Clark v. Roemer

FINDINGS OF FACT AND CONCLUSIONS OF LAW REMEDY PHASE JOHN V. PARKER, Chief Judge. A. PRELIMINARY STATEMENT In view of the holding of the Fifth Circuit in Chisom v. Edwards, 839 F.2d 1056 (5th Cir.) cert. denied sub nom. Roemer v. Chisom, 488 U.S. 955, 109 S.Ct. 390, 102 L.Ed.2d 379 (1988), this court assumed that its path was rather clearly staked by the appellate court. However, on May 11, 1990, the decision of the Fifth Circuit came down in League of United Latin Am. Citizens Council No. 4434 v. Clements, 902 F.2d 293 (5th Cir.1990). That ease effectively undercuts the teaching of Chisom because it holds that, although Section 2 of the Voting Rights Act, 42 U.S.C. § 1973, does indeed apply to judicial elections, the use of at-large election districts in the election of Texas trial judges does not violate Section 2. Since Louisiana’s trial judges are also elected at-large, that holding, if it becomes final, will have a significant impact upon the decision which this court must hand down in this Louisiana case. This case has been tried and decided on the violation issues and the Louisiana Legislature was granted time to present a remedy which the Legislature failed to do. This case has also been tried on the remedy phase and the court was on the verge of presenting its decision on the remaining issues when the decision in the Texas case came down. The Fifth Circuit has now granted rehearing en banc in the Texas case which has the effect of vacating the opinion of the panel. League of United Latin Am. Citizens Council No. 44346 v. Clements, 902 F.2d 322 (5th Cir.1990). The Fifth Circuit has scheduled oral argument for June 19, 1990. The LULAC opinion is thus suspended and as of this writing, the only appellate guidance available to this court is Chisom v. Edwards. Under ordinary circumstances this court would simply delay its decision pending final appellate court judgment. These are not ordinary circumstances, however. The court has concluded that delaying its decision in this case would disserve the cause of justice. Some of the factors which lead to this conclusion are: Louisiana has judicial elections scheduled generally in 1990. Time is required to gear up for those elections both by the candidates and the state officials charged with conducting the elections. There are vacancies in the state judicial system which need to be filled but which, because of the state’s consent injunction, cannot be filled. As will be seen, infra, this court does not find a Section 2 violation in every district *450 and it is imperative that elections be allowed to proceed free of federal interference in those districts in which no Section 2 violation is found. Throughout the course of this litigation, this court has marched to the solitary beat of its own drum, insisting that if Louisiana’s system for selecting its judges produces violations of Section 2 of the Voting Rights Act, the remedy should be to revise the system, not to make adjustments in a few “guilty” districts which may well need adjusting again in the future. (See, for example, the views expressed at Clark v. Edwards, 725 F.Supp. 285 at 294-295) (M.D.La.1988). None of the parties — plaintiffs, defendants, or intervenors — have joined the court’s march. Each insists that, as a finding of a Section 2 violation is district specific, so any remedy must also be limited to districts in which specific violations have been found. 1 For reasons stated later, the court concludes that it lacks the power to impose a systemic remedy upon the state and that any remedy is indeed limited to “guilty” districts. Both the parties and the appellate court need to be aware of this court’s findings of fact. If, because of the Fifth Circuit’s resolution of the Texas case, revisions in this court’s conclusions of law are required, so be it. Accordingly, it is imperative that the state be permitted to fill judicial vacancies as rapidly as possible, in these districts in which no violation is found. B. BACKGROUND OF CASE This is a class action brought by black voters and black lawyers who possess the qualifications to be elected to the offices of Louisiana district judge, family court judge and court of appeal judge. Plaintiffs claim that the use by Louisiana of multimember election districts to elect these judicial officers operates to dilute black voting strength in violation of Section 2 of the Voting Rights Act of 1965, as amended, 42 U.S.C. § 1973. Plaintiffs also assert claims of racial voter discrimination under the Fourteenth and Fifteenth Amendments. Trial was had on the liability issues and the court concluded that the use of multimem-ber election districts has indeed produced violations of Section 2 of the Voting Rights Act in some of Louisiana’s judicial districts. In view of the finding of statutory violations, the constitutional claims were not addressed. See Clark v. Edwards, 725 F.Supp. 285 (M.D.La.1988). On December 15, 1988, the State of Louisiana requested and was granted time to attempt to structure remedial legislation during the 1989 Regular Session of the Legislature and for such additional time as might be required to submit to the electorate any amendments to Louisiana’s Constitution which might be required relative to revising the State’s judicial selection process. Plaintiffs opposed the request. At the same time, the State consented to an injunction which prohibits filling by election positions “now vacant” for the family court, district courts and courts of appeal, “pending the issuance of further orders by this Court.” It should also be noted that a three-judge court was convened in this case under Section 5 of the Voting Rights Act, 42 U.S.C. § 1973c, because Louisiana had created judgeships and had taken other actions regarding judicial election procedures which had not been submitted for approval to either the Attorney General of the United States or to the United States District Court for the District of Columbia, Allen v. State Board of Elections, 393 U.S. 544, 89 S.Ct. 817, 22 L.Ed.2d 1 (1969). The three-judge court enjoined implementation of a number of such changes and the State authorities proceeded to submit a number of such changes to the Attorney General. In due course, letters of “no objection” issued as to several changes which the State was then allowed to implement. On September 27, 1988, the three-judge court enjoined the implementation of La. Act 801 of 1987 which created four new judgeships in the Second and Third Circuit Courts of Appeal. That injunctive order was predicated upon the formal objection of the Attorney Gener *451 al to implementation of Act 801. That in-junctive order remains in effect and will be unaffected by any order of this single-judge court. See Allen v. State Board of Elections, supra. Eventually, the Legislature did come up with a number of specific changes in judicial election procedures in some judicial districts. Some of the changes required revisions to Louisiana’s Constitution and, under the direction of the Legislature, the entire “package” was submitted to a vote of the people of Louisiana. In November, 1989, the people rejected the proposed revisions. The Legislature has submitted nothing to this court. The matter has now been heard on the remedy phase and numerous proposals have been presented to the court by the original parties as well as by a host of intervenors. Other proposals have been submitted by way of amicus curiae briefs, mostly on behalf of individual courts. As noted, however, no proposal has been submitted by the Louisiana Legislature. C. PENDING MOTIONS Shortly before trial, a motion was filed on behalf of the Governor and other State defendants to “Recall and Recast Findings of Fact and Conclusions of Law as Entered on the 15th Day of August, 1988.” Also, shortly before trial, The Orleans Trial Judges Association, as intervenor, filed a motion to reconsider the court’s findings as to Orleans Parish (Orleans Civil and Criminal District Courts). Both motions were taken up at trial and are submitted on the record as made at the original trial and the trial on remedy. Plaintiffs vigorously oppose both motions. The motions present several issues. Issues similar to those presented hereby were raised regarding Louisiana’s system for electing justices of its Supreme Court in Chisom v. Edwards, 839 F.2d 1056 (5th Cir.) cert. denied sub nom. Roemer v. Chisom, 488 U.S. 955, 109 S.Ct. 390, 102 L.Ed.2d 379 (1988). There the district court held that Section 2 of the Voting Rights Act, as amended, did not apply to judicial elections. The Fifth Circuit reversed, holding: “Clearly, judges are ‘candidates for public or party office’ elected in a primary, special, or general election; therefore, section 2, by its express terms, extends to state judicial elections. This truly is the only construction consistent with the plain language of the Act.” 839 F.2d at 1060. Thus, under Chisom, this court is bound to apply the provisions of Section 2 of the Act to the judicial elections at issue in this litigation. So long as Louisiana chooses to select its judges by popular election, the elections must all be conducted in accordance with Section 2, which forbids the states from using any voting standard or practice, “which results in a denial or abridgement of the right of any citizen ... to vote on account of race or color_” The Act goes on to specifically recognize claims of vote dilution as a violation of the statute. The 1982 amendment to the Act rejects the “intent to discriminate” test in favor of the “results” test. If, under the “totality of circumstances” it is established that “the political processes ... are not equally open to participation by ... [a protected minority] ... 