WALLACE v. HOUSE

Rehearing and Rehearing En Banc Denied September 29, 1975.

GOLDBERG, Circuit Judge:

This case and a companion case, Perry v. City of Opelousas, 5 Cir. 1975, 515 F.2d 639, also decided today, form another chapter in the long and difficult struggle to ensure equal voting rights for all citizens. In the proceedings below, the district court determined that an all-at-large aldermanic election scheme in a small Louisiana town operated to dilute the votes of the town's black citizens in an unconstitutional fashion. The court also reasoned that the proposal of the Board of Aldermen [the Board] to adopt an election plan with a single at-large member also failed the constitutional test, and so ordered the implementation of an all-single-member selection process. Finally, the district court awarded attorney's fees to the black plaintiffs. 377 F.Supp. 1192. Although the all-at-large election scheme is clearly unconstitutional in the circumstances of this case, we believe that the Board's mixed election plan is not unconstitutional, and that the district court should therefore have deferred to the municipality's legislative judgment and adopted that plan. We affirm in part and reverse in part.

I

Ferriday, Louisiana, is a town of 5,200 people in Concordia Parish, in the northeastern part of the state. As is common in other towns in the area, Ferriday's population is closely divided between blacks and whites: the 1970 census counted about 3,000 blacks (58%) and 2,200 whites (42%). In March, 1972, the voters of Ferriday went to the polls to elect various local officials, including five aldermen, all of whom were to be elected at-large, with no residence requirements. It is fair to say that the town is both highly politicized and racially polarized, so that when the voters were faced with a choice of five white candidates and five black candidates, they apparently opted right down the line for racial solidarity, with whites voting for whites and blacks voting for blacks. Since whites enjoyed a very slight edge in voter registration over blacks (1,571 (50.5%) to 1,538 (49.5%)),1 and since 83% of the eligible voters turned out on election day, no one should have been surprised to learn that all five white candidates had been elected and all five black candidates defeated.

Even if they were not surprised, the defeated blacks were very unhappy with absolutely no black representation on a Board of Aldermen in a town with a black population majority. The black candidates accordingly filed this 42 U.S.C. § 1983 class action in federal district court on June 13, 1972, charging that Ferriday's all-at-large voting scheme impermissibly diluted the votes of local blacks, and asking for appropriate declaratory and injunctive relief. The court ordered each party to submit alternative redistricting plans, and a bench trial was held on April 24 and 25, 1974, after which the district court concluded that only single-member aldermanic districts would sufficiently guarantee to the black voters the full efficacy of their right of suffrage.

II

There is no question that Ferriday's all-at-large aldermanic election scheme operated to dilute the votes of the black citizens of the town, in violation of the Fourteenth and Fifteenth Amendment rights of that near-majority of the local electorate. As this Court noted in Howard v. Adams County Bd. of Supervisors, 5 Cir. 1972, 453 F.2d 455, 457, aggrieved voters may establish the existence of an unconstitutional districting scheme either by showing a racially motivated gerrymander or a plan drawn along racial lines, or by demonstrating that, designedly or otherwise, the particular scheme operates "to minimize or cancel out" the voting strength of minority elements of the voting population. The second type of cognizable grievance set out in Howard — the grievance of plaintiffs here — is generally denominated "dilution." We sketched the parameters of this complex doctrine in Zimmer v. McKeithen, 5 Cir. (en banc) 1973, 485 F.2d 1297:

485 F.2d at 1305. See also Turner v. McKeithen, 5 Cir. 1973, 490 F.2d 191, 194.

In this case, plaintiffs presented evidence that, with one recent and fortuitous exception,2 no black has ever been elected to municipal office in Ferriday. They showed that, in this thoroughly Democratic town (there are only 85 registered Republicans in the entire parish), no black man or woman has ever been selected to run as the candidate of the Democratic Party. Plaintiffs then chronicled the all-too-familiar story of racial segregation and other racial discrimination in every facet of local public life, including public education and public employment, which discrimination is abating only now. With respect to the distribution and quality of municipal services, the district court found that the streets and sidewalks, sewers and public recreational facilities provided by the Town for its black citizens are clearly inferior to those which it provides for its white citizens. In all these years, the Ferriday Board of Aldermen has failed miserably in its responsibilities to its many black constituents.

