Full opinion text
TJOFLAT, Chief Judge: Section 2(a) of the Voting Rights Act, 42 U.S.C. § 1973 (1988), states that “[n]o voting qualification or prerequisite to voting, or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner 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.” Under section 2(b) of the Act, “[a] violation of [section 2(a) ] is established if, based on the totality of the circumstances, it is shown that the political processes leading to nomination or election in the State ... are not equally open to participation by members of a class of citizens protected by [section 2(a) ] 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.” Section 2 applies to state judicial elections. Chisom v. Roemer, 501 U.S. 380, 404, 111 S.Ct. 2354, 2368, 115 L.Ed.2d348 (1991); Houston Lawyers’ Ass’n v. Attorney Gen., 501 U.S. 419, 423-24, 111 S.Ct. 2376, 2379, 115 L.Ed.2d 379 (1991). In this case, the appellants, black voters and an association of black attorneys, challenge the system used to elect the judges of Florida’s Fourth Judicial Circuit Court, which encompasses Duval, Clay, and Nassau counties, and the judges of the Duval County Court. The appellants contend that the use of at-large elections in those trial court jurisdictions dilutes the voting strength of the black minority in violation of section 2; they seek a remedy, such as the creation of sub-districts, that will ensure their ability to elect black judges of their choice. The appellees contend that the appellants are entitled to no relief. First, the appellants and the black voters they represent have suffered no racial vote dilution. Second, assuming that vote dilution exists, the relief the appellants seek would so alter the structure of the Fourth Judicial Circuit and Duval County courts as to undermine the ability of those courts to administer justice. Following a five-day bench trial, the United States District Court for the Middle District of Florida dismissed the appellants’ case. The court did so on two grounds: (1) the appellants failed to establish an essential element of a vote dilution case — racially polarized voting in the relevant communities— as required by Thornburg v. Gingles, 478 U.S. 30,106 S.Ct. 2752, 92 L.Ed.2d 25 (1986); and (2) the appellees’ rebuttal evidence demonstrated that, under “the totality of the circumstances,” racial discrimination was not playing a role in the judicial elections under challenge. The district court’s finding of no section 2 liability made it unnecessary for the court to consider the matter of remedy. On appeal, a panel of this court concluded that the evidence before the district court demonstrated racially polarized voting in the Fourth Judicial Circuit and Duval County as a matter of law, and thus a section 2 violation. Without considering, as required by Houston Lawyers’ Ass’n, 501 U.S. at 426, 111 S.Ct. at 2380-81, the appellees’ argument that the relief the appellants sought would undermine the administration of justice in those jurisdictions, the panel remanded the case to the district court for the imposition of a remedy. Nipper v. Smith, 1 F.3d 1171, 1184 (11th Cir.1993). The ease is now before us on rehearing en banc. Nipper v. Smith, 17 F.3d 1352 (11th Cir.1994). To determine whether the district court erred in holding that the appellants failed to establish a case of vote dilution, we must address a question the Supreme Court has not decided and our divided en banc court in Solomon v. Liberty County, 899 F.2d 1012 (11th Cir.1990) (en banc) (per curiam), cert. denied, 498 U.S. 1023, 111 S.Ct. 670, 112 L.Ed.2d 663 (1991), has precluded us from answering: whether section 2 plaintiffs, in order to establish that the challenged electoral scheme is diluting their right to vote “on account of [their] race or color,” must demonstrate that their diminished opportunity to participate in the political process and to elect representatives of their choice is being caused by the interaction of racial bias in the voting community and the challenged scheme. Without an answer to this question, we cannot determine what evidence is relevant, and the weight it should be accorded, in the totality of the circumstances inquiry of section 2(b). Specifically, we cannot determine the weight to be accorded the state policies underlying the challenged judicial electoral scheme. Houston Lawyers’ Ass’n, 501 U.S. at 426, 111 S.Ct. at 2380-81. Nor can we determine, given those policies and the structure of the courts involved, whether the remedy sought is feasible. In part I of this opinion, we set forth the facts and procedural history of the case. In part II, we address the question left unanswered in Solomon: whether the existence of racial bias in the voting community necessarily forms the basis of a section 2 violation. We hold that the totality of the circumstances must demonstrate that the voting community is driven by racial bias and that the electoral scheme in question permits that bias to dilute the plaintiff minority’s voting strength. In part III, after noting several important differences between judicial and legislative elections, we discuss the factors, including the policies advanced by Florida’s method of selecting its trial court judges, involved in the totality of the circumstances analysis in judicial election eases. In parts IV and V, we apply the principles set forth in parts II and III to the case at hand. In part IV, we examine the district court’s finding of no vote dilution and conclude, as did the panel, that the appellants established a case of dilution. In part V, we consider the matter of remedy. We find that the type of relief the appellants seek would undermine the administration of justice in the trial courts at issue; we therefore affirm the district court’s denial of section 2 relief. I. Voting rights cases are inherently fact-intensive, particularly those section 2 vote dilution claims alleging that, due to the operation of a challenged voting scheme, minority voters are denied an equal opportunity to participate in the political process and to elect representatives of their choice. In such cases, courts must conduct a “searching practical evaluation of the ‘past and present reality’ ” of the electoral system’s operation. Gingles, 478 U.S. at 45, 106 S.Ct. at 2764 (quoting S.Rep. No. 417, 97th Cong., 2d Sess. 30 (1982), reprinted in 1982 U.S.C.C.A.N. 177, 208). Accordingly, because a claim of vote dilution must be evaluated with a functional, rather than a formalistic, view of the political process, the Supreme Court has emphasized the importance of “ ‘an intensely local appraisal of the design and impact’ ” of the electoral structure, practice, or procedure at issue. Id. at 79, 106 S.Ct. at 2781 (quoting Rogers v. Lodge, 458 U.S. 613, 621, 102 S.Ct. 3272, 3278, 73 L.Ed.2d 1012 (1982)). Given the necessity of a nuanced understanding of the election system in evaluating a claim of racial vote dilution, we set out the factual background of this case in some detail. A. Florida’s judiciary employs two tiers of trial courts. The circuit courts have general jurisdiction over civil and criminal cases while the jurisdiction of the county courts is restricted to certain statutorily defined classes of misdemeanor and small claims cases. Fla. Const, art. 5, §§ 5-6 (West