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” then a violation of the Act has been shown. In Thornburg v. Gingles, 478 U.S. 30, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), the Supreme Court explained the requirements for a vote dilution case under Section 2. In view of the square holding in Chisom that Louisiana’s judicial elections fall within the requirements of Section 2 of the Act, it is appropriate that the court apply the teachings of Thornburg to determine whether plaintiffs have proved any violations. That is what the court did following the trial on liability issues. The decision in League of United Latin Am. Citizens v. Clements, supra, is predicated upon the notion that trial judges hold “single-member” offices, more closely akin to offices such as president or governor or mayor than to multi-member bodies such as legislatures or municipal councils or school *452 boards. Noting that each district (trial) judge individually hears and decides each case on his own docket without consultation or input from any other district (trial) judge in the judicial district, regardless of the number of judges for that district, the panel concludes that there can be no violation in the at-large election of such officers: “There cannot be a violation of Section 2(b) ... through at-large elections of the trial judges who sit on the Texas district courts.... ” This is so, the panel reasons, because, “... the full authority of a trial judge’s office is exercised exclusively by one individual, and there can be no share of such a single-member office.” (slip op. at 4087) This reasoning apparently would not extend to appellate court judges who do decide cases in panels, one vote each. The reasoning would, however, squarely apply to Louisiana’s district (trial) judges since their functions are exercised in precisely the same manner as are those in Texas and they are elected district-wide (at least one parish in every judicial district). This court is not moved to embrace the reasoning of the panel in the Texas case at this time for two reasons: (1) the decision is not final; indeed, it is at present suspended; and (2) no party to this litigation has at any time advanced the reasoning of the Texas case. An additional factor is that court of appeal judges are involved in this case, hence claimed Section 2 violations must be dealt with at that level in any event. This court can find no applicable teaching to use in determining whether Section 2(b) violations have been proved except Thornburg v. Gingles and that standard will be applied. This court must admit that its discussion of the Thornburg threshold factors was minimal at best. The Fifth Circuit on more than one occasion has stressed the need for district courts to make specific, detailed findings of fact in vote dilution cases. See e.g., Westwego Citizens For Better Government v. City of Westwego, 872 F.2d 1201 (5th Cir.1989). Because the court’s findings of fact were not sufficiently specific as to all judicial districts, the court will reconsider its findings and will make additional findings. All previous findings are adopted, except as they may be specifically changed herein. D. THE APPLICABLE LAW Several matters need to be addressed before undertaking a district by district review of the evidence. In Thornburg v. Gingles, the Supreme Court made it plain that plaintiffs in a vote dilution case bear the burden of proving that “the use of a multimember electoral structure operates to minimize or cancel out their ability to elect their preferred candidates.” 106 S.Ct. at 2765. The Court also made it plain that three factors must be established before it becomes necessary to undertake an examination of the totality of the circumstances (the so-called “Zimmer” factors): (1) The minority group must be sufficiently large and geographically compact to constitute a (voting) majority in a single-member district; (2) The minority group must be politically cohesive; and (3) The minority must demonstrate that the white majority votes sufficiently as a bloc to enable the majority to usually defeat the cohesive minority’s preferred candidate. While satisfying these threshold factors does not prove a violation of Section 2, failure to establish all of them may eliminate the necessity to even conduct a totality of the circumstances examination. Monroe v. City of Woodville, Miss., 881 F.2d 1327 (5th Cir.1989), on petition for rehearing, 897 F.2d 763 (5th Cir.1990); Overton v. City of Austin, 871 F.2d 529 (5th Cir.1989). Although the Fifth Circuit has apparently not yet squarely so held, it seems rather clear that the majority population with which Thornburg v. Gingles is concerned is a voting majority, not simply a population majority. The Court of Appeals has at least implied that the single-member district which is created must contain at least a voting age majority of the minority group. See Houston v. Haley, 859 F.2d 341 (5th Cir.1988), vacated on other grounds, 869 F.2d 807 (1989), (where the court referred to this issue as “critical”); Brewer v. Ham, 876 F.2d 448, 452 (5th *453 Cir.1989); Overton v. City of Austin, 871 F.2d 529, 542 (5th Cir.1989), Jones, J., concurring. This court concludes that in order to be viable under the Thornburg v. Gingles rationale any such district must contain at least a voting age majority of the minority group. In the absence of sufficient votes to effectively compete, clearly it is not the existence of the multimember district that deprives the minority group of the opportunity to elect its preferred candidates. In this case the defendants’ expert, Dr. Weber, had accurate voter registration data available. In drawing subdistricts he attempted to form black voter registration majorities. Although it may not be appropriate to use registration figures in every vote dilution case, it strikes this court that, when such information is available, it provides valuable assistance in determining whether there is a violation of Section 2 in a judicial district, as well as in fashioning a remedy which will actually work. Relying upon Wells v. Edwards, 347 F.Supp. 453 (M.D.La.1972), aff'd, 409 U.S. 1095, 93 S.Ct. 904, 34 L.Ed.2d 679 (1973), plaintiffs, in both the liability trial and the remedy phase, have frequently ignored “one-man, one-vote” principles. As' the Fifth Circuit pointed out in Chisom v. Edwards, 839 F.2d 1056, 1060 (5th Cir.1988), such equal representation principles do not apply per se to judicial elections, but Wells clearly has no application to claims of racial discrimination in judicial elections. This court certainly accepts the notion that there is no federal requirement that judicial districts must contain approximately equal numbers of people. Indeed, judicial districts are jurisdiction specific, not population oriented. Many times the limits of a judicial district are historical and tied directly to a unit of local government. However, where one is drawing subdistricts within existing judicial districts, either for the purpose of determining violation under Section 2 or for remedy purposes, equitable considerations (and possibly constitutional questions) arise. A subdistrict which allows a much smaller number of people to elect one or more judges whose jurisdiction extends throughout the judicial district, including those who reside in the district but cannot vote for or against that judge or those judges, raises rather perplexing questions. In any event, this court concludes, as did Judge Barbour in Martin v. Allain, 658 F.Supp. 1183, 1198 (S.D.Miss.1987) that equity demands that all such subdistricts — for whatever purpose created — must contain substantially equal populations. Plaintiffs as well as some of the other parties have implied that this court is free to create additional judgeships. The State Legislature followed that approach in the proposal which was rejected by Louisiana’s voters. Clearly in determining whether a violation has been proven, the court is limited to existing circumstances. As Judge Jones pointed out in her concurring opinion in Overton v. City of Austin, 871 F.2d 529, 543 (5th Cir.1989): “The district court assumed arguendo that it might order an expansion of the city council to accommodate single-member districts. This reasoning puts the cart before the horse, by authorizing a remedy for voter dilution before a violation of Section 2 has been found. Actionable vote dilution must be measured against the number of positions in the existing governmental body rather than some hypothetical model based upon whatever size is necessary to accomplish proportional representation....” As far as this court is concerned, the proper body to determine the optimum number of Louisiana judges is the Louisiana Legislature. Since the only additional judgeships created by the Legislature (excluding those enjoined by the three-judge court) were contingent upon voter approval of the entire package, this court will make no attempt to create additional judgeships for either violation or remedy purposes. DISCUSSION OF EACH JUDICIAL DISTRICT IN WHICH A VIOLATION WAS FOUND Each side had an expert who analyzed election results to estimate the number of black and white votes which were cast for *454 various candidates. Both utilized bivariate ecological regression analysis to obtain their estimates and their results are largely compatible with each other where the same elections are considered. Dr. Engstrom, the plaintiffs’ expert, analyzed only elections in which a black candidate ran against one or more white candidates. His view was that in over ninety percent of the judicial elections he had reviewed, black voters strongly supported the black candidate. Dr. Engstrom felt that he did not receive much useful information from the results of an election in which black voters are given their choice of white candidates. Dr. Weber, for the defense, insisted that all elections should be included and he included every district court, family court and court of appeal election since 1978. The court’s view is that some helpful information is provided by the results in all elections but that the most significant results are those involving black candidates. Plaintiffs have attempted to generalize Dr. Engstrom’s findings in some specific judicial elections to judicial elections statewide. The court would agree that there is, for example, evidence of strong racial polarization in Louisiana statewide but that generalization does not substitute for the district specific evidence of vote dilution which must be presented before a Section 2 violation may be found in any district. Both experts also analyzed the results of the 1987 Louisiana Secretary of State election, a state-wide election which included several black candidates, and the 1988 Louisiana presidential primary in which Rev. Jesse Jackson was a candidate. While plaintiffs’ expert, Dr. Engstrom, attributed more significance to those elections than did Dr. Weber, the court accepts them simply as general indications of voter behavior. For example, the evidence shows that the black Secretary of State candidates were almost unknown outside of the New Orleans area but almost everyone knows that Jesse Jackson is black. The results of those elections do, however, provide some useful information. 