A large part of the explanation for this inexcusable neglect of black interests is found in the fact that most blacks were not able to vote in Ferriday until fairly recently. Although no federal registrars have ever been sent to Ferriday, there were very few blacks registered to vote until the Voting Rights Act of 1965 was enacted and enforced, and the legacy of intimidation and inferior educational opportunity inhibits many blacks in the exercise of the suffrage even when they are able to vote. Moreover, two features of Louisiana municipal election law coincide with the brief experience of black voting to minimize the impact of black votes upon the outcome of local elections, and therefore, presumably, upon the conduct of local officials. First, Lousiana law provides for a majority rather than a plurality requirement in primary elections. La. Rev.Stat. § 18:358. This seemingly nondiscriminatory provision3 requires any candidate who receives a mere plurality in the primary to run in a second primary. Where the voters vote overwhelmingly along racial lines, where a majority of the registered voters are white and where all of the officials are elected at-large, no black candidate is likely to achieve a majority in either the first or the second primary. And here, the primary is the election. As if the majority vote requirement were not enough of an impediment to black candidates, Louisiana law poses another severe obstacle to all minority voting interests, racial and otherwise, in the form of the "anti-single shot" or "full slate" requirement. La.Rev.Stat. § 18:351.4 This provision forces a voter in an at-large election to vote for as many candidates as there are places to be filled, on pain of having his ballot invalidated as to all of the at-large positions for that particular office. Where a minority interest group does not boast a full slate of candidates, the anti-single shot law requires supporters of the minority group to cast ballots for at least some of the group's opponents, thereby rendering the minority's task that much more difficult.5 See generally, Derfner, Racial Discrimination and the Right to Vote, 26 Vand.L.Rev. 523 (1973).

The district court was thus presented with a history of a dearth of black representation in Ferriday municipal government, with clear indications that white aldermen have not been in the habit of rendering effective representation for blacks, and with evidence that the past practice of disenfranchisement of blacks has combined with present provisions of Lousiana's at-large election machinery to make it almost impossible for blacks in Ferriday to compel any alderman to consider their interests. The court decided that at-large government as practiced in Ferriday was not constitutionally representative government insofar as the Town's black citizens were concerned, and that the plaintiffs had made out a prima facie dilution case under the Zimmer guidelines. Although the Board attempted to rebut plaintiffs' case at trial by arguing that Ferriday is a friendly small town where everyone knows everyone else and votes for the best man, the Board admitted to this Court at oral argument that plaintiff's dilution case had not been rebutted with respect to the all-at-large aldermanic districting scheme, and that the traditional scheme is unconstitutional. The able trial judge properly found that the all-at-large plan is irremediably defective, and that he correctly ordered the Board to submit a constitutional plan for his approval. The problem on appeal is whether the district court made the proper choice between the Board's alternative plans.

III

A

The Board submitted two redistricting plans to the district court.6 The first [or mixed] plan divided Ferriday into four reasonably compact single-member districts and one at-large district encompassing the entire community; the second created five compact single-member aldermanic districts. The maximum population percentage variations are 6.7% for the first plan and 4.3% for the second, and the average percentage variations are ± 2.11% and ± 1.02%, respectively. Both plans clearly fall within the permissible zone of district population variances allowed in state and local government units by the rule of Mahan v. Howell, 1973, 410 U.S. 315, 93 S.Ct. 979, 35 L.Ed.2d 320, and plaintiffs do not claim that the single-member districts contained in either plan have been manipulated in an illegal fashion. The discrete problem presented by this case is whether the one at-large district in the Board's first plan ineluctably taints the plan with the unconstitutional aura of dilution.

