cannot justly be blamed for a minority group's lack of success at the polls.
In the case of low voter turnout, the electoral system may not always be so easily absolved. For one thing, even with a modest turnout, the actual votes cast may be probative of minority cohesion if a sufficiently strong pattern emerges. See, e.g., United States v. Dallas County Comm'n, 739 F.2d 1529, 1536 n. 4 (11th Cir. 1984). For another thing, low 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. In such instances, low turnout itself may actually be probative of vote dilution. See, e.g., Gingles, 478 U.S. at 69 (opinion of Brennan, J.); see also Gomez, 863 F.2d at 1416 n. 4 (suggesting that voter apathy traceable to past discrimination is "evidence of minority voters' lack of ability to participate effectively in the political process"); Kirksey v. Board of Supervisors, 554 F.2d 139, 145 n. 13 (5th Cir.) (observing that failure to register may be a residual effect of previous lack of access or feelings of futility in light of white bloc voting), cert. denied, 434 U.S. 968 (1977); see also Buckanaga v. Sisseton Indep. Sch. Dist., 804 F.2d 469, 475 (8th Cir. 1986). When low turnout results from the very problems that the Voting Rights Act is intended to ameliorate, it would be mindless for courts to ignore the evidence of minority cohesion that can be culled from the actual ballot tallies.
This is not to say, as the plaintiffs would have it, either that low voter turnout is altogether irrelevant to a vote dilution inquiry, or that courts should look only to actual voting results. The cause of poor turnout is often difficult to detect. If minority voters have not made reasonable efforts to organize and participate in the electoral system, courts cannot accurately gauge the effects of the system on the minority group's political aspirations. See City of Columbia, 850 F. Supp. at 415-16. Furthermore, low turnout sometimes may be an indicium of ebbing community support for a particular minority candidate. See id. at 418-19. Hence, evidence of this nature may — or may not — be probative on the issue of minority cohesion.
In the final analysis, the question of whether low minority voter turnout helps or hurts a claim of vote dilution, and the related question of whether actual votes cast provide a sufficient (or better) measure of minority political cohesion without regard to turnout, both depend on the facts and circumstances of the particular case. Consequently, courts cannot resort to the easy visibility of a bright-line rule. On this delicate, fact-sensitive issue, only a case-by-case approach satisfactorily permits courts to peel away the layers and conduct the functional vote dilution inquiry that the VRA requires.
In the case at hand, the district court made reasonably detailed findings concerning the relationship between depressed turnout among Hispanics and the structural attributes of Holyoke's electoral system. The court determined that the City imposed — or neglected to remove — a variety of obstacles to Hispanic political participation. The court mentioned, inter alia, the City's niggardly deployment of bilingual registrars and poll workers, its removal from voter registration rolls of Hispanics who did not fill out English-language census forms, and its failure to print ballot information posters in Spanish. See Holyoke, 880 F. Supp. at 925. In the court's estimation, these deficiencies, along with downtrodden socioeconomic conditions, accounted for the low turnout among Hispanic voters. Id. And to cap matters, the court found that the actual turnout, though small, was adequate to reflect political cohesion in the Hispanic community. Id.
We believe that these findings are supportable. In a vote dilution case characterized by meager turnout among minority voters, plaintiffs need not show that the sole cause of low numbers is the interaction between racial divisions in the community and identifiable elements of the electoral system. It is sufficient if the plaintiffs persuade the trial court that considerations implicating race contributed substantially to repressing minority participation. In light of the aggregate facts and circumstances, coupled with the district court's explicit findings, we believe that the plaintiffs satisfied this burden here. Thus, the evidence of low Hispanic turnout does not undercut the court's ultimate conclusion that the plaintiffs established minority political cohesion.
We leave open the possibility that especially low minority voter turnout in a particular election may be evidence that factors other than racially based politics (say, poor political organization or lack of minority community support) were the cause of the minority community's inability to elect its preferred candidate in that election.
D. Adequacy of the Findings .
The City's most telling point involves the lower court's application of relevant legal principles to discerned facts. In condoning the necessary appraisal, we are mindful that a district judge sitting without a jury cannot paint with too broad a brush. Rule 52(a) requires the judge to make findings of fact and conclusions of law that are sufficiently detailed to permit a reviewing court to ascertain the factual core of, and the legal foundation for, the rulings below. See Touch v. Master Unit Die Prods., Inc., 43 F.3d 754, 759 (1st Cir. 1995); Pearson v. Fair, 808 F.2d 163, 165-66 n. 2 (1st Cir. 1986) (per curiam).