Supp.1994); Fla.Stat.Ann. §§ 26.012, 34.01 (West 1988 & Supp.1994). Florida currently has twenty circuit courts (designated by number); they range from single-county circuits (in highly populated areas like Miami or Tampa) to circuits comprised of as many as six or seven counties (many of which are rural). See Fla.Stat.Ann. § 26.021 (West 1988). Each circuit judge has jurisdiction or authority within the entire circuit; similarly, the jurisdiction of each judge of the county court extends throughout the county. Although the circuit and county courts have multiple members, like most trial courts they do not operate as collegial bodies; rather, the judges exercise independent judicial authority, engaging in coordinated decision-making only for the handling of some administrative matters. Circuit judges serve six-year terms and county judges serve four-year terms, Fla. Const, art. 5, § 10; in all other respects, the office of judge on the two benches is virtually identical. The judges of the circuit and county courts are elected at-large in nonpartisan circuitwide and countywide elections, respectively; they must reside within the territory of the court on which they serve. The at-large judicial elections for both courts are characterized by a majority vote requirement, a numbered place system (meaning that candidates must run for a particular seat on the court), and the use of staggered terms. In order to qualify to run for a judgeship, a candidate must pay a filing fee (although a candidate can avoid the fee by obtaining a number of signatures from registered voters on a petition), see Fla.Stat.Ann. §§ 105.031(3) & .035 (West 1992), and must have been a member in good standing of The Florida Bar for at least five years immediately preceding the election in question, see Fla. Const, art. 5, § 8. In the event of a midterm vacancy on either court, the governor appoints a replacement to serve until the next general election. Id. § 11. In the instant case, the appellants challenge the electoral system for both trial courts in and around Jacksonville, Florida. The Fourth Judicial Circuit is located in northeast Florida and consists of Duval, Clay, and Nassau Counties. Duval County encompasses the City of Jacksonville; the county and city governments were consolidated in 1968. According to the 1990 census, the total population of the Fourth Circuit is 822,928; the black population is 173,937, or approximately twenty-one percent of the total. At the time of trial in 1991, the Fourth Circuit had twenty-eight circuit judges, only one of whom was black. The total population of Duval County (as of the 1990 census) is 672,971, of which 163,902, or approximately twenty-four percent, are black. Two of the twelve judges on the Duval County Court at the time of trial were black judges. Although the voting age populations of the Fourth Circuit and Duval County are nineteen and twenty-two percent black, respectively, no black candidate for a judgeship on either court has ever been elected to office in a contested election. Every black judge who has served on one of these courts originally reached office by way of a mid-term appointment to a vacant seat. Although forming a powerful initial sketch of the role played by race in the relevant judicial elections, this bare description of circuit and county court election results paints only part of the total picture. To place the elections at issue in this litigation into their proper context, we first detail the historical evolution of the Florida scheme of judicial elections. We then describe the operation of the current model since its adoption in 1972 and examine the special circumstances and relative infrequency with which black candidates have sought judgeships on the Fourth Circuit and the Duval County Court. B. . Article 5 of the Florida Constitution, which governs the judicial branch of the state government, has evolved over the last twenty-five years, thereby changing the method of selecting state court judges dramatically. For much of this century, judicial positions in Florida were filled through partisan elections. Florida voters adopted a revised constitution in 1968, but the new version temporarily retained the old system of partisan judicial elections. In 1971, the Florida legislature adopted its first nonpartisan electoral system for the state judiciary. The statute, the contemporary counterpart of which is now codified as Fla.Stat.Ann. § 105.011(2) (West 1992), provided essentially that “[a] judicial office is a nonpartisan office, and a candidate for election or retention thereto is prohibited from campaigning or qualifying for such an office based on party affiliation.” At that time, the process of selecting judges became fundamentally different from the methods used to choose other elected state officials. Article 5 was significantly amended in 1972, although the revised version continued to provide that all judges would be elected “by vote of the qualified electors within the territorial jurisdiction of their respective courts.” When a vacancy occurred on any court of the state, however, the governor would make an interim appointment until a permanent replacement for the remainder of the predecessor judge’s term could be elected. Although the ultimate decision regarding such an interim appointment was entrusted to the discretion of the governor, the appointee nonetheless had to be selected from a slate of not fewer than three candidates submitted to the governor by a judicial nominating commission. One commission was created for each court in the state to accept applications from interested eligible attorneys when vacancies arise and to make recommendations thereon. The nominating commission system was designed to inaugurate a merit selection process for members of the state judiciary through the submission of qualified nominees to the governor. The system of judicial elections currently in place was created by a further amendment to Article 5 that was adopted by the citizens of Florida at the general election in November of 1976. At that time, Florida adopted the “Missouri Plan” for the selection of supreme court justices and the judges of the district courts of appeal, the intermediate appellate courts of the Florida judiciary. Initial selection of those justices and judges was transferred away from the electorate to the nominating commissions and to the governor. The new selection process empowers the governor to fill any vacancy (not just an unexpired term) on the supreme court or the district courts of appeal, the only condition being that the selection must be made from a slate of “three persons nominated by the appropriate judicial nominating commission.” Fla. Const, art. 5, § 11(a). These initial appointments are for partial terms of not less than one year, id., at the end of which each justice or judge “may qualify for retention [and a full six-year term] by a vote of the electors.” Id. § 10(a). The justices and judges of Florida’s appellate courts then stand for retention at the end of every term. The retention elections, however, do not involve challenges to the incumbent by another competitor for office; rather, each candidate for retention under this approach runs against his or her established record of performance. The 1976 constitutional amendments, however, did