1. District Court and Family Court Districts First Judicial District Caddo Parish. Total population, 252,358. Black, 95,242 or 37%. The parties have stipulated that at least one compact single-member subdistrict can be drawn which would have a majority black population. Indeed, it is undisputed that such a subdistrict would have a voting age majority. Hence, the first of the Thornburg requirements are met. This district has nine district judges. Two black lawyers have been elected district judge in recent years in Caddo Parish. In 1974 Paul Lynch ran but lost to a white candidate. Neither expert included that election in his analysis. In 1978 Mr. Lynch ran again and was elected over a white candidate. The experts estimated that Lynch received more than ninety percent of the black votes cast and about 45% of the white votes. Judge Lynch died in office in 1982. In 1985 Carl Stewart was elected in the primary over two white candidates, with 50.7% of the total vote. The experts estimated that Stewart received nearly all the black votes and between fourteen and nineteen percent of the white votes. Evidence was presented of campaign appeals to racial prejudice by the successful white candidate in a 1982 election for Shreveport (which contains most of the people in Caddo Parish) City Court. Dr. Engstrom estimated that the black candidate received more than ninety-nine percent of the black votes in both the primary and general elections but only fourteen percent and 8.6 percent of the white votes. The ten district court elections considered by Dr. Weber, together with the elections included by Dr. Engstrom provide a factual basis for concluding that black voters in Caddo Parish are politically cohesive and frequently support the same candidate in judicial elections. When the candidate is black, black support is near unanimous. Despite the election victories by Judge Lynch and Judge Stewart, there is evidence to support the conclusion that the white majority usually does vote sufficient *455 ly as a bloc to prevent the minority’s preferred candidate from winning. The Zimmer factors discussed more fully in the original opinion support the conclusion that the totality of the circumstances demonstrate vote dilution and hence a Section 2 violation in this district. Second Judicial District Jackson, Claiborne and Bienville Parishes. Total population, 50,803. Black, 20,463 or 40.3%. The minority population of this three-judge district is not grouped compactly, but instead is spread over all three parishes in the district. At the first hearing, plaintiffs produced a district (as drawn by defendants, Ex. D-4) which was not at all compact (Dr. Weber counted 24 sides) and snaked through all three parishes. The proposed subdistrict produced a majority black population but not a majority black voter registration. At the remedy hearing, plaintiffs produced a very similar subdis-trict with a 54.3% black voter registration which was fairly equal in population. Like the first proposal, the second is a gerrymander, making it impossible to draw two more (or even one more) reasonably compact subdistricts, and two of the principal towns in the district, Arcadia and Jones-boro, are split. The lack of compactness and the gerrymander are clearly shown on Ex. P-1, a copy of which is attached hereto. The proposed majority black district is district 2. Thornburg requires that there be a compact district in which the minority may be the majority. That threshold requirement has not been established and it is not necessary to discuss the other Thornburg requirements or the Zimmer factors. The court erred in finding a violation of Section 2 in this judicial district and that finding is hereby vacated and set aside. Third Judicial District Lincoln and Union Parishes. Total population, 60,930. Black, 20,756 or 34.1%. The subdistrict created here suffers some of the same problems as that created in the Second Judicial District. Plaintiffs have attempted to create two subdistricts in this two-judge district, one majority black. The subdistrict submitted at the liability trial was not contiguous. The new subdistrict is more contiguous than the old but lacks compactness. It is simply not possible to draw a compact, population equal subdistrict having a majority black voter registration. Here again the first Thornburg threshold factor has not been established and the court erred in finding a violation of Section 2. That finding is hereby vacated and set aside. Fourth Judicial District Ouachita and Morehouse Parishes. Total population, 174,044. Black, 54,562 or 31.3%. The parties have stipulated that at least one single-member subdistrict can be drawn in this district which would have a majority black voter registration. This district elects seven judges. Since there were no district judge elections in which blacks were candidates, Dr. Eng-strom included 1984 elections for City Court judge in Bastrop and Monroe, the largest cities in Morehouse and Ouachita Parishes, respectively. He also analyzed a 1984 election for district attorney in the same district as the district court. There were black candidates in all three elections. Dr. Weber considered eight district court elections, one of which was a runoff of a primary. None of the elections Dr. Weber reviewed involved any black candidates. It was Dr. Weber’s opinion that the preferred candidate of minority voters won in seventy percent of the district court elections. The eight elections analyzed by Dr. Weber included three district court elections held at the same time as the city court and district attorney elections. In the district attorney and city court elections, each black candidate received strong (well over a majority in each case) black support and almost no white support. Evidence was presented of overt appeals to racial prejudice by the white city court candidate in Monroe. All the election results tend to show minority political cohesion in this district and racially polarized voting. These election *456 results, together with the totality of the circumstances discussed in the original opinion lead the court to find that vote dilution and hence a Section 2 violation has been established in this district. Sixth District East Carroll, Madison and Tensas Parishes. Total population, 36,272. Black, 21,224 or 58.5%. This two-judge district is majority black and the parties have stipulated that a single-member, compact, roughly equal in population black voter majority subdistrict may be drawn. Defendants’ “Motion to Recall and Recast” does not challenge the court’s finding of a Section 2 violation in this district. The court’s interlocutory ruling will be reconsidered despite the fact that defendants have not specifically challenged it. The only judicial election reviewed in this district was a 1984 race involving white only candidates considered by Dr. Weber. It was his opinion that the black “minority’s” preferred candidate won. Dr. Eng-strom also considered the results of the 1987 Louisiana Secretary of State election about which the court has already commented and the 1988 Louisiana presidential primary about which the court has also already commented. Plaintiffs bear the burden of proof as to both political cohesion and white bloc voting. The meager election results in this district are simply not sufficient to demonstrate the existence of the last two Thorn-burg threshold factors. The court erred in finding a violation of Section 2 in this district and that finding is hereby vacated and set aside. Seventh Judicial District Catahoula and Concordia Parishes. Total population, 35,268. Black, 11,231 or 31.8%. The subdistrict originally proposed by plaintiffs in this two-judge district was neither compact nor contiguous. At least three small areas were noncontiguous either to the proposed district or to the remaining district. Moreover, the proposed subdistrict had a population deviation of more than twelve percent. On the remedy phase, plaintiffs submitted a revised sub-district which apparently simply incorporates all black precincts, regardless of location. The revision has a population deviation of more than forty percent, splits the towns of Ferriday and Jonesville, and barely achieves a majority in voter registration. The court concludes that, after two tries, plaintiffs have failed to establish even the first prong of the Thornburg factors. The court erred in finding a violation in this district and that finding is hereby vacated and set aside. Ninth Judicial District Rapides Parish. Total population, 135,-282. Black, 36,308 or 26.8%. The parties have also stipulated on this district as to the first of the Thornburg factors. The expert for the defense, Dr. Weber, considered eight judicial elections since 1978 from this district. All except two were white only races and Dr. Engstrom considered the two with black candidates. Dr. Weber conceded that in this district black preferred candidates tended to lose, even in white only