This bedrock rule has particular force in cases of this genre. Vote dilution claims are often marked by a significant degree of complexity. Typically, the resolution of such claims demands a careful sifting of imbricated, highly ramified fact patterns. The legal principles that must be applied are convoluted, and they almost always touch upon constitutional precepts, together with important issues of federalism and the separation of powers. Accordingly, a trial court that decides a vote dilution case must be scrupulous in chronicling the relevant facts and delineating the linkages between those facts and the ultimate conclusion of vote dilution vel non. See Cousin v. McWherter, 46 F.3d 568, 574-75 (6th Cir. 1995); Buckanaga, 804 F.2d at 472; Velasquez v. City of Abilene, 725 F.2d 1017, 1020 (5th Cir. 1984). To this end, the district court must discuss "not only the evidence that supports its decision but also all the substantial evidence contrary to its opinion." Harvell v. Ladd, 958 F.2d 226, 229 (8th Cir. 1992) (quoting Buckanaga, 804 F.2d at 472); see also Houston, 56 F.3d at 612 n. 6 (similar; collecting cases). Despite the district judge's obvious investment of time and effort in the proceedings below, and his thoughtful analysis of difficult legal issues, the findings of fact in the instant case fail to satisfy these demanding criteria.
In any claim brought under VRA Section(s) 2, the Gingles preconditions are central to the plaintiffs' success. Here, the trial court sounded an uncertain trumpet in respect to both the second and third preconditions. This uncertainty stems from a lack of congruence between the court's subsidiary findings anent the particular elections it studied and its overall findings of minority cohesion and white bloc voting in Holyoke. We explain briefly.
The lower court analyzed fifteen elections in which Hispanic candidates ran for office. Of these, only four were at-large elections; the rest were ward elections for either city council or school committee seats. In four of the eleven ward elections, Hispanic candidates ran unopposed. These elections reveal little about either minority cohesion or white bloc voting. The district court found neither minority cohesion behind Hispanic candidates nor racially polarized voting in seven of the eleven elections in which Hispanic candidates ran against non-Hispanic opponents. The court found minority cohesion in the four remaining elections, and found white bloc voting only in the three that occurred before 1988. See Holyoke, 880 F. Supp. at 921-24.
Although the VRA does not require for a successful section 2 showing that minority-preferred candidates be members of the minority group, see Clarke v. City of Cincinnati, 40 F.3d 807, 810 n. 1 (6th Cir. 1994), cert. denied, ___ U.S. ___, 115 S.Ct. 1960 (1995); Sanchez v. Bond, 875 F.2d 1488, 1495 (10th Cir. 1989), cert. denied, 498 U.S. 937 (1990), elections in which minority candidates run are often especially probative on the issue of racial bloc voting. See, e.g., Jenkins, 4 F.3d at 1128; Magnolia Bar Ass'n, Inc. v. Lee, 994 F.2d 1143, 1149 (5th Cir.), cert. denied, ___ U.S. ___, 114 S.Ct. 555 (1993). But evidence exhumed from "white only" elections may still be relevant in assessing the totality of the circumstances in a vote dilution case, especially if it tends to rebut the evidence of cohesion or white bloc voting extracted from "mixed" elections. See Jenkins, 4 F.3d at 1128.
To be sure, the district court implied that blank ballots cast in three of these four elections (the 1989 and 1993 school committee elections in Ward 2, and the 1993 city council race in the same ward) evinced white bloc voting. See Holyoke, 880 F. Supp. at 923-24. But the record furnishes no foundation for the implication that white voters cast blank ballots as a protest against unopposed Hispanic candidates. In 1989, for example, the highest percentage of blank ballots was recorded in the precinct that had the highest percentage of Hispanic voters.
The district court also found that the minority community had cohered behind a non-Hispanic candidate, Elaine Pluta, in her successful 1991 bid for an at-large seat on the city council. In fact, Pluta ranked ahead of a Hispanic candidate on Hispanic voters' ballots. However, she received strong support from non-Hispanic voters as well; that segment of the electorate ranked her fifth (out of sixteen) among at-large candidates. Thus, while there may have been minority cohesion behind Pluta, the record reflects no evidence of white bloc voting against the candidate that minority voters preferred.