not alter the method of selecting circuit and county court judges. Therefore, under the present system, which constitutes a partial “Missouri Plan,” judges of the circuit and county courts continue to be elected (in nonpartisan elections) to six- and four-year terms, respectively, by vote of the qualified electors in their jurisdictions. Id. § 10(b). When a mid-term vacancy occurs, the governor appoints a replacement from a slate provided by the appropriate nominating commission, with a single commission handling nominations for vacancies on a circuit court as well as the county courts in that circuit’s territory. Id. § 11. As a result, subsequent elections, even at the trial court levels, are affected by the merit selection of interim appointments by the governor upon the recommendation of the judicial nominating commissions; the momentum gained from initial selection by such an independent review panel has a telling effect upon later races in which the incumbent is a candidate. Under the current system of electing trial court judges in Florida, therefore, merit selection plays a part in all but a few: those to fill vacancies on the circuit or county court benches that occur at the end of a term— when, for example, a judge retires and two challengers compete to fill his or her seat. These changes in Florida’s constitution were clearly designed to eliminate the vices of partisan, electoral politics from the process of selecting state court judges. The goal of merit selection of judges, naturally, is to insulate them from popular pressure and to make them more willing to decide an unpopular case fairly and impartially while, at the same time, raising the level of qualifications of judicial officers. C. A total of five black candidates have run for seats on either the Fourth Circuit or the Duval County Court in six elections since 1972. We now highlight the circumstances surrounding each of those electoral races while describing the methods by which the judges on those courts have generally reached office during the same period. The first of the elections relied on by the parties in litigating this case predates the inauguration of Florida’s partial merit selection system for trial court judges. In 1972, Leander Shaw became the first black candidate for a judgeship on a court at issue in this litigation. A black lawyer who now serves as a justice of the Florida Supreme Court, Shaw ran for a vacant seat on the Fourth Circuit bench. In the nonpartisan primary, he was opposed by two white candidates, John S. Cox and George L. Proctor. Shaw and Cox advanced to a runoff election, which Cox ultimately won. The Shaw election was one of only four Fourth Circuit electoral races that were contested during the 1972 election cycle; the other fifteen Fourth Circuit judgeships at stake in that year’s elections were filled by candidates who ran unopposed. The parties in this case introduced a wealth of statistical evidence that indicates how Florida’s scheme of trial court elections has operated in the Jacksonville area under the current model. A pattern readily emerges that has prevailed throughout the period at issue in this case: relatively few judicial elections have been contested in the Fourth Circuit and in Duval County. Between January 1,1973, when the current system for selecting circuit and county court judges went into effect, and the end of 1990, the last year for which information was available at the time of trial, the record reflects that seventy-eight circuit judgeships were filled by election. Of those positions, only ten (or approximately thirteen percent) were contested; the rest were filled by candidates who ran unopposed. Only one black candidate ran for a seat on the Fourth Circuit bench during this period: Harrell T. Buggs in 1978. In the primary, Buggs and two white candidates challenged an incumbent white judge, Lawrence D. Fay, who had been appointed to the court through the merit selection process approximately one hundred days before the election. Fay and one of the white challengers, John E. Palmer, advanced to the runoff; Buggs, who had obtained over eighty percent of the black vote but only approximately four percent of the white vote, was eliminated in the primary. Focusing solely on the elections described above, however, paints an incomplete picture of how judges come to the circuit court bench in the Fourth Circuit. According to the Du-val County Clerk’s Office, a total of fifty-four judges have served on the Fourth Circuit since 1972. Fourteen were holdovers from the pre-reform system (six of whom were appointed by the governor before the merit selection process was instituted and eight of whom were elected). With respect to the remainder, twenty-eight circuit court judges began their judgeships after being appointed through the commission system although only twelve originally were elected in nonpartisan elections. Furthermore, although it is true that no black candidates have been elected to judicial office in the Fourth Circuit, two black judges have been appointed under the nominating commission system (Judges Adams and Davis, see supra note 6). The increasing reliance on the appointment process for selecting trial court judges is highlighted by an examination of the composition of the current Fourth Circuit Court. Of the twenty-eight judges presently sitting on the court, nineteen were appointed through the commission process, eight initially were elected in nonpartisan elections, and one was appointed by the governor in 1967 under the old system. Moreover, at the time of trial, only one of the judges then serving on the Fourth Circuit bench had ever been opposed for reelection (Judge Fay by Harrell Buggs in 1978), although the twenty-eight judges collectively had stood for reelection over twenty times. A similar pattern emerges from an examination of the county court selection process. From 1973 to 1990, fifty-six judgeships on the Duval County Court were decided by election; only twelve (or approximately twenty-one percent) of those electoral races were contested. Black candidates participated in primaries for three of those elections and were defeated each time. As the district court noted, however, special circumstances surrounded two of these three races. Alfred Washington, a black candidate, lost his primary race for a seat on the Duval County Court in 1978 to Giles P. Lewis, despite garnering eighty-eight percent of the black vote. Prior to the election, however, Washington had resided in the community for only a short period of time whereas Lewis was a longtime resident with an established legal practice. In 1984, two black candidates sought positions on the county court bench. Denise Prescod was defeated in her primary by incumbent, Edward P. Westberry, even though she garnered a similar proportion of the black vote. Judge Westberry was not a lawyer, but he was permitted to serve on the court under a grandfather clause in the 1972 constitutional revision that allowed former justices of the peace to become county court judges. He had, by all accounts, received consistently poor ratings from the local bar but was acknowledged to be a vigorous and extremely effective campaigner. Judge Westberry also had been a judicial officer since 1966, thus claiming the mantle of