elections. Dr. Engstrom provided strong evidence of racial polarization in voting in this district and that there is political cohesion among black voters. The court concludes that plaintiffs have established all three of the Thornburg factors in this district. For the reasons originally set forth, the totality of the circumstances demonstrates a Section 2 violation in this district. Tenth Judicial District Nachitoches Parish. Total population, 39,863. Black, 14,450, or 36.3%. Having reviewed all the evidence from both hearings, the court concludes that a compact, contiguous single-member majority black district of roughly equal population cannot be created in this two-judge district. About 8,800 of the total black population lives in the City of Nachitoches but the remaining black population is disbursed throughout the district. Plaintiffs’ revised proposal on the remedy phase presents an affirmative gerrymander which *457 has 52.5% black voter registration but splits the City of Nachitoches across several subdistricts to no good purpose. It also presents a population deviation of nearly thirteen percent. The court erred in finding a Section 2 violation in this district and that finding is hereby vacated and set aside. Eleventh Judicial District DeSoto and Sabine Parishes. Total population, 51,007. Black, 16,250 or 31.9%. In this rural two-judge judicial district, the black population again is not grouped compactly. It is spread over both parishes. At the liability trial, plaintiffs proposed a subdistrict which included the City of Many as an isolated island, which contained a thirty-one percent population deviation and which contained forty sides. Its configuration made it impossible to create a second district with compactness. The revised proposal submitted at the remedy hearing contained one subdistrict with a population of forty thousand and the other of ten thousand (a deviation of about 115%) and still produced a black voter majority of only 50.5%. The subdistrict is not compact, is a gerrymander and the population deviation is unacceptable. Dr. Weber had originally listed this judicial district as one in which the black population is neither large enough nor geographically compact enough to create a black majority single-member district. This court concludes that Dr. Weber was and is correct and that the initial Thornburg factor has not been proved. The court erred in finding a Section 2 violation in this district and that finding is hereby vacated and set aside. Fourteenth Judicial District Calcasieu Parish. Total population, 167,223. Black, 35,284 or 21.7%. There are eight judges in this district and the parties have stipulated that the first Thornburg factor may be met. Dr. Weber testified that he reviewed six elections in this district. There were six different election days but there were actually eight separate judicial elections, one of which involved a black candidate in 1984. Dr. Engstrom considered only the one election involving the black candidate plus the statewide elections previously mentioned. Dr. Weber was of the opinion that the minority preferred candidate won 67% of the time. Dr. Engstrom was of the opinion that there was strong evidence of racial polarization in voting in the district, that the black minority was politically cohesive and that white racial bloc voting usually prevented the black preferred candidate from winning. Mr. Gray, the black lawyer who was the candidate, testified that he sought the support of white lawyers and public officials. He was the first black to run for a judicial office in the district and virtually no white lawyer or office holder would publicly endorse his campaign. He was told repeatedly that “Calcasieu Parish is not ready for a black judge,” he had great difficulty raising funds, and he had great difficulty placing any yard signs in white areas (only two signs were placed in the entire parish). Mr. Gray campaigned almost exclusively in the white community; he received almost all of the black votes and very few white votes against two white opponents. Plaintiffs have demonstrated political cohesion among black voters in this district as well as white bloc voting in frustration of minority preference. The “Zimmer” factors originally discussed together with this evidence establish a Section 2 violation in this district. Fifteenth Judicial District Acadia, Lafayette and Vermillion Parishes. Total population, 254,902. Black, 46,-661 or 18.3%. The parties have stipulated on the first Thornburg factor in this eleven judge district (and it is only because of the high number of judges that such a subdistrict is possible). Dr. Engstrom, for the plaintiffs, presented an analysis of what he called three elections involving black candidates versus white candidates. Dr. Weber presented an analysis of six judicial elections in the district, including what he called two elections involving black candidates. They are the *458 same elections; the witnesses just counted them differently. Dr. Weber was of the opinion that the black preferred candidate won only 33% of the time. Mrs. Sylvia Cooks, the black lawyer who was the candidate in both (or all three) of those races, testified. She ran in 1983 for a newly created vacancy against one white male lawyer. She received over ninety percent of the black votes and a few white votes but lost. She was the first black person to run for judicial office in that district. Mrs. Cooks tried again in 1987 against three white male lawyers. Again she received solid black voter support; she made the run-off but lost in the general election. Mrs. Cooks had fund raising problems in the white community similar to those of Mr. Gray in the Fourteenth District and similar problems in obtaining public white support. She also testified about the overt and covert racial appeals in both elections by candidates and the public. By all accounts, Mrs. Cooks ran a professional, well organized campaign, using a public relations specialist (who also testified), polls, television, etc. She spent large amounts of her own money. The court can say, without depreciating the qualifications of any of her opponents, that Mrs. Cooks was at least as well qualified professionally as any opponent in either race. As she put it, however, there was nothing that she could do about the color of her skin. There is no doubt that the white majority in the Fifteenth Judicial District voted against her in both races because of the color of her skin and that the bloc vote prevented the black preferred candidates from winning. It is beyond dispute that all three Thorn-burg factors have been established in this district and that evidence, together with the “Zimmer” factors discussed in the original opinion, support the court’s finding of a Section 2 violation in this district. Sixteenth Judicial District Iberia, St. Martin and St. Mary Parishes. Total population, 168,219. Black, 49,340 or 29.3%. This district elects seven judges. Plaintiffs submitted two different proposals for subdistricts which defendants mapped. (Ex. D-9 and D-10). Each created a sub-district which contained a barely majority black population (52.5 and 51.7 percent, respectively) but there is no evidence that either proposed subdistrict would actually contain a majority black voting age or registered voter population. Both proposals contained or created other problems as well. At the remedy hearing, plaintiffs submitted seven subdistricts, one of which achieved a black population of 58.1 percent and a black voter registration of 56.2 percent. The black majority district is not compact, although the populations of the districts are fairly even. The City of New Iberia is split among four subdistricts to no good purpose and most of the other six districts are irregularly shaped, as well. The court concludes that plaintiffs have failed to demonstrate the first Thornburg factor and it is not necessary to consider any other factor. The court erred in finding voting dilution in this district and that finding is hereby vacated and set aside. Eighteenth Judicial District Iberville, West Baton Rouge and Pointe Coupee Parishes. Total population, 75,290. Black, 33,041 or 43.9%. With four judges, the parties have stipulated that the first Thornburg factor has been established. Dr. Weber considered four judicial elections in this district, two of which involved the same black candidate. Dr. Engstrom considered only the races involving black candidates. Dr. Weber conceded that the candidate preferred by black voters lost in three of the four elections and that there is strong evidence of racial polarization in voting. (Although Dr. Weber’s report, Ex. D-2a, Table 2 indicates that he considered only two elections in this district, Appendix A actually lists four separate judicial elections on three different dates, two of which involved the same black candidate.) Despite the limited number of judicial elections in this district, the results of those elections, Dr. Weber’s testimony, the *459 results in other elections and Dr. Eng-strom’s testimony, convince the court that the finding of a Section 2 violation is correct. Nineteenth Judicial District East Baton Rouge Parish. Total population, 366,191. Black, 114,741 or 31.