The practical politics of the controversy are simple. The first plan will create two "safe" white seats and two "safe" black seats, white the critical fifth aldermanic slot will be filled by the votes of all the registered voters of Ferriday. Since the whites have an over-all voting majority, however slim, and since almost every voter in the community seems to vote for persons of his or her own color, the at-large alderman will almost certainly be white and the Board will thus continue to be controlled by the Town's white residents. On the other hand, both parties agree that any fairly-drawn division of the community into five single-member districts will yield two "safe" white seats and three "safe" black seats, so that control of the Board will pass to the black residents of Ferriday if the second plan is adopted.7

Plaintiffs argue that the first plan — the plan the Board prefers — is an illegal "institutional" gerrymander, for the use of the one at-large district enables the whites to retain the political control that they would surely lose under an all-single-member plan. Defendants rejoin that the all-single-member plan would merely yield the opposite racial and political result, and that the adoption of the second plan would constitute reverse discrimination. Although there was no direct evidence that the Board drew up the first plan with the specific intention of disenfranchising its black constituents, the district court could not help but conclude that the aldermen knew what the different results of the two plans would be. The court then decided that the Board's strong preference for the first plan with its one at-large district was "merely a slightly more sophisticated device . . . for accomplishing what obviously would be impermissible if done by manipulation of district lines," 377 F.Supp. at 1200 — that is, the dilution of the black vote — so that the one at-large district operated to minimize and cancel out the local black vote for the same reasons as the traditional all-at-large system. The district court therefore rejected the Board's mixed plan and adopted its all-single-member plan. Unfortunately, the learned trial judge was forced to apply some very broad generalizations of voting rights law to a particularized problem of first impression, and we believe that he misapprehended some of the legal problems upon which his conclusions were based.

B

The Supreme Court first set a normative standard for the "one man, one vote" doctrine in Reynolds v. Sims, 1964, 377 U.S. 533, 84 S.Ct. 1362, 12 L.Ed.2d 506, when it held that "an individual's right to vote for state legislators is unconstitutionally impaired when its weight is in a substantial fashion diluted when compared with votes of citizens living in other parts of the State." 377 U.S. at 568, 84 S.Ct. at 1385, 12 L.Ed.2d at 531. In Reynolds, the Court was faced with the then-common situation wherein, as a result of legislative malapportionment, one district (usually an urban one) might contain two, five or ten times as many people as another district (usually a rural one), and yet have only the same number of representatives in the legislature. For the purposes of state government, it was as if each rural voter had two, five or ten times as many votes as each urban voter, and such "overweighting and overvaluation of the votes of those living here has the certain effect of dilution and undervaluation of the votes of those living there." 377 U.S. at 563, 84 S.Ct. at 1382, 12 L.Ed.2d at 528. The resulting discrimination against the underrepresented voters was obvious, for such malapportionment allowed most legislators to ignore the problems and interests of a substantial portion, or even of a majority, of the state's population and yet run no risk of defeat at the polls.8