experience. Prescod, a lawyer, received several important endorsements, including those from Jacksonville’s only daily newspaper and the public school teachers’ organization. At the time she ran for judicial office, however, Prescod had been admitted to The Florida Bar for only two years. Under the current five-year experience requirements, Prescod would not have been eligible to run for county court judge in 1984. The district court concluded that “neither candidate’s qualifications ... were particularly helpful to their respective campaigns, and in all probability incumbency ultimately proved to be the deciding factor.” Nipper, 795 F.Supp. at 1542 n. 18. In the remaining county court election involving a black candidate and the only one in which the black candidate did not challenge an incumbent, Dietra Micks won three-quarters of the black vote but failed to make the runoff for her position against a field of three white candidates. Again, this time in the county court context, looking solely at elections tells only part of the story. Since 1972, a total of thirty judges have served on the Duval County Court. Ten initially were appointed by the governor under the commission nominating process while twelve originally obtained their positions by winning nonpartisan elections. Nine other county court judges were holdovers from the prior (pre-1972) system. As for the twelve county court judges serving at the time of trial, five were originally appointed to the bench, including both of the black judges. As the district court concluded, the evidence of black electoral success in the jurisdictions at issue in this case is mixed at best. The court found: The black circuit and county judges who have faced reelection have been unopposed and, of course, successful in their reelection bids. In a similar judicial election, Joseph Hatchett’s 1976 Supreme Court [reflection, the incumbent Hatchett defeated a white candidate. Hatchett garnered a majority of the white vote in both the Fourth Judicial Circuit and Duval County. No black candidate has won a contested election for circuit or county judge at least since 1972. Id. at 1537. In addition, a review of nonjudicial elections in the Jacksonville area reveals a dearth of black electoral success: “[N]o black candidate has won a contested circuitwide or countywide election for any office since 1979,” and only one city council member and three members of the civil service board in Jacksonville appear to have done so prior to that time. Id. Black candidates fared better in the Fourth Circuit appointment process. Henry Coxe, the chairman of the Fourth Circuit Judicial Nominating Committee from 1987 to 1991, testified at trial concerning minority participation in the appointment process during his tenure on the committee. Based on that testimony, the district court found: Statistics provided by the [Fourth Circuit Judicial Nominating] Commission indicate that from 1987 to 1991, the Commission accepted applications for eleven circuit and county court vacancies. Of the total number of applicants, approximately eleven percent were black attorneys. An almost identical percentage of blacks were among the total number of nominees sent by the Commission to the Governor. Nine percent of the total appointments made by the Governor during this period were black (specifically, the county judge appointment in the summer of 1991). Id. at 1538. Moreover, the court emphasized, “[i]n judicial elections, incumbency and name recognition are the primary factors behind electoral success.” Id. Several witnesses testified to this effect. Indeed, only one of the twenty-eight incumbent circuit judges serving at the time of trial had ever faced electoral opposition subsequent to his or her initial appointment or election; none of the twelve incumbent county judges had ever been opposed in a reelection bid. Judge Adams agreed that incumbency is an important factor in determining the outcome of judicial elections and testified revealingly that “I don’t think that a black could win a county-wide judicial election without the benefit of incumbency.” By all accounts, the electoral benefits of incumbency accrue equally to black and white judges as none of the black judges on either the Fourth Circuit or Duval County Court benches has ever been opposed for reelection. D. The complaint in this case was filed in the United States District Court for the Middle District of Florida on July 5, 1990. The plaintiffs (the appellants here) challenged the method of electing judges for the Fourth Circuit and the Duval County Court under section 2 of the Voting Rights Act as well as under the Fourteenth and Fifteenth Amendments of the Constitution; they alleged that those election systems unlawfully diluted the electoral strength of black voters. Trial originally was scheduled for March 1991, but the district court stayed the proceedings pending the outcome of cases before the United States Supreme Court concerning the applicability of section 2 to judicial elections. After the Court confirmed that section 2 applies to the election of state court judges, see Chisom v. Roemer, 501 U.S. 380, 404, 111 S.Ct. 2354, 2368, 115 L.Ed.2d 348 (1991); Houston Lawyers’ Ass’n v. Attorney Gen., 501 U.S. at 419, 423-24, 111 S.Ct. 2376, 2379,115 L.Ed.2d 379 (1991), the trial of this ease was reset and commenced on December 12, 1991. To resolve the issue of whether voting in circuit and county judicial elections was racially polarized, both sides offered expert testimony and statistical evidence. The appellants relied on the testimony of Dr. Allan Lichtman, Professor of History at American University in Washington, D.C. The appel-lees offered two expert witnesses: Dr. Ronald Weber, Professor of Government at the University of Wisconsin, and Dr. Joan Ha-worth, former Professor of Economics at Florida State University and currently the president of a private consulting company. Although the ultimate conclusion reached by the two sides was different, the experts agreed on most of the foundational data. Indeed, all of these experts used the same or similar statistical techniques to analyze the election results at issue. The experts divided sharply, however, over which types of elections should be analyzed in order to determine whether voting in Fourth Circuit and Duval County judicial elections has been, and continues to be, racially polarized. The appellants’ expert, Dr. Lichtman, studied the six circuit and county judicial elections in which black candidates have competed against white candidates since 1972. He also analyzed numerous elections for offices not at issue in this litigation (referred to as “exogenous” elections) that involved black candidates, placing particular emphasis on Justice Hatchett’s 1976 Florida Supreme Court reelection bid. Almost all of the other exogenous elections studied by Dr. Lichtman were of a partisan nature (including Jesse Jackson’s presidential primary campaigns in 1984 and 1988 and Alcee Hastings’ various bids for statewide office). Based on his review of these elections, Dr. Liehtman concluded that voting in the Fourth Circuit and in Duval County is racially polarized in judicial elections: Black voters prefer black candidates by overwhelming margins, but those candidates are not elected because they