-3%. This district has thirteen judges and the parties have stipulated to the first Thorn-burg factor. Dr. Engstrom, in his analysis, considered five judicial elections involving black candidates. He also considered two elections for the City Court of the City of Baton Rouge, which includes a significant portion of all residents of the Parish. Dr. Weber considered eleven district court elections, three of which involved black candidates. (Appendix A actually lists fourteen separate elections, counting primary and general elections.) In only one of the five elections was a black candidate elected — in 1987 Judge Pitcher won. In the 1984 election, Rayford, a black candidate, received very strong black support but almost no white votes and lost. In two of the elections there was little support for the black candidate among either black or white voters. In the fifth race, the black candidate had a majority but not nearly the solid black support that Rayford and Judge Pitcher received. Judge Pitcher’s election was unusual. It was held at a time when this litigation was pending in Baton Rouge, where the suit had received substantial news coverage. Pitcher was already a judge of the City Court and had the advantage of campaigning with the title “judge” before his name. It was also unusual in that Pitcher received substantial public support from white lawyers and community leaders. His margin of victory was only 52.7% although he received nearly all of the black votes in a race against a virtually unknown white lawyer. This evidence, together with Dr. Eng-strom’s testimony and the Zimmer factors discussed previously support the court’s finding that a violation has been established in this district. The Family Court for the Parish of East Baton Rouge The election district for this court of three judges is East Baton Rouge Parish. The parties have stipulated as to the first prong of Thornburg factors in this district. All that the court has said concerning the Nineteenth Judicial District applies equally to the Family Court election district. The evidence supports the court’s finding of vote dilution, hence a Section 2 violation. Twentieth Judicial District East and West Feliciana Parishes. Total population, 31,201. Black, 16,301 or 52.2%. At the liability hearing, plaintiffs submitted a proposed subdistrict which was not compact, had a bare majority black population but less than a majority black voter registration. At the remedy phase, plaintiffs submitted a revised proposal which is more compact and is equal in population to the other district. It still does not achieve a majority black voter registration. Plaintiffs’ greatest problem in this rural, two-judge district is that they ignore the presence in West Feliciana Parish of the Louisiana State Penitentiary, the State’s largest penal institution. At any given time, this prison houses about five thousand convicted felons — none of whom have the right to vote. The prison is mostly black in population. To ignore a prison population in a large metropolitan area is one matter but in a parish such as West Feliciana which has only 12,000 inhabitants, is quite another. Neither proposed subdis-trict, because each contains so many disenfranchised residents, achieves a majority black voter registration. The first Thornburg factor has not been established. The court erred in finding a Section 2 violation and that finding is hereby vacated and set aside. Twenty-First Judicial District Livingston, St. Helena and Tangipahoa Parishes. Total population, 149,331. Black, 33,306 or 22.3%. *460 There are six judges in the district and plaintiffs proposed six subdistricts at the remedy trial. Three of them, including the proposed majority black district, are not compact. Plaintiffs achieved a 52% majority black voter registration but there is a 6.86% population deviation, the Town of Independence is split across two subdis-tricts and because of a poor history of black voter turn out, blacks are unlikely to outvote whites. The court erred in finding a statutory violation in this district and that finding is hereby vacated and set aside. Twenty-Third Judicial District Ascension, Assumption and St. James Parishes. Total population, 93,647. Black, 28,413 or 30.3%. The parties have stipulated that a compact, equal population majority black voter registration subdistrict can be created in this district. The defendants’ “Motion to Recall and Recast” did not challenge the court’s original finding of a Section 2 violation in this district. The court nevertheless reviews the evidence. There have been no judicial elections in this district in recent years. Thus, the only election information presented to the court was Dr. Engstrom’s testimony and report of 1987 Secretary of State and the 1988 presidential primary. According to Dr. Engstrom, Rev. Jackson received the overwhelming majority of black votes and almost no white votes in this district. No evidence of any other election, mayor, school board, district attorney, sheriff, police jury or any of the host of political offices Louisiana elects was presented. Any finding of a Section 2 violation here would simply be an assumption as to voter attitudes extrapolated from one presidential primary and the Secretary of State election. The Fifth Circuit has cautioned the district courts about reading too much into “exogenous election results.” Monroe v. City of Woodville, Miss., 881 F.2d 1327, 1329 (5th Cir.1989), on petition for rehearing, 897 F.2d 763 (5th Cir.1990); Westwego Citizens for Better Government v. City of Westwego, 872 F.2d 1201, 1208, n. 8 (5th Cir.1989). Although the presidential primary particularly demonstrates state-wide racial polarization in voting, that alone is not sufficient to prove vote dilution. A presidential primary is not akin to a judicial election in this judicial district, particularly since Louisiana is a late-comer to presidential primaries. There is simply insufficient evidence to support a finding of a Section 2 violation here. The court erred in making such a finding and it is hereby vacated and set aside. Twenty-Fourth Judicial District Jefferson Parish. Total population, 454,592. Black, 63,001 or 13.9%. This district has fifteen district judges and the parties have stipulated as to the first Thornburg factor. A black lawyer has served as a district judge in Jefferson Parish. Lionel Collins was appointed to fill a vacancy and then in 1978 was the only candidate and was elected without opposition. He was re-elected in 1984 without opposition and died in office in 1988. Such elections without opposition do not reveal very much about the electorate. Dr. Weber included seven district judge elections in his analysis (although he only refers to six in Table 2). He concluded that in 85% of the elections, the candidate favored by the minority voters won. He listed no elections involving black candidates. Dr. Engstrom included a 1987 election for Juvenile Court in Jefferson Parish in which a black candidate received 75% of the black votes but less than 5% of the white votes. This is a parish wide election with exactly the same district lines as the district court and, of course, is a judicial election. The court simply cannot accept Dr. Weber’s theory that, given a choice of white candidates, black voters win 85% of the time. The presidential primary showed over 95% of the black voters for Rev. Jackson and about 5% of the white voters. The evidence of racial polarization is strong. Here it seems rather clear that blacks cannot successfully field a candidate if the candidate has to run district wide and an 86% white majority will bloc vote against him. This is a district where the minority *461 has not been able to field but one candidate. The minority in a smaller district could elect candidates of their choice. The evidence supports the court’s finding of a Section 2 violation in this district. Twenty-Sixth Judicial District Bossier and Webster Parishes. Total population, 124,352. Black, 29,173 or 23.-5%. This district has four judges. (Act no. 174 of the 1989 Legislature adds a fifth judge but that legislation has apparently not yet been submitted for “preclearanee” under Section 5 of the Voting Rights Act). A review of all the evidence has convinced the court that no Section 2 violation has been proved in this district. The proposed subdistrict was nearly eleven percent below an equal population figure; the district was not compact; the district splits a voting precinct; and, finally, the district is not majority black — black registered voters are only 43.3% of the total. The court erred in concluding that a statutory violation has been established in this district and that finding is hereby vacated and set aside. Twenty-Seventh Judicial District St. Landry Parish. Total population, 84,128. Black, 31,965 or 38.0%. The parties have stipulated that a majority black registered voter district may be drawn in this four-judge district. The only judicial election in this district in recent years (at least the only one presented to the court) was a 1982 election considered by Dr. Weber which involved only white candidates. It was Dr. Weber’s opinion that the black preferred candidate did not win. The only other election results presented to the court were the aforementioned presidential primary and race for Secretary of State. There is simply not sufficient evidence. The Twenty-seventh Judicial District may be an area where minority voters are politically cohesive but have been unable to field judicial candidates. Plaintiffs, however, bear the burden of proving these facts. Here, as in the Sixth and Twenty-third districts, plaintiffs have not presented sufficient evidence to establish the second and third Thornburg factors and the court is unwilling to extrapolate conclusions from the sketchy election results presented. The court erred in concluding that a Section 2 violation has been proved in this district and that finding is hereby vacated and set aside. Twenty-Ninth Judicial District St. Charles Parish. Total population, 37,-259. Black, 9,499 or 25.4%. Based on the testimony of Dr. Weber at the first hearing, the court concludes that it is not possible to draw a majority black subdistrict in this three-judge district which would meet the first Thornburg prong. The court erred in finding a Section 2 violation in this district and that finding is hereby vacated and set aside. Fortieth Judicial District St. John the Baptist Parish. Total population, 31,924. Black, 12,175 or 38.1%. The parties have stipulated that the first Thornburg factor is met in this two-judge district. Defendants’ “Motion to Recall and Recast” does not challenge the court’s finding of a Section 2 violation. Dr. Weber considered four judicial