Viewed from a different angle, the court's finding that so few elections exhibited telltale signs of minority cohesion and/or white bloc voting seems to be tantamount to a finding that those characteristics were absent from approximately two-thirds of the analyzed elections. The finding also seems to contradict the district court's conclusion that the plaintiffs established the second and third Gingles preconditions. Of course, it is possible that the apparent contradiction can be explained away: we recognize that determining whether racial bloc voting exists is not merely an arithmetic exercise that consists of totting up columns of numbers, and nothing more. To the contrary, the district court should not confine itself to raw numbers, but must make a practical, commonsense assay of all the evidence. See Magnolia Bar Ass'n, Inc. v. Lee, 994 F.2d 1143, 1147 (5th Cir.), cert. denied, ___ U.S. ___, 114 S.Ct. 555 (1993); see also S. Rep. No. 417, supra, at 29-30 n. 118, reprinted in 1982 U.S.C.C.A.N. at 207-08. But allowing for the possibility of a sophisticated evaluative judgment does not dissipate the need to explain that judgment.
The district court, forced to juggle several issues at once, offered no explanation of this seeming contradiction. The court not only glossed over the raw numbers but also failed to clarify why evidence reflecting racially polarized voting in at most three or four elections (out of eleven) justified a finding of vote dilution. While we are unprepared to say, here and now, that such a finding is incorrect as a matter of law, we cannot accept it without a better articulated rationale. Thus, because we are unable to follow the district judge's thought processes in this regard, we must return the case to him for a more detailed explication of his reasoning. See Houston, 56 F.3d at 612-13 n. 8 (remanding because the "district court findings are too general to allow us to conduct our appellate review") (citing cases); Cousin, 46 F.3d at 575 (remanding because the "record fails to provide the bases for the district court reasoning"); Velasquez, 725 F.2d at 1021 (similar); cf. Westwego Citizens for Better Gov't v. City of Westwego, 946 F.2d 1109, 1119 (5th Cir. 1991) (ascribing error when district court's facially inconsistent findings were not explained).
We take this step reluctantly, mindful that district courts have heavy workloads and that appellate tribunals should not stand unduly on ceremony, but should fill in blanks in the district court's account when the record and the circumstances permit this to be done without short-changing the parties. See Applewood Landscape Nursery Co. v. Hollingsworth, 884 F.2d 1502, 1503-04 (1st Cir. 1989) (collecting cases). In this situation, however, the record does not lend itself to curing the omissions in this fashion.
We are fortified in this cautious approach by what we envision as the distinct possibility that the district court may have undervalued the import of Holyoke's rapidly changing political environment. During the decade analyzed by the court, 1983 to 1993, the embryonic Hispanic community grew to maturity, gathering both numbers and political muscle. Hispanic leaders mounted a "successful community-based voter registration drive" in the mid-1980s and boosted voter turnout dramatically. Holyoke, 880 F. Supp. at 922. In 1985, Holyoke voters elected an Hispanic to political office for the first time in Massachusetts' history. See id. at 921. Hot on the heels of this signal victory, the 1987 municipal elections witnessed the "most successful city-wide campaign ever run by an Hispanic in Holyoke." Id. at 922. Those elections also witnessed the last contest in which the district court supportably found white bloc voting. From that time forward, Hispanics have maintained political dominance over two wards and have represented those wards on both the school committee and the city council. See id. at 921-24.
Success is, of course, relative; the Hispanic candidate came close but nevertheless lost. While some might say that close only counts in horseshoes, hand grenades, and ballroom dancing, we think that progress of this sort, even short of an electoral win, is significant.
This rise in the Hispanic community's political fortunes is significant. The ultimate question in any section 2 case must be posed in the present tense, not the past tense. The court must determine whether the challenged electoral structure deprives a racial minority of equal opportunity to participate in the political process at present. Though past elections may be probative of racially polarized voting, they become less so as environmental change occurs. In particular, elections that provide insights into past history are less probative than those that mirror the current political reality. See LULAC, 999 F.2d at 891; Meek v. Metropolitan Dade County, 985 F.2d 1471, 1482-83 (11th Cir. 1993).
In this instance, the district court alluded to Holyoke's political evolution, see, e.g., Holyoke, 880 F. Supp. at 927, but does not appear to have given it weight in evaluating either the Gingles preconditions or the strength of any inference to be drawn therefrom. Under these circumstances, we think it is incumbent upon the court to explain more fully its view that vote dilution persists in spite of improved political conditions.