do not receive sufficient crossover votes from the white electorate. According to the ecological regression studies, a large majority of black voters strongly supported the black candidate in each of the six circuit and county judicial elections. At the same time, the majority of white voters supported the white candidate, thereby resulting in the defeat of all of the black candidates except for Leander Shaw in the 1972 primary. Shaw, however, went on to lose in the general election. In these six elections, black support for the black candidates ranged from seventy-three to ninety-eight percent, while white support for those same black candidates varied from three to thirty-three percent. Black voter support for the white candidates never exceeded twenty-seven percent while white support for the white candidates ranged from sixty-seven to ninety-seven percent. The district court adopted Dr. Lichtman’s statistical estimates concerning these six elections as findings of fact. Nipper, 795 F.Supp. at 1534 n. 4. The principal expert witness for the defense, Dr. Weber, also reviewed the six circuit and county court campaigns in which black candidates participated. Dr. Weber also examined a set of elections that Dr. Lichtman had not. Declining to confine his analysis of Fourth Circuit and Duval County judicial elections to those races in which a black candidate had competed, Dr. Weber instead examined all of the judicial elections in the relevant jurisdictions between 1972 and 1990, even those in which only white candidates participated. In further refining his study sample, Dr. Weber “argued that only those non-judicial elections for low-visibility offices similar to judgeships would be of importance in a judicial elections case.” Id. at 1533. Accordingly, he criticized Dr. Lichtman’s reliance on high-profile elections for statewide or national office. Dr. Weber concluded that the candidate of choice of black voters wins elections in the Fourth Circuit and Duval County because of the high degree of black cohesion and sufficient white crossover voting — but he admitted on cross-examination that his conclusion should be qualified as follows: “These data show very clearly that black candidates of choice are regularly elected in both jurisdictions and white voters do not usually vote as a bloc to deny black voters the opportunity to elect candidates of choice to these judicial posts— so long as these candidates are white and not black.” There were a total of nineteen contested elections for circuit judge between 1972 and 1990, of which sixteen involved only white candidates. Dr. Weber’s analyses, which the district court “adopt[ed] as findings of fact for statistical purposes only,” id. at 1534, revealed that the candidate of choice of black voters won thirteen, or sixty-eight percent, of these contested elections. During this same time period, there were twenty-four contested Duval County Court elections, three of which involved black candidates. Dr. Weber’s study revealed that the candidate of choice of black voters won fourteen, or fifty-eight percent, of those elections. A somewhat different picture emerges, however, when “split preference” statistics are examined. Looking at both circuit and county court races, Dr. Lichtman testified that the candidate preferred by black voters differed from the candidate preferred by white voters in nineteen of the thirty-seven contested judicial elections that involved only white candidates. In those “split preference” elections, the black voters’ preferred candidate lost eighty-four percent of the time. At trial, the appellants also introduced non-statistical evidence of Florida’s history of discrimination against black citizens, including its legacy of disfranchisement and segregation in most areas of life. In particular, Florida employed various franchise restrictions — from the poll tax to the white primary — for decades in an attempt to restrict the access of black voters to the ballot. Those classic discriminatory devices have been eliminated, however, and the consensus at trial was that there is currently little disparity in voter registration in the Fourth Circuit and in Duval County by race. Blacks of voting age are registered at higher rates than whites of voting age in both the Fourth Circuit and Duval County. Despite the removal of overt badges of segregation, the district court nonetheless found that “black citizens in Florida still suffer in some ways from the effects of Florida’s history of purposeful discrimination,” particularly in terms of socio-economic disparities, such as family income and high school graduation rates. Id. at 1536. Black citizens in the region covered by the Fourth Circuit have lower median incomes than whites and are more likely to be unemployed and to fall below the poverty line. In addition, the limited evidence presented at trial (reflected in a consensus among the experts) suggested that, although little disparity exists in voter registration, black voter turnout appears to be slightly lower than white turnout. And the “rolloff’ effect — which measures the number of voters who sign in at the polls but fail to cast a vote for a particular election on the ballot — is greater among black voters than white voters. The appellants also introduced into evidence a report issued in 1990 by the Florida Supreme Court Racial and Ethnic Bias Study Commission that, according to the district court, “documented numerous features of Florida’s justice system that allegedly have an adverse effect on the dispensation of justice to minority citizens, including the under-representation of minorities in the judiciary in comparison to the percentage of minorities in the total population.” Id. at 1535. The Study Commission concluded: Clearly, the current election process, which provides for circuit-wide, at-large elections, is not yielding sufficient representation of minorities on Florida’s bench. At the same time, the dramatic underrepresentation of minority judges reflected in the ... statistics compels the conclusion that the appointive system, as currently structured and implemented, has itself failed to achieve racial and ethnic diversity. The Commission strongly believes that serious measures need to be considered for implementation — in both systems — which are aimed at producing a more racially and ethnically sensitive judiciary. To that end, the Study Commission recommended that the state legislature study the feasibility of utilizing subdistricts for judicial elections as a means of redressing prior discrimination and increasing minority representation on the bench. The appellants established that the low number of minority judges on the trial courts cited by the Study Commission affects the perception, if not the reality, of the system’s fairness; all of the witnesses acknowledged, however, that the judges of the Fourth Circuit and the Duval County Court are fair and impartial in their administration of justice without regard to race. E. The record in this case therefore reveals a somewhat conflicted picture of the judicial selection process in the Fourth Circuit and Duval County. On the one hand, all of the black candidates for judicial office in those jurisdictions have been defeated by their white opponents. Several of the elections involving