elections, all involving only white candidates. It was his opinion that the candidate preferred by black voters won each election. Dr. Engstrom considered no judicial election results. He did consider the Secretary of State and presidential primary elections. Dr. Weber’s weighted regression analysis indicates that in all four elections black voters exhibited political cohesion, as they did in the presidential primary. The court concludes that in this two-judge district, black voters have simply been unable to field a candidate. All Thornburg factors are present and that evidence, together with the Zimmer evidence discussed in the first opinion, supports the court’s finding of a Section 2 violation. Orleans Parish Total population, 557,515. Black, 308,149 or 55.3%. As of January, 1988 blacks were 53.17% of the registered voters. The black voting age population has *462 been and still is increasing and the black majority in registration is increasing. Blacks have been a majority of voters since March 31, 1985. The parties have stipulated as to the first Thornburg prong. The Parish of Orleans serves as an election district for numerous public offices, including civil and criminal district courts, juvenile court, one district of the Court of Appeal, Fourth Circuit, mayor, councilman-at-large, district attorney, civil sheriff, criminal sheriff, school board, civil clerk of court and criminal clerk of court. There are twelve civil district judges and ten criminal district judges. Orleans elects eight court of appeal judges. At present there are three black civil district court judges, no black criminal court judges and one black court of appeal judge. In the past, the late Ernest Morial was elected juvenile judge, after having been appointed to that office. Later he was elected to the court of appeal. He resigned his judicial position to run for mayor of New Orleans and was twice elected. Also in the past, Judge Israel Augustine (now retired) was appointed criminal court judge and was later elected to that position over two white candidates, and was elected as court of appeal judge without opposition. The current mayor of the City is black and has just been re-elected to a second term. Blacks have been elected to many other offices in New Orleans where they are currently serving. Some blacks were elected before Orleans had a majority black population. They were elected with solid black support and some white “cross-over” votes. Dr. Engstrom, plaintiffs’ expert, testified at the first hearing that Orleans Parish is now racially competitive: Q. You have said that a district can have polarized voting between blacks and whites, both; and yet because of the outcome of elections caused by crossover votes by both whites and blacks, that the black vote is not diluted. A. I have said that, yes. Q. Isn’t that true in Orleans Parish? A. I have said that the Parish of Orleans is today a racially competitive election district. Tr. vol. 1, p. 86 Further Dr. Engstrom testified: Q. Will you kindly do it? A. Well, I would say that a competitive district is an election in which both the black preferred and the white preferred candidates have reasonable opportunities to expect to be elected. Q. And that is the state of Orleans Parish at this time? A. That is the state of Orleans Parish today, yes, I believe so. Tr. vol. 1, p. 104 Dr. Renwick, one of the defendants’ experts, submitted a report, OTJA Ex. 12, which details the large number of black candidates who have been elected to office, including judicial office, in Orleans Parish. There is no question that black voters are politically active and organized. They can be politically cohesive but they do not always vote as a racial bloc. There is cross over voting among both black and white voters, although less among white voters. Plaintiffs cannot dispute the facts which past election results and voter registration figures establish and they do not attempt to do so. They argue instead that the court should impose remedial action because of the effects of “present and past dilution of black voting strength.” 2 This *463 court finds that, although there is still significant racial polarization in voting in Orleans Parish, there is no longer dilution of black voting strength. Plaintiffs contend that drawing smaller, blacker subdistricts would decrease the cost of elections and make it easier to elect candidates preferred by black voters to judicial office. That would no doubt be the result of taking plaintiffs’ suggestion. The court’s response is that Section 2 of the Voting Rights Act specifically does not guarantee proportional representation. The evidence shows that Orleans Parish is at present a level playing field for voters of both races. That is supposed to be the aim of the Voting Rights Act and of this court. Louisiana may draw its judicial election districts in any fashion it chooses so long as the use of the district does not dilute minority voting strength. The court erred in finding a Section 2 violation in Orleans Parish and that finding is hereby vacated and set aside. What has been said here applies to all judicial elections for which Orleans Parish is the district. 2. Courts of Appeal This court’s discussion of the Thornburg factors as related to the courts of appeal was even more sparse than the discussion relating to the district courts. As we have already noted, Thornburg requires district courts to make factual and district specific findings in vote dilution cases. That deficiency will be corrected by an individual examination of each district in which the court has made a finding of a Section 2 violation. Courts of Appeal, Circuit-wide The court found a Section 2 violation, “using a circuitwide analysis [in] all five of the existing circuit courts of appeal.” 725 F.Supp. at 304. The primary underpinning for that finding was the testimony of Mr. Floyd (Tr. Vol. 8, p. 14-17) that at least one compact, contiguous majority black district could be drawn in each circuit by ignoring existing district lines. As the Fifth Circuit subsequently held in Overton v. City of Austin, supra, that approach “puts the cart before the horse.” 871 F.2d at 543. The determination of vote dilution begins with examining the existing election district and the existing number of positions. Whether a court ought to consider changes in either, as a part of the remedy should a violation be found, is no part of determining whether there is vote dilution, for if there is vote dilution it is the existing district which must be its cause in order for there to be a Section 2 violation. It is clear that the court erred in finding that even the first of the Thornburg threshold requirements could be satisfied under the circuitwide analysis employed by Mr. Floyd. The court also erred in finding circuitwide statutory violations and that finding is hereby vacated and set aside. First Circuit Each of the three districts within this circuit elects four judges. District 2 East Baton Rouge Parish. The parties stipulated to the drawing of a district here as they did for the Nineteenth Judicial District. The findings of the court as to the district court are equally applicable to the court of appeal district. Dr. Weber’s inclusion of three elections in this district involving all white candidates does not alter the finding of vote dilution. District 3 East and West Feliciana, Livingston, St. Helena, St. Tammany, Tangipahoa and Washington Parishes. Total population, 335,608. Black, 76,776 or 22.9%. Plaintiffs were unable to draw a compact, majority black district within this district. The district proposed (depicted graphically upon Ex. D-15) is not compact *464 and has a black voter registration of only 43%. The first Thornburg requirement has not been met. The court erred in finding a violation of the statute in this district and that finding is hereby vacated and set aside. Second Circuit Each of the three districts elects two judges and one judge is elected at-large. (A second at-large judgeship was added by Act No. 801 of 1987 but that legislation has not yet been “precleared” under Section 5 of the Voting Rights Act, hence it cannot be put into effect.) Plaintiffs were unable to create subdis-tricts in any district of this court that actually had a majority black voter registration without allocating one of the at-large judges. In District 2 plaintiffs reallocated the at-large judgeship but still could produce only a 47.6% black voter registration. Under Overton, supra, of course, one may not add positions in deciding the question of violation. The court erred in finding Section 2 violations in districts 1, 2 and 3 of the Second Circuit and those findings are hereby vacated and set aside. Third Circuit Each of the three districts in this circuit elects two judges and three are elected at-large. (Act No. 801 of 1987 added an additional judgeship in each district but, as noted, the legislation has not been “pre-cleared”.) In this circuit plaintiffs were unable to propose a compact, majority black subdis-trict in any district. In order to try to create such subdistricts, they added the Act 801 judgeships and reallotted the at-large judgeships, one to each district. Even then it is not at all clear that the subdistricts would contain majority black voter registrations. In any event, plaintiffs approach is contrary to the teaching of Overton, supra. The court erred in finding violations of the statute in districts 1, 2 and 3 of the Third Circuit and those findings are hereby vacated and set aside. Fourth Circuit District 1 (Orleans Parish) elects eight judges and each of the other two districts (Plaquemines and St. Bernard Parishes) elects one. In addition, two judges are elected at-large. Plaquemines Parish, total population, 26,-049. Black, 5,540 or 21.3%. St. Bernard Parish, total population, 64,-097. Black, 2,411 or 3.8%. At-Large The only evidence offered was Mr. Floyd’s testimony that it is possible to draw majority black districts. It is quite apparent