V. ADDITIONAL MATTERS
Because remand is required, we take this opportunity to comment briefly on two other areas of continuing interest.
First, the shortcomings we have catalogued in the district court's findings cloud the relationship between evidence of racially polarized voting in the ward elections and the trial court's conclusion that the at-large component of the electoral system unlawfully dilutes the Hispanic vote. Though we do not quarrel with the court's decision to consider evidence from the ward elections in analyzing racial polarization in the at-large elections — as we have indicated supra, a court has a duty to ponder all available evidence concerning racially polarized voting that promises to cast light on the factors at work in a particular electoral scheme, see, e.g., Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 502 (5th Cir. 1987) (approving use of data from exogenous elections when other evidence is sparse), cert. denied, 492 U.S. 905 (1989) — we remain at a loss, on the record as it stands, to comprehend why and how the court thought that the evidence from the ward elections informed the analysis of what had transpired in the contests for at-large seats on the city council. These questions, too, demand more specific findings. See Monroe v. City of Woodville, 881 F.2d 1327, 1330 (5th Cir.) (holding that when a trial court relies on information from exogenous elections, it should undertake fact-specific assessments of their relevance and probative worth), modified in other respects, 897 F.2d 763, cert. denied, 498 U.S. 822 (1990).
Second, we think that the district court, which made no reference to Ward 4 in its initial assessment, must meet head-on the City's contention that this ward (in which Hispanics comprise approximately 28% of the voting age population) constitutes a so-called influence district and therefore should be taken into account in evaluating whether Hispanic voting strength has been illegally diluted.
Although "society's racial and ethnic cleavages sometimes necessitate majority-minority districts to ensure equal political and electoral opportunity, that should not obscure the fact that there are communities in which minority citizens are able to form coalitions with voters from other racial and ethnic groups, having no need to be a majority within a single district in order to elect candidates of their choice." De Grandy, ___ U.S. ___, 114 S.Ct. at 2661; see also Chisom v. Roemer, 501 U.S. 380, 397 n. 24 (1991) (rejecting "the erroneous assumption that a small group of voters can never influence the outcome of an election"); Gingles, 478 U.S. at 87-88 (O'Connor, J., concurring) (intimating that a group's voting strength should be assessed with reference not only to its prospects for electoral success but also in terms of "other avenues of political influence"). These precedents merely confirm the lessons of practical politics: the 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. See Latino Political Action Comm., Inc. v. Boston, 609 F. Supp. 739, 747-48 (D.Mass. 1985), aff'd, 784 F.2d 409 (1st Cir. 1986); see also Rural W. Tenn. African-Am. Affairs Council, Inc. v. McWherter, 877 F. Supp. 1096, 1105 (W.D.Tenn.) (three-judge court) (holding that an influence district exists if a minority group constitutes at least one-quarter of the voting age population because the group then "ha[s] significant influence on candidates in virtually every election"), aff'd, 116 S.Ct. 42 (1995); City of Columbia, 850 F. Supp. at 429 (discussing minority's "shared influence" over at-large seats in districts where the minority comprises 40% of the total voting age population).
Although we are unwilling to prescribe any numerical floor above which a minority is automatically deemed large enough to convert a district into an influence district, we believe that when, as now, a minority group constitutes 28% of the voting age population, its potential influence is relevant to a determination of whether the group lacks a meaningful opportunity to participate in the electoral system. Accord McWherter, 877 F. Supp. at 1102. As is true of other factors, the district court should make a searching evaluation of the degree of influence exercisable by the minority, consistent with the political realities, past and present, and should enter its findings and conclusions as to how (if at all) the voting strength of Hispanics in Ward 4 affects the section 2 calculus.
In requiring that influence districts be considered in section 2 cases, we are guided by the Court's recent admonition that the VRA's goals include "eradicating invidious discrimination from the electoral process and enhancing the legitimacy of our political institutions." Miller v. Johnson, ___ U.S. ___, 115 S.Ct. 2475, 2494 (1995). These goals are poorly served by balkanizing electorates and carving them into racial fiefdoms. See id. Influence districts, on the other hand, are to be prized as a means of encouraging both voters and candidates to dismantle the barriers that wall off racial groups and replace those barriers with voting coalitions. In fine, influence districts bring us closer to "the goal of a political system in which race no longer matters." Shaw v. Reno, 113 S.Ct. 2816, 2832 (1993); see also De Grandy, ___ U.S. ___, 114 S.Ct. at 2661 (reflecting that "minority voters are not immune from the obligation 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").