black candidates also involved other special circumstances, such as incumbency, that cloud the issue of whether race was the factor most responsible for the outcome. On the other hand, black judges have been appointed to the bench under the merit selection nominating commission system roughly in proportion to the number of minority applicants; once in office, moreover, those black judges have enjoyed the same benefits of incumbency as their white counterparts (including standing unopposed for reelection). The influence of the merit appointment system has been significant: Since its institution in 1972, more judges initially have reached the bench by appointment than by election. Based upon this evidence, the district court issued an order containing findings of fact and conclusions of law on June 2, 1992. In that order, the court rejected all of the appellants’ statutory and constitutional claims and denied relief. Id. at 1548. The court first ruled against the appellants on the critical issue of racially polarized voting, concluding that the statistical evidence, viewed in light of the special circumstances surrounding the judicial elections in question, failed to demonstrate sufficient racial bloc voting such that the white majority usually defeats the minority’s preferred candidate. Id. at 1543. The court then concluded alternatively that: [Ejven if Plaintiffs had made a sufficient showing of racial polarization, Defendants have offered overwhelming proof of objective factors in rebuttal that demonstrate, under the totality of the circumstances, that the social conditions in the Fourth Judicial Circuit and Duval County are such that their interaction^] with the electoral scheme do not, and will not, result in voting discrimination in the judicial elections under challenge. Id. at 1548. Judgment was entered for the appellees that same day. In this appeal, the appellants challenge only the district court’s denial of their section 2 claim; this court is not asked to review the court’s rulings on the constitutional issues. In an opinion issued on September 15, 1993, a panel of this court reversed, and remanded the case to the district court for the imposition “forthwith [of] an appropriate remedy.” Nipper v. Smith, 1 F.3d 1171 (11th Cir.1993). Much of the panel’s opinion was devoted to evaluating the evidence before the district court on the issue of racially polarized voting. According to the panel opinion, the district court had discounted the appellant’s statistical evidence (which the district court had acknowledged would ordinarily be sufficient to establish racially polarized voting) for three reasons: because “(1) the judicial elections involving black candidates were stale; (2) black voters occasionally elected candidates of choice in elections involving only white candidates; and (3) two of the black-white elections involved incumbents.” Id. at 1178-79. The panel “rejected] as clear error each of the district court’s reasons for concluding that appellants failed to show racial polarization” and held that “the evidence in the entire record demonstrates racially polarized voting in the Fourth Judicial Circuit and Duval County.” Id. at 1181-82. In the panel’s view, the district court also “erred in ruling that appellees discharged their burden of proving that the voting communities were not driven by racial bias”; specifically, the appellees had not shown “an absence of racially biased voting in the Fourth Judicial Circuit and Duval County.” Id. at 1184. Subsequently, we vacated the panel opinion and granted rehearing en banc. Nipper v. Smith, 17 F.3d 1352 (11th Cir.1994). II. Section 2 of the Voting Rights Act, adopted in 1965 and amended in 1982, outlaws election practices that result in racial discrimination. Of particular concern are electoral structures, such as at-large elections in areas with white majorities, that produce racial vote dilution because “[t]he right to vote can be affected by a dilution of voting power as well as by an absolute prohibition on casting a ballot.” Allen v. State Bd. of Elections, 393 U.S. 544, 569, 89 S.Ct. 817, 883, 22 L.Ed.2d 1 (1969) (emphasis added). Since racial and ethnic groups, tend, at times, to be geographically compartmentalized in our society, a districting plan drawn without regard to the physical distribution of such groups may nevertheless operate to distort their relative voting strengths, as well as to provide the opportunity for subtle discrimination through the manipulation of electoral structures. Vote dilution cases involve allegations that the location of district lines or the use of other electoral practices, such as at-large voting, may ‘“interact with social and historical conditions,’ [thereby] impair[ing] the ability of a protected class to elect its candidate of choice on an equal basis with other voters.” Voinovich v. Quitter, — U.S. -, -, 113 S.Ct. 1149, 1155, 122 L.Ed.2d 500 (1993) (quoting Thornburg v. Gingles, 478 U.S. 30, 47, 106 S.Ct. 2752, 2764, 92 L.Ed.2d 25 (1986)). Cases alleging a distortion of group voting power of this type have been termed “qualitative” (as opposed to quantitative) reapportionment cases because they focus “not on population-based apportionment but on the quality of representation.” Whitcomb v. Chavis, 403 U.S. 124, 142, 91 S.Ct. 1858, 1868, 29 L.Ed.2d 363 (1971). To prevail on a claim of vote dilution under section 2, plaintiffs generally must meet certain threshold requirements that the Supreme Court first identified in Gingles. Specifically, plaintiffs in vote dilution cases must establish as a threshold matter: (1) that the minority group is “sufficiently large and geographically compact to constitute a majority in a single-member district”; (2) that the minority group is “politically cohesive”; and (3) that sufficient racial bloc voting exists such that the white majority usually defeats the minority’s preferred candidate. Gingles, 478 U.S. at 50-51, 106 S.Ct. at 2766-67; Growe v. Emison, — U.S. -, -, 113 S.Ct. 1075, 1084, 122 L.Ed.2d 388 (1993) (quoting Gingles). The second and third of these threshold preconditions, or factors, relate to the question of whether the challenged electoral scheme is abridging the right of the plaintiff minority group “to vote on account of race or color.” The first precondition, or factor, asks whether the court can fashion a remedy for a demonstrated abridgement. In addition to the second and third preconditions, the Gingles Court identified other factors that may, in “the totality of the circumstances,” support a claim of racial vote dilution. Derived from the Senate Report accompanying the 1982 amendment to section 2, those factors include: 1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process; 2. the extent to which voting in the elections of the state or political subdivision is racially polarized; 3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, anti-single shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group; 4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process; 5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process; 6. whether political campaigns have been characterized by overt or subtle racial appeals; [and] 7. the extent to which members of the minority group have been elected to public office in the jurisdiction. Gingles, 