that a large number of Orleans precincts would have to be utilized for this purpose and the court has now held that there is no vote dilution there. Accepting Mr. Floyd’s testimony, there is still no evidence as to the second and third Thorn-burg requirements. The evidence is simply insufficient to establish vote dilution. The court erred in finding vote dilution in the at-large elections in this circuit and that finding is hereby vacated and set aside. Fifth Circuit This circuit elects six judges from district one and one from each of the other two. District 1 Jefferson Parish. The court found vote dilution in the Twenty-fourth Judicial District, which is also Jefferson Parish, based in part upon the stipulation of the parties that if the Twenty-fourth Judicial District were to be: “... subdivided into single-member sub-districts — with the number of single-member subdistricts equalling the number of judges in the original district, and with the subdistricts being of roughly equal population when compared to the other subdistricts within the original district — the black population ... is sufficiently large and geographically compact to constitute a majority in at least one of the single-member subdistricts.” (Stipulation No. 17) There is no such stipulation for the Court of Appeal District, relating to the first Thornburg threshold factor. The finding at the district court level does not automat *465 ically dictate such a finding for the court of appeal district. Plaintiffs were unable to draw a compact, contiguous district using the authorized number of positions (six) which was majority black in population or voter registration and thus failed to establish the first Thornburg factor. The court erred in finding a Section 2 violation in district 1 of this circuit and that finding is hereby vacated and set aside. REMEDY Having found Section 2 violations in some judicial districts, the court must now consider the appropriate remedy. 1. Legislative Plan Initially, the court considers plaintiffs’ argument that the court accept the “package” of bills, which the Legislature submitted to the voters as a constitutional amendment, as a “legislative plan” to which the court must defer. Plaintiffs insist that, because the “package” was designed by the Legislature, the court accept it as a legislative proposal to the court. As the defendants point out, the “plan” was rejected by the voters. As the court has pointed out, that “plan” was not submitted to the court. The Legislature has submitted nothing to the court. Much of the proposal submitted to the people did not, under Louisiana law, require amending the state constitution, yet the Legislature chose to base the entire package on a vote of the people. No contingency plan was prepared or submitted. The Legislature had no restrictions upon it as to what it might propose to the court. The court has no information indicating a legislative preference as to remedy because none has been submitted. Accordingly, the court concludes that no proposal before the court is entitled to preference over any other proposal. 2. Systemic Remedy Throughout these proceedings, this court has sought to avoid creating sub judicial districts in “guilty” districts, insisting that, if the election system employed by the state produced violations of Section 2 of the Voting Rights Act, the proper remedy would be to change the system, not tinker with a few judicial districts. As noted earlier, that view has found no support from any party to the litigation. All plaintiffs, all defendants and all intervenors have specifically declined to embrace the court’s notion. The Louisiana Organization For Judicial Excellence, as amicus curiae, however, in an eloquent brief, has urged the court to seize the “unique and rare opportunity to serve the people of Louisiana by establishing for the long term a judicial system designed ... ‘to establish justice for all’.” The Louisiana Organization For Judicial Excellence suggests that the court devise and order implemented a system of merit selection involving appointment and retention election for judges similar in many respects to the so-called “Missouri Plan” which is presently used in thirty-three states plus the District of Columbia and Puerto Rico at some or all levels of the judiciary. The Louisiana Organization For Judicial Excellence insists that such a merit system would not only improve the quality of the state judiciary but also would totally eliminate any possibility of Section 2 violations. Alternatively, LOJE suggests that if the court is restricted to providing remedies only in “guilty” judicial districts, then merit selection can and should be employed in those districts, thus encouraging the Legislature to extend the process statewide. The court is convinced that it is not free to write upon a clean slate; neither is it proper to anticipate and make provision for violations which will probably arise in the future, given the system employed by the state. It is well settled that a federal district court, in fashioning a remedy for voting rights violations should not “intrude upon state policy any more than necessary.” White v. Weiser, 412 U.S. 783, 795, 93 S.Ct. 2348, 2355, 37 L.Ed.2d 335 (1973) (reapportionment plan). Federal courts should follow policies expressed in state statutory and constitutional provisions whenever adherence to state policy would not detract *466 from federal constitutional requirements. Id. This same rule of deference applies in Section 2 cases. McGhee v. Granville County, N.C., 860 F.2d 110 (4th Cir.1988). As was noted by Judge Barbour in considering the Mississippi judicial election system, a federal court in fashioning a remedy in a judicial election case should conform to state policies as set forth in “existing state election and political laws” as much as possible. Martin v. Mabus, 700 F.Supp. 327, 330 (S.D.Miss.1988). All of the parties to this Louisiana litigation appear to agree that the basic goal of judicial deference to state policies is applicable. Not surprisingly, however, the parties disagree on what constitutes a state policy and which policy should be furthered over others. There are essentially three policies expressed in the Louisiana Constitution of 1974 which have been identified by the parties. Plaintiffs contend that Article 5, Section 22(A), which provides that “all judges shall be elected” with certain minor exceptions, should be paramount. The state defendants point to Article 5, Sections 9 and 14 as setting forth a policy against subdistricting judicial districts. Those sections provide: Section 9: Each circuit shall be divided into at least three districts, and at least one judge shall be elected from each. The circuits and districts and the number of judges as elected in each circuit on the effective date of this constitution are retained, subject to change by law enacted by two-thirds of the elected members of each house of the legislature. Section 14: The state shall be divided into judicial districts, each composed of at least one parish and served by at least one district judge. The Orleans Trial Judges have suggested that “Article 1 Section 10 of the state constitution expresses a more paramount state policy” of protecting the right to vote. The Louisiana Revised Statutes contain numerous sections relating to the courts of appeal, the district courts and the Family Court for the Parish of East Baton Rouge. Section 312 of Title 13 of the Louisiana Revised Statutes establishes the five court of appeal circuits and subdivides the state by parishes; it further subdivides each circuit into districts composed of at least one parish. (The only parish to be subdivided is St. John the Baptist Parish.) Sections 312.1, 312.2 and 312.3 establish the number of judges for each circuit and sets forth the number of judges which are to be elected from each election district or at-large by the “qualified electors” of that circuit. Section 321 creates divisions of the districts of the courts of appeal for the “purpose of nomination and election only.” Candidates for nomination and election to any court of appeal must “at the time of filing his declaration as a candidate ... designate only one division of the district ... for which he is a candidate.” Section 477 establishes forty judicial districts excluding the Parish of Orleans. Each district is composed of one or more parishes. Section 582 provides for “separate and distinct divisions of the various district courts having two or more elected or appointed judges, for the purpose of nomination and election.” Under Section 584, each candidate must designate the division for which he is a candidate. Under Section 585, the nomination and election of candidates for district judgeships is to be held and conducted in accordance with the primary and general election laws of the state. Sections 1136, 1335, 1346 and 1403 set forth the number of judges for Orleans Civil District and Criminal District Courts and the Family Court for the Parish of East Baton Rouge. Title 18, the Election Code, also contains numerous provisions relating to election of judicial officers. Sections 511 and 512 are the only pertinent provisions and they deal with the election of candidates in primary and general elections: both sections require a majority vote. This court concludes that Louisiana’s constitutional and statutory policies demonstrate a strong preference for popular election of judicial officers by majority vote. That conclusion precludes the use of the gubernatorial appointment and retention election system advocated by the Louisiana *467 Organization for Judicial Excellence. While a retention election (“Shall Judge Doe be retained in office?”) is technically an election, it is an election preceded by gubernatorial appointment and hence not a popular election in the Louisiana sense of the term. Federal courts in fashioning remedies for violations of federal law must always be aware that the court has authority only to remedy