Since Reynolds, the Supreme Court and the other federal courts have faced various problems connected with the "one man, one vote" principle,9 but nowhere have the courts faced such difficulty in application of the rule as when asked to determine the constitutionality of electoral systems utilizing multi-member or at-large districts. Multi-member districts are particularly troublesome because they may satisfy the "one man, one vote" standard where raw population data are concerned and yet effectively negate the voting strength of large numbers of voters. Of course, since there are winners and losers in every election, it might be said that losers' votes are always "lost," and there is clearly nothing unconstitutional about such a result in itself. But multi-member districts have a peculiar capacity to deny representation to racial or political minority groups which could or would obtain such representation if the polity in which they live were divided instead into single-member constituencies. Among the defects generally assigned to multi-member districts are: first, their tendency — by virtue of their winner-take-all aspect — to submerge minorities and to over-represent the majority party or group as compared with that party or group's polity-wide electoral position; second, a general preference for legislatures which reflect community interests as closely as possible; third, the fact that since at-large members necessarily represent their multi-member district as a whole, identifiable constituencies within the district have no one member specifically charged with representing them; and fourth, the fact that ballots in at-large elections may be bulky and confusing to voters. See, e. g., Chapman v. Meier, 1975, 420 U.S. 1, 95 S.Ct. 751, 760, 42 L.Ed.2d 766, 778; Whitcomb v. Chavis, supra, 403 U.S. at 158-59, 91 S.Ct. at 1877, 29 L.Ed.2d at 384-85; Carpeneti, Legislative Apportionment: Multi-Member Districts and Fair Representation, 120 U.Pa.L.Rev. 666 (1972); Banzhaf, Multi-Member Electoral Districts — Do They Violate the "One Man, One Vote" Principle, 75 Yale L.J. 1309 (1966); Note, Ghetto Voting and At-Large Elections: A Subtle Infringement Upon Minority Rights, 58 Geo.L.J. 989 (1970). The question is at what point these undesirable effects coalesce so as to deny legislative representation to minorities on a scale of constitutional dimensions.

For the first few years after Reynolds, the Supreme Court declined to analyze the problems presented by multi-member districts, aside from noting in Fortson v. Dorsey, 1965, 379 U.S. 433, 85 S.Ct. 498, 13 L.Ed.2d 401, Burns v. Richardson, 1966, 384 U.S. 73, 86 S.Ct. 1286, 16 L.Ed.2d 376, and Kilgarlin v. Hill, 1967, 386 U.S. 120, 87 S.Ct. 820, 17 L.Ed.2d 771, that such schemes were not unconstitutional per se, while cautioning that particular uses of the device might not withstand inspection.10 In Connor v. Johnson, 1971, 402 U.S. 690, 91 S.Ct. 1760, 29 L.Ed.2d 268, however, the Court decided that multi-member districts are sufficiently less desirable than single-member districts so that when federal courts are forced to fashion apportionment plans, they ought as a general rule to prefer single-member districts to multi-member solutions.

Then, in Whitcomb v. Chavis, 1971, 403 U.S. 124, 91 S.Ct. 1858, 29 L.Ed.2d 363, the Court was asked to decide the constitutionality of a multi-member plan for the election of state legislators in Marion County, Indiana. The plan provided that all the voters of the County would vote for all eight of the County's state senators and all fifteen of its assemblymen, and since a majority of the County's voters were in the habit of voting for white, Republican candidates, the result was that black and Democratic legislators were relatively rare in the Marion County delegation. A group of black plaintiffs attacked the scheme, on the theory that it produced relatively fewer black legislators than there were black residents in the County. The district court held that the multi-member districts illegally minimized and cancelled out black voting strength. On appeal, the Supreme Court found the multi-member plan to be constitutional and reversed the judgment of the district court.

The Court began its discussion by stressing that multi-member districts may in some cases dilute minority voting rights in an impermissible manner, particularly if the districts are large or if the election scheme lacks residence requirements (so that all of a district's representatives might live in one area of the district), thus enhancing the prospects of majoritarian monopoly of representation. That said, the Court noted that challengers of election schemes have the burden of demonstrating the unconstitutionality thereof, and concluded that the Chavis plaintiffs had failed to discharge their burden of proof of dilution. The plaintiffs and the district court had assumed that black citizens could be adequately represented only by black officials and had reasoned that a comparison of population figures and legislative strength made it clear that the multimember plan produced fewer black legislators than the black population deserved, that one black person's vote was in this way less effective than one white person's vote, and that this inequality of voting power was the very sort of abridgment of the right of suffrage forbidden by Reynolds and its progeny.