478 U.S. at 37, 106 S.Ct. at 2759 (quoting S.Rep. No. 417, at 28-29, reprinted in 1982 U.S.C.C.A.N. at 177, 206-07). Additional factors that may be probative of vote dilution in some eases are: whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group; [and] whether the policy underlying the state or political subdivision’s use of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous. Id., at 37, 106 S.Ct. at 2759. The Gingles Court emphasized that “this list of typical factors is neither comprehensive nor exclusive,” id. at 45, 106 S.Ct. at 2763, although evidence of the third precondition, racially polarized voting, is “the linchpin of a § 2 vote dilution claim,” Citizens for a Better Gretna v. City of Gretna, 834 F.2d 496, 499 (5th Cir.1987), cert. denied, 492 U.S. 905, 109 S.Ct. 3213, 106 L.Ed.2d 564 (1989), and thus forms the basis of the Gingles threshold inquiry on the issue of liability (as opposed to remedy). Regardless of the particular factors that the plaintiff in a vote dilution case is able to demonstrate, however, the essence of a section 2 claim is that certain electoral characteristics, in conjunction with “social and historical conditions,” operate to eviscerate the ability of minority voters to elect their candidates of choice. Gingles, 478 U.S. at 47, 106 S.Ct. at 2764; see also Carrollton Branch of the NAACP v. Stallings, 829 F.2d 1547, 1555 (11th Cir.1987), cert. denied, 485 U.S. 936, 108 S.Ct. 1111, 99 L.Ed.2d 272 (1988). In 1990, this court divided on the issue of whether voting rights plaintiffs can establish a section 2 violation merely by satisfying the second and third Gingles threshold factors or whether the ultimate inquiry is always under the totality of the circumstances, such that the defendants have the opportunity to rebut the plaintiffs’ showing of vote dilution — and thereby avoid liability — by demonstrating a lack of racial bias in the voting community. See Solomon v. Liberty County, 899 F.2d 1012, 1017 (11th Cir.1990) (en banc) (per curiam) (division detailed in the concurring opinions of Kravitch, J., and Tjoflat, C.J.), cert. denied, 498 U.S. 1023, 111 S.Ct. 670, 112 L.Ed.2d 663 (1991). The Supréme Court’s recent decision in Johnson v. De Grandy, — U.S. -, 114 S.Ct. 2647, 129 L.Ed.2d 775 (1994), has essentially resolved this- issue as the following discussion indicates. A. Proof of the three core factors emphasized in Gingles is necessary, Gingles, 478 U.S. at 50, 106 S.Ct. at 2766, but not always sufficient, to establish a claim for relief under -section 2. Rather, plaintiffs in such cases also must show that, under the totality of the circumstances, “they do not possess the same opportunities to participate in the political process and elect representatives of their choice enjoyed by other voters.” League of United Latin Am. Citizens, Council No. 4434 v. Clements, 999 F.2d 831, 849 (5th Cir.1993) (en banc) (“LULAC”), cert. denied, — U.S. -, 114 S.Ct. 878, 127 L.Ed.2d 74 (1994). In Gingles, the Court stated that, “[a]s both amended § 2 and its legislative history make clear ... the trial court is to consider the totality of the circumstances and to determine, based upon a searching practical evaluation of the past and present reality, whether the political process is equally open to minority voters.” Gingles, 478 U.S. at 79, 106 S.Ct. at 2781 (internal quotation marks and citation omitted). The Gingles threshold factors (specifically, the second and third) constitute “essentially a ‘gloss’ on the Senate factors and a limitation on the interpretation of those factors in proving a vote dilution claim.” Stallings, 829 F.2d at 1555; accord Collins v. City of Norfolk, 816 F.2d 932, 935 (4th Cir.1987). Thus, the Gingles Court did not decree that a plaintiff who proves the three preconditions invariably will win; instead, it merely held that the three factors are prerequisites to a successful claim. See Gingles, 478 U.S. at 48 n. 15, 106 S.Ct. at 2765 n. 15. The Supreme Court’s recent decision in De Grandy, — U.S. -; 114 S.Ct. 2647, resolved any doubt as to the threshold nature of the Gingles factors. In De Grandy, a case involving a challenge to Florida’s legislative reapportionment plan, the Court definitively characterized the “Gingles factors (compactness/numerousness, minority cohesion or bloc voting, and majority bloc voting) as ‘necessary preconditions,’ for establishing vote dilution.” Id. at -, 114 S.Ct. at 2657 (citation omitted). The Court went on to say, however, that “the ultimate conclusions about equality or inequality of opportunity were intended by Congress to be judgments resting on comprehensive, not limited, canvassing of relevant facts.” Id. at -, 114 S.Ct. at 2657. Thus, under De Grandy, reviewing courts are required — not just invited — to look beyond the Gingles threshold factors when evaluating vote dilution claims. As the Court stressed, “[l]ack of electoral success is evidence of vote dilution, but courts must also examine other evidence in the totality of circumstances, including the extent of the opportunities minority voters enjoy to participate in the political processes.” Id. at -, 114 S.Ct. at 2657. In De Grandy, the Court assumed that the Gingles preconditions had been satisfied, but nevertheless concluded that section 2 relief should not be granted because, notwithstanding the presence of continued discrimination and racial bloc voting, minority voters were able to form effective voting majorities in a number of legislative districts that were roughly proportional to their respective shares in the voting age population. See id. at -, 114 S.Ct. at 2663. As an initial matter, therefore, proof of the Gingles threshold factors is a necessary precondition to section 2 relief; such a showing, however, will not guarantee relief. A defendant in a vote dilution case may always attempt to rebut the plaintiffs claim by introducing evidence of objective, non-racial factors under the totality of the circumstances standard. Indeed, [i]f the Court meant to deny the defendant an opportunity to rebut the plaintiffs case after the plaintiff has offered evidence of the three factors, then the three factors would be both necessary and always sufficient to win under section 2.... The Court’s adherence to the totality-of-the-circumstances test must mean that the defendant can rebut the plaintiffs claim— even after the plaintiff has offered proof of the three Gingles factors. Solomon, 899 F.2d at 1035 (Tjoflat, C.J., specially concurring). Because the ultimate inquiry is under the totality of the circumstances, courts must consider not only evidence of the Gingles threshold factors but also any other evidence offered by the parties that is relevant to the statutory test. It may well be true, as the Third Circuit has suggested, that “it will be only the very unusual case in which the plaintiffs can establish the ... Gingles [threshold] factors but still have failed to establish a violation of § 2 under the totality of the circumstances.” Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1135 (3d Cir.1993). Nevertheless, the Supreme Court cautions: To be sure, some § 2 plaintiffs may have easy cases, but although lack of equal electoral opportunity may be readily imagined and unsurprising when demonstrated under circumstances that include the ... essential Gingles factors, that conclusion must still be addressed explicitly, and without isolating any other arguably relevant facts from the act of judgment. De Grandy, — U.S, at --, 114 S.Ct. at 2657. It is, therefore, under the totality of the circumstances that courts must decide, in the final analysis, whether a violation of section 2 has been established. B. Given that courts evaluating section 2 claims must look beyond the Gingles threshold factors, we must determine what constitutes a violation of the statute’s “results” test under the totality of the circumstances. As the Supreme Court has stated, section 2 “make[s] clear that certain practices and procedures that result in the denial or abridgement of the right to vote are forbidden even though the absence of proof of discriminatory intent [in the adoption or maintenance of those practices and procedures] protects them from constitutional challenge” under the Fourteenth and Fifteenth Amendments. Chisom v. Roemer, 501 U.S. 380, 383-84, 111 S.Ct. 2354, 2358, 115 L.Ed.2d 348 (1991). The concern thus becomes what proof is required to satisfy the results test, as well as what form of intent inquiry is forbidden thereunder. We hold that section 2 prohibits those voting systems that have the effect of allowing a community motivated by racial bias to exclude a minority group from participation in the political process. Therefore, if the evidence shows, under the totality of the circumstances, that the community is not motivated by racial bias in its voting patterns, then a ease of vote dilution has not been made. Our interpretation of the Voting Rights Act is supported by the text of the statute, the legislative history accompanying the 1982 amendment to section 2, and the Supreme Court’s vote dilution cases. We now discuss these sources in greater detail. 1. At first glance, the language of section 2, as amended in 1982, appears somewhat inconsistent and unclear. Subsection (a) forbids the use of electoral structures that “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. § 1973(a). Subsection (b), on the other hand, appears more generous, explaining that a violation of subsection (a) is established if members of a minority group “have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.” Id. § 1973(b). We believe, however, that the statutory language, when properly understood, circumscribes the universe of prohibited electoral structures and requires more than a mere showing of electoral losses at the polls by minority candidates. The appellants argue that the language of section 2 requires nothing more than proof of disparate election results — i.e., that, when observed from a numerical/success rate point of view, minorities have less opportunity than their white counterparts to participate in the political process and to elect representatives of their choice. Such proof, they argue, sufficiently demonstrates the required denial or abridgement of a minority group’s right to vote. This reading of section 2, however, ignores crucial portions of the language in both subsection (a) and subsection (b). It is an axiomatic principle of statutory construction that statutes are to be read in their entirety, not in the piecemeal fashion employed by the appellants. Accordingly, we “follow the cardinal rule that a statute is to be read as a whole, see Massachusetts v. Morash, 490 U.S. 107, 115, 109 S.Ct. 1668, 1673, 104 L.Ed.2d 98 (1989), since the meaning of statutory language, plain or not, depends on context.” King v. St. Vincent’s Hosp., 502 U.S. 215, 221, 112 S.Ct. 570, 574, 116 L.Ed.2d 578 (1991). When the entire statutory provision is examined, it becomes clear that the well-known first portion of subsection (b) can only be understood in light of subsection (a), which requires that the denial or abridgement of the right to vote be “on account of race or color,” and the final clause of subsection (b), which makes clear that the 1982 amendment was not designed to create a right of proportional representation. This comprehensive reading plainly precludes the appellants’ interpretation of section 2. First, the language of section 2 as amended, by prohibiting voting structures that “result[ ] in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color,” explicitly retains racial bias as the gravamen of a vote dilution claim. 42 U.S.C. § 1973(a) (emphasis added). The existence of some form of racial discrimination therefore remains the cornerstone of section 2 claims; to be actionable, a deprivation of the minority group’s right to equal participation in the political process must be on account of a classification, decision, or practice that depends on race or color, not on account of some other racially neutral cause. As the Fifth Circuit has explained, “[t]he scope of the Voting Rights Act is indeed quite broad, but its rigorous protections, as the text of § 2 suggests, extend only to defeats experienced by voters ‘on account of race or color.’ ” LULAC, 999 F.2d at 850. Furthermore, this linguistic conclusion is supported by the fact that any other reading might well render section 2 outside the limits of Congress’ legislative powers and therefore unconstitutional. It is important to remember that the Voting Rights Act was adopted pursuant to Congress’ authority to enforce the Fourteenth and Fifteenth Amendments. The Civil War Amendments to the Constitution were designed to remedy pervasive racial discrimination, and Congress has broad power to enforce those amendments by appropriate legislation. Congressional actions, however, must remain rooted in the purpose of the amendments. Accordingly, the Supreme Court has upheld the ban on electoral changes in section 5 of the Voting Rights Act (the preclearance provision) because “the Act’s ban on electoral changes that are discriminatory in effect is an appropriate method of promoting the purposes of the Fifteenth Amendment, even if it is assumed that § 1 of the Amendment prohibits only intentional discrimination in voting.” City of Rome v. United States, 446 U.S. 156, 177, 100 S.Ct. 1548, 1562, 64 L.Ed.2d 119 (1980). Thus, Congress’ decision to include and retain the “on account of race or color” language was not merely fortuitous. Indeed, “[t]his limitation was not so much the product of legislative discretion as constitutional imperative, given that the scope of Congress’ remedial power under the Civil War Amendments is defined in large part by the wrongs they prohibit.” LULAC, 999 F.2d at 854. As the Senate Judiciary Committee concluded, the proposed amendment to section 2 was constitutional because of “the very terms and operation of the provision, which confine its application to actual racial discrimination.” S.Rep. No. 417, at 43, reprinted in 1982 U.S.C.C.A.N. at 177, 221. In addition, to accept the appellants’ interpretation of section 2 — that is, to read the “on account of race” language out of the statute by allowing section 2 plaintiffs to establish a violation merely by proving numerical differences in representation levels— would be to create a defacto right to proportional representation, a result explicitly prohibited by section 2 itself. See 42 U.S.C. § 1973(b) (“[NJothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.”). For, under the appellants’ reading, courts would do no more than sim