the specific federal violation. The federal court is not free to impose its own notion of whether a better system than that employed by the state could be devised. Federal authority becomes involved only because of a federal violation and, while the court has broad equitable authority to remedy the federal violation, that authority ends when the remedy is devised. Thus, the remedy to be fashioned in this case must: (1) be district specific, that is, the remedy may be imposed only in those specific districts where violations have been proven and (2) follow state policies except to the limited extent necessary to remedy the federal violations. White v. Weiser, 412 U.S. 783, 795, 93 S.Ct. 2348, 2354, 37 L.Ed.2d 335 (1973); Martin v. Mabus, 700 F.Supp. 327 (S.D.Miss.1988). This court has received a plethora of proposals on remedy. Plaintiffs have proposed subdistricts in judicial districts which are drawn so as to maximize minority voting strength. Creation of such subdistricts obviously generates other problems, such as compactness, equality of population, need for future modification because of additional judgeships or population changes and the potential or perceived potential for small election districts leading to “home town” justice, to name a few. In an effort to avoid the problems inherent in any subdistrict scheme involving judicial elections, the Louisiana District Judges Association and the Governor propose a simple solution — plurality voting in judicial elections by division — at both district and court of appeal levels. The advantages cited by the advocates of plurality voting are many: no need to draw or redraw compact, contiguous, equal population subdistricts; it retains Louisiana’s strong preference for popular election of judges; no problems are caused by adding additional judgeships; there is no concern about future violations — as minority voting strength increases in any particular district so will go the elections. Plaintiffs assert that the proposal has at least one significant disadvantage; it has not been demonstrated that plurality voting will actually provide a remedy for minority vote dilution. The temptation is great to embrace a seemingly simple, almost painless solution as the remedy in this case. The court has carefully considered the evidence for and against the plurality voting proposal. Ultimately, however, the court must conclude that elimination of the majority vote requirements in judicial elections will not actually remedy the Section 2 violations which have been established in this case. The court is mandated to fashion a remedy which will actually work and there is not sufficient evidence to sustain a conclusion that plurality voting will actually eliminate minority vote dilution. Dr. Weber has speculated on a number of possible scenarios in which black voters could win plurality elections. The problem with that approach is that it is, indeed, speculation. As plaintiffs have pointed out, plurality voting, by its very nature, requires more than two candidates and the overwhelming number of Louisiana judicial elections involve only two candidates. Plurality voting has every advantage except that of actually providing a remedy. For that reason the court must reject it. The Governor has presented what might be best described as a smorgasbord of possible remedies. They include a plan for multi-member subdistricts plus retention elections, a series of counter-subdistricts to those proposed by plaintiffs and a mixture of several different approaches. As an alternative to the drawing of subdistricts, about which this court has voiced reservations from the beginning of the case, plaintiffs have proposed the institution of a system of limited voting. This proposal is offered as a counter to the other side’s suggestion of plurality voting. This system was described by Dr. Eng- *468 strom and essentially would limit the number of votes to be cast by each elector in multimember districts to some number less than the number of positions to be filled. The theory is that a minority in such an election can overcome the effects of majority vote dilution by putting their strength behind a particular candidate. The limited number of votes make it less likely that the majority will be able to determine the winners of all positions, thus eliminating the “winner-take-all” effect. This proposal would require elimination of Louisiana’s majority vote requirement and would require all candidates to run, without designation by division, in a pool. Elections would be required at the conclusion of every term since all candidates would qualify for the same office and the top vote getters would fill the available positions. This court declines to accept limited voting for several reasons. First, it makes more changes in Louisiana’s chosen election system than required. At present, a competent judge frequently has no opposition for his re-election and no election is held, rewarding the judge for his service and saving the cost of an election. Under limited voting, an election would be necessary unless, for example, there were four positions to be filled and only four people qualified. The majority vote requirement has already been mentioned. Second, this court finds it difficult to comprehend how we expand access to the ballot box by limiting the number of votes each elector may cast. The court reluctantly concludes that the subdistrict approach suggested by plaintiffs, with all of its attendant problems, represents the only proposal which will actually remedy the violations of Section 2 (short of devising an entirely different system). Election subdistricts will be drawn in those judicial districts where violations have been found. State residence requirements shall be maintained in the judicial district but there shall be no residence requirement as to any election sub-district. All other state election requirements shall be maintained. Dr. Weber testified that in proposing alternate subdistricts to those opposed by plaintiffs, he relied upon eight criteria. As the court understands them, they are: (1) All subdistricts must be compact; (2) All subdistricts must be contiguous; (3) Parish and municipal boundaries should be maintained; (4) Current parish precinct lines should be followed; (5) Each subdistrict must be drawn using the existing number of judgeships; (6) Deviation in population across subdis-tricts must be kept to a minimum; (7) Gerrymandering to dilute minority voting strength must be avoided; (8) “Packing” or gerrymandering to concentrate minority voting strength in an attempt to achieve proportional representation must be avoided. These factors are appropriate and will be applied in determining subdistricts to be drawn in this case. The court has reviewed each subdistrict proposed by the parties in the eleven districts in which violations are found. As noted elsewhere, many subdistricts drawn by plaintiffs ignore the factor of equal population. Many of their subdistricts are not compact or contiguous and some of them do not comply with other criteria. Some of those proposed by Dr. Weber also present problems. The court is prepared, if necessary, to draw the districts itself; however, it would be of great benefit to the court if the parties could stipulate on those factors. The court requests the parties to confer with each other with the aim of entering a stipulation: (a) as to the configuration of subdistricts, (b) as to the number of judges to be elected in each subdistrict and (c) as to the specific divisions which shall be allotted to each subdistrict for election purposes. If such a stipulation is entered, it is understood that it will be done with full reservations of rights as to all parties and the stipulation shall not preclude any party *469 from appealing the court’s findings as to any judicial district. For the foregoing reasons, there will be judgment in favor of plaintiffs and against defendants making the injunction previously issued herein permanent as to the first, fourth, ninth, fourteenth, fifteenth, eighteenth, nineteenth, twenty-fourth and fortieth judicial districts and as to the Family Court for the Parish of East Baton Rouge and as to the Court of Appeal First Circuit, District 2. The injunction will be vacated as to all other judicial districts. *470 [[Image here]] Footnotes [1] This is not the argument advanced by the Louisiana Organization for Judicial Excellence, amicus curiae, however. Their argument is set forth in more detail later. [2] More particularly, plaintiffs argue that the court correctly found in its opinion on liability that there is a current dilution of black voting strength in Orleans Parish. Plaintiffs contend that the fact that Orleans Parish has a majority black population is not dispositive under the case law and because of state-sanctioned historical discrimination and resulting socioeconomic burdens, blacks are a minority of those voting in at-large judicial elections. Plaintiffs further contend that black electoral success in nonjudicial offices should not defeat an otherwise meritorious Section 2 claim. Alternatively, plaintiffs argue that even if black voting strength is not presently diluted, the “power of incumbency" of white judges (elected when at-large elections diluted black voting strength) perpetuates the effects of prior dilution and the system must be changed under Kirksey v. Board of Supervisors of Hinds County, 554 F.2d 139, 147 (5th Cir.1977). *463 In view of Dr. Engstrom’s testimony at the liability phase and again at the remedy phase that there is presently no dilution of black votes in Orleans Parish, plaintiffs simply have not established a Section 2 claim under Gingles. Plaintiffs’ reliance upon Kirksey which deals with perpetuation of intentional and purposeful discriminatory denial of access in violation of the Constitution is not persuasive here. Nor is it clear from the evidence presented that dilution will be perpetuated due to the incumbency of white judges.