The Supreme Court held, however, that the disproportion between black residents and black legislators did not prove invidious discrimination absent evidence and findings that blacks in Marion County

403 U.S. at 149-50, 91 S.Ct. at 1872, 29 L.Ed.2d at 379-80. On the contrary, blacks were regularly slated as candidates and were elected with some frequency. Most importantly, the evidence showed that candidates and elected representatives, white as well as black, "avowed a substantial commitment to the substantive interests of black people," 403 U.S. at 150 n. 30, 91 S.Ct. at 1873, 29 L.Ed.2d at 380, and there was no evidence that the legislators charged with representing blacks in the formulation of state policy and programs had ignored the interests of their black constituents. Faced with this record, the Court determined that the major source of plaintiffs' difficulties lay in the fact that most Marion County blacks were Democrats in a Republican area, and that plaintiffs' problems were not so much in gaining access to the political process as in losing elections.

403 U.S. at 153-56, 91 S.Ct. at 1874, 29 L.Ed.2d at 381-82.

Although Chavis did not provide an opportunity for the Supreme Court to demonstrate precisely wherein multi-member plans may suffer from constitutional infirmities, a better occasion soon presented itself in White v. Regester, 1973, 412 U.S. 755, 93 S.Ct. 2332, 37 L.Ed.2d 314. In that case, the Court encountered a massive challenge to a redistricting plan for the Texas state legislature. One of the questions presented was whether multi-member districts in the Dallas and San Antonio areas operated to dilute the votes of blacks and Mexican-Americans in those localities. The evidence presented in Regester was of a very different sort than that offered in Chavis. In Regester, the district court found that Texas blacks and browns had experienced a long and recent history of pervasive discrimination with respect to voter registration and voting. The court found further that Dallas blacks were never slated as candidates by the political organizations in that city, that white candidates had often injected racial issues into campaigns, that blacks in Dallas were "generally not permitted to enter into the political process in a reliable and meaningful manner," 412 U.S. at 767, 93 S.Ct. at 2340, 37 L.Ed.2d at 325, and that Mexican-Americans in San Antonio suffered similar political disabilities. The multi-member district device was intimately related to the perpetuation of this effective disenfranchisement of minority voters, for the plan allowed the white majorities in Dallas and San Antonio totally to negate black and brown voting strength, and the device also allowed the state legislators who were elected by the white majority totally to ignore the interests of their minority constituents. Faced with an election scheme which did not then and might not ever provide any voice — white, black or brown — in the legislature for Dallas blacks and San Antonio Mexican-Americans, the district court concluded that in this case the multi-member districts could not be tolerated, for they operated to exclude substantial racial minorities not only from political victory but even from political consideration, and that such a result was inimical to those constitutional notions of fair political representation that are bound up in the concept of "the right to vote."

The Supreme Court affirmed the judgment of the district court, and so held for the first time that a multi-member plan devised by a legislature was unconstitutional. The Court rejected the notion proposed by plaintiffs that "every racial or political group has a constitutional right to be represented in the state legislature," 412 U.S. at 769, 93 S.Ct. at 2341, 37 L.Ed.2d at 326, but agreed with the district court that the totality of circumstances in Regester supported a finding of an illegal dilution of the voting rights of minority citizens. The Court repeated its stricture, first enunciated in Chavis, that in order to sustain claims that the votes of a minority group have been minimized or cancelled out:

412 U.S. at 765-66, 93 S.Ct. at 2339, 37 L.Ed.2d at 324.

The right to vote considered in Chavis and Regester implies that the voter have some reasonably meaningful participation in the choice of candidates and of policies. When the Court condemned political systems where substantial minority groups are effectively excluded from the nomination and election process, it meant that to give a black voter a choice of voting for one of three white candidates who know nothing and care less about his or her interests is to render the vote nugatory and the right meaningless. Of course, such a situation could and does occur in single-member districts where the aggrieved minority may be so small as to command no consideration by elected public officials, although in such a case it might be argued that where a minority is sufficiently small, there can be no substantial dilution of its politically insignificant vote. The particular vice of multi-member districts, however, is their tendency to minimize minority representation even at the lowest political levels in a way that could not occur if single-member districts existed in their stead. Multi-member districts thus pose a problem of degree of fair representation — "fair" not in the sense either of "considerable" or of "proportionate," but rather in a general sense of equity. Chavis and Regester represent efforts to decide whether a particular political system is fair in this general equitable sense; the two cases make the dilution doctrine an intensely practical, factually-oriented rule against fundamental unfairness. It is for this reason that the Court placed so great a reliance on the records to support or refute plaintiffs' contentions that their respective interest groups were precluded from exercising that amount of political power to which they ought in all fairness to have had access. There was nothing in the record in Chavis to indicate that this fundamental political unfairness existed in Marion County, Indiana. Conversely, in Regester, the Court made no sweeping statements about the nature of the right to vote and what specific measures must be taken to preserve it intact, but rather concluded that:

412 U.S. at 769-70, 93 S.Ct. at 2341, 37 L.Ed.2d at 326.

C

The federal district courts and courts of appeal have had many opportunities to apply the dilution doctrine to various fact situations and to distill from the Supreme Court's cases additional guidelines for decision. Chavis and Regester hold explicitly that no racial or political group has a constitutional right to be represented in the legislature in proportion to its numbers,11 so it follows that no such group is constitutionally entitled to an apportionment structure designed to maximize its political advantages. Gilbert v. Sterrett, 5 Cir. 1975, 509 F.2d 1389; Turner v. McKeithen, supra; Cousins v. City Council of City of Chicago, 7 Cir. 1974, 503 F.2d 912. Neither does any voter or group of voters have a constitutional right to be included within an electoral district that is especially favorable to the interests of one's own group, or to be excluded from a district that is dominated by some other group. Taylor v. McKeithen, 5 Cir. 1974, 499 F.2d 893; Gunderson v. Adams, S.D. Fla.1970, 328 F.Supp. 584, aff'd, 1971, 403 U.S. 913, 91 S.Ct. 2225, 29 L.Ed.2d 692; see Ferrell v. State of Oklahoma ex rel. Hall, W.D.Okl.1972, 339 F.Supp. 73, aff'd, 406 U.S. 939, 92 S.Ct. 2045, 32 L.Ed.2d 328. The critical question under Chavis and Regester is not whether the challenged political system has a demonstrably adverse effect on the political fortunes of a particular group, but whether the effect is invidiously discriminatory, that is, fundamentally unfair.

This Court has decided several dilution cases in recent years, and although we have consistently adhered to the proposition that "access to the political process and not population [is] the barometer of dilution of minority voting strength," Bradas v. Rapides Parish Police Jury, 5 Cir. 1975, 508 F.2d 1109, 1112, we believe that our decision in this case will be better understood after a discussion of two recent cases involving multi-member electoral districts: Zimmer v. McKeithen, 5 Cir. (en banc) 1973, 485 F.2d 1297, and Turner v. McKeithen, 5 Cir. 1973, 490 F.2d 191.12 Zimmer presented us with the question of whether dilution of black voting rights could occur where blacks constituted a majority of the population of a given polity but a minority of the registered voters, and where the redistricting plan provided that all the members of the school board and police jury in a rural Louisiana parish would be elected at-large.13 The district court and a panel of this Court decided that there could not be any dilution of the black vote in such a case. After an exhaustive review of the factual situation underlying the multi-member plan, the en banc Court found that dilution of the black vote could and did exist and would be perpetuated by the all-at-large plan. The existence of a black population majority was not dispositive of the dilution issue, for it is not population but access to the political process that determines whether an interest group enjoys the full vigor of its political rights. Nor could the small size of the parish ensure against dilution, for the size of a multi-member district bears only on the possibility for voter confusion and lack of voter identification with the elected representatives, neither of which is necessary to a finding of dilution. The record in Zimmer revealed a long history of racial discrimination in all aspects of local life, a lack of access by blacks to the nomination process, a majority vote requirement, an anti-single shot voting requirement and a history of bloc voting by race, all of which compelled the conclusion that the combination of the white voter majority and the all-at-large system would result in all-white school boards and all-white police juries which would pay little attention to black interests. Also of considerable importance to our result was the fact that there had been a strong Louisiana policy against at-large elections for school boards and police juries until 1968, when the pertinent state statute was changed to allow at-large elections for those governmental bodies. Although we did not find it necessary to decide whether this abrupt change in policy — which coincided with increased black voter registration — was racially motivated, we did decide that the new state policy in favor of all-at-large elections was not entitled to great deference where the record showed that the effect of the policy in this parish was to dilute black voting rights. We concluded that the facts in Zimmer were indistinguishable from those of Regester and that although some multi-member election plans might be constitutional, this one was not.