Factoring influence districts into the calculus also helps ease the tension between Congress's desire to permit vote dilution claims to be brought under section 2 and its intent to avoid creating a right to proportional representation. See Gingles, 478 U.S. at 84 (O'Connor, J., concurring) (discussing "inherent tension between what Congress wished to do and what it wished to avoid").
It is important to realize that influence districts serve these goals only to the extent that they reflect a meaningful opportunity for minority voters to participate in the political process. Consequently, before the existence of an influence district is given significant weight in the balance, the evidence must reveal that minority voters in the district have in fact joined with other voters to elect representatives of their choice. Moreover, the record must show that elected representatives from such a district serve, at least in part, the interests of the minority community and vie for its support.
To recapitulate, the district court's opinion is well-considered and in many respects deftly navigates the marshy terrain of voting rights jurisprudence. Yet, at the risk of seeming unappreciative of a job well done, we believe the court's opinion lacks essential clarity in its factual findings. For one thing, the court neither acknowledges nor discusses critical evidence that appears to contradict its ultimate conclusion of vote dilution. For another thing, it never adequately explains the relevance of some evidence upon which it relies quite heavily to support this conclusion. And, finally, it omits any meaningful mention of potentially salient factors (such as influence districts). Rather than guess at the missing elements, we think that the course of prudence is to vacate and remand.
We leave the procedure to be followed on remand to the lower court's informed discretion, without endeavoring to set an outer limit on its range of options. See Lussier v. Runyon, 50 F.3d 1103, 1115 (1st Cir.), cert. denied, ___ U.S. ___, 116 S.Ct. 69 (1995). At a minimum, the court must discuss the evidence we have identified as troubling (or as possibly overlooked) and explain the relationship of this evidence to the issue of vote dilution. The court need not stop there, however; it is free to reopen the record, to take additional evidence, and/or to reconsider all (or any part) of its findings in light of the comments contained in this opinion. To this end, while we neither require nor anticipate an entirely new trial, the court in its discretion may permit the parties to supplement the existing record with additional facts (including, but not limited to, evidence gleaned from the new round of municipal elections that have recently been completed). See Charles Alan Wright Arthur R. Miller, Federal Practice and Procedure Section(s) 2577 (2d ed. 1995).
We are mindful that, in addition to the assignments of error that we have discussed, the City strenuously objects to the remedy fashioned by the court below. We do not address this objection today. If the district court, after further consideration, again finds that Holyoke's electoral structure violates section 2 of the VRA — and we do not intimate any expectancy in this regard — we anticipate that it will then revisit the question of how best to mold an appropriate remedy. Withal — and, perhaps, at the expense of remarking the obvious — we offer two brief bits of general guidance that may be helpful if this contingency materializes.
First, the court must be sure to analyze the question of remedy in light of any new findings that it makes on remand. Second, the court now has — and should take advantage of — the luxury of time. The court originally attempted to craft a remedy in time for the 1995 municipal elections. That cycle has turned, and the next is well in the future. Given this window of opportunity, the option of choice (assuming that the court finds a section 2 transgression) is to give the defendant the first chance to assemble a remedial plan. We think it is a fundamental tenet of voting rights law that, time permitting, a federal court should defer in the first instance to an affected state's or city's choice among legally permissible remedies. See Cane v. Worcester County, 35 F.3d 921, 927 (4th Cir. 1994), cert. denied, ___ U.S. ___, 115 S.Ct. 1097 (1995); Westwego, 946 F.2d at 1124.
If, and only if, the City fails to formulate a satisfactory remedial plan should the district court step in and fashion the appropriate anodyne ex proprio vigore. See Miller, ___ U.S. ___, 115 S.Ct. at 2488. It goes almost without saying that this authority must be exercised responsibly and with due attention to the Supreme Court's recent warnings about the social and political costs of dividing communities along racial lines in the name of improving electoral systems. See, e.g., Shaw, 113 S.Ct. at 2832 (observing that "[r]acial gerrymandering, even for remedial purposes, may balkanize us into competing racial factions").
Vacated and remanded. All parties will bear their own costs on this appeal .