In Turner v. McKeithen, supra, we dealt with another multi-member election scheme for a Louisiana parish police jury. The record in Turner contained the usual history and lingering effects of racial discrimination and revealed that although blacks constituted a substantial portion of the parish population, no black had ever been elected to the police jury. The evidence also showed that blacks were neither considered nor consulted in the candidate slating process, but that the black vote was instead solicited "at a stage when the actual candidate selection has already occurred and the possibility for meaningful influence is significantly diminished." 490 F.2d at 195. Finally, the local policy supporting multi-member districts was the same new and dubious one examined in Zimmer. We agreed with the district court that the plan foreclosed effective black participation in the political life of the parish, and we affirmed the trial court's invalidation of the all-at-large scheme.

Zimmer and Turner are good examples of the proper application of the principles of the dilution doctrine. The Court in both cases paid close attention to the facts of the particular situations at hand, to the history of studied neglect by elected representatives of the interests of a large number of their own constituents, to the practical effects of electoral schemes which were likely to perpetuate that shameful failure of representation and to the apparent absence of any rational state or local policy in support of the all-at-large plans. We found in both cases that the only way in which blacks would be able to secure a fair hearing of their needs was to order elections to be held on the basis of those single-member districts which the defendant governing bodies themselves had only recently abandoned.

IV

With respect to the case at bar, we have already discussed the reasons why Ferriday's traditional all-at-large aldermanic election system dilutes the voting rights of local blacks in a substantial and therefore unconstitutional fashion. The situation here is in many respects very like those of Zimmer and Turner. The long history and continuing effects of racial discrimination, the failure of the local political organizations to consult blacks regarding the slating of candidates, the unresponsiveness of the elected aldermen to the needs of the black community and the Louisiana anti-single shot and majority vote requirements have combined to give the white voting majority an absolute dominion over local politics by means of the all-at-large system.14 When the district court found this to be so, the experienced trial judge drew the further conclusion that even a single at-large aldermanic slot would have the same pernicious effect that the all-at-large system had, and the court thus held that only an all-single-member plan would ensure fair political representation for Ferriday's black citizens.

We believe that the trial court was mistaken in adopting its per se rule in an area so factually-oriented and practically-based as this one, and we find that a thorough study of the Board's mixed single-member-at-large-member plan requires the conclusion that the scheme is not unconstitutional. If the mixed plan were to be adopted, the political life of Ferriday would differ in several crucial respects from its present configuration. Of primary importance is the fact that the plan assures local blacks of at least two aldermen who will necessarily be accountable to the overwhelming black population of their single-member districts. The Board's mixed plan is a great improvement over its predecessor for other reasons as well. Where there is only one at-large member to be selected in a given polity, neither the anti-single shot law nor the majority vote requirement can invidiously discriminate against minority voters. If there is only one at-large place to be filled and there are minority candidates, then minority voters can vote for minority candidates. If there are no minority candidates, the minority voter can refrain from voting for a majority candidate without voiding his other votes. In a one-party community such as Ferriday, the requirement that a candidate receive a majority of the votes cast in the first primary if he is to avoid a second primary has historically enabled the white majority in Ferriday to defeat in the second primary all black candidates fortunate enough to survive the first primary. If there is only one at-large seat, however, the majority vote requirement is no longer objectionable. Even if the voting in Ferriday continues to be along racial lines and a white candidate defeats a black candidate for the position, it would be difficult to complain about such a result since a majority of the voters are white. Of course, the same thing could be said about the majority vote requirement in an all-at-large setting; however, the effects of this electoral device, standing alone, are not particularly discriminatory in any case and are minimized where there is only one at-large position effected by the rule. The majority vote requirement will no longer be fundamentally unfair to black voters if the Board's mixed plan is implemented. Furthermore, in a community where there is a precarious balance between white and black voting strength and where black voters are in the habit of voting in their own interest, it is improbable that white candidates for the one at-large position will be so confident of the size and racial solidity of the white vote that such candidates could afford to ignore or offend the nearly 50 percent of the electorate which happens to be black.

Plaintiffs would urge that none of these practical considerations can refute the district court's conclusion that the one at-large position renders the Board's mixed plan unconstitutional, for no one denies that defendants' insistence on the one at-large seat will almost certainly prevent blacks from gaining control of the Board of Aldermen. Plaintiffs contend that the effect of the plan is ample proof of its discriminatory purpose. These arguments might be entitled to great consideration if the governmental interest advanced in support of the at-large device were of the new and questionable sort found in Zimmer and Turner, for those cases show that a tenuous state policy underlying the preference for at-large districting is a good indication that the scheme is intended to dilute the vote of minority interests. But the situation here is very different from those of Zimmer and Turner in that respect.

At-large voting in aldermanic elections has been the state policy of Louisiana since 1898; the policy is presently codified as La.Rev.Stat. 33:381. The reason usually given in support of at-large elections for municipal offices is that at-large representatives will be free from possible ward parochialism and will keep the interests of the entire city in mind as they discharge their duties. While this theory does not always hold true in practice, as the experience of Ferriday's black citizens attests, we cannot say that the rationale is so tenuous that it can be disregarded. Nor have plaintiffs demonstrated that the at-large device here was conceived as a tool of racial discrimination as appeared to be the case in Zimmer and Turner. When the enabling legislation was passed in 1898, and for the almost 70 years thereafter when the policy was in force across the state, there could have been no thought that the device was racially discriminatory, because very few blacks were allowed to vote in Louisiana during that period. As Judge Wisdom noted in Taylor v. McKeithen, supra:

499 F.2d at 896.

We would be callous indeed to tell plaintiffs that seventy years of illegality somehow legitimizes continued dilution of black voting rights, but that is not the thrust of our discussion. In order for there to be substantial — and thus illegal — impairment of minority voting rights, there must be some fundamental unfairness in the electoral system, some denial of fair representation to a particular class. The seventy years of consistent state support of at-large elections for municipal offices strongly suggests that the broad political judgment upon which the policy is based is not racially-motivated, and the lack of racial motivation is at least some evidence of a lack of discriminatory effect. The only evidence that plaintiffs have produced to demonstrate that the Board's plan will have an invidiously discriminatory effect is the suggestion that the very existence of the one at-large position will enable the white voters of Ferriday to control three aldermanic seats instead of two.15 Plaintiffs' argument is thus that black votes will be diluted unless their effect is maximized, but we have already seen that the Constitution does not require such a result. See City of Richmond v. United States, ___ U.S. ___, ___, 95 S.Ct. 2296, 2302-2304, 45 L.Ed.2d 245, 246.

Under the mixed plan, the black citizens of Ferriday will certainly command the allegiance of forty percent of the Town's aldermen, a share only slightly lower than the proportion of black voters in the entire electorate; such representation will give blacks that access to the political process which was denied to them under the all-at-large plans here, in Zimmer and in Turner. This new-found black political power is unlikely to permit the new Board to neglect black interests as former boards have done.16 We also believe that the state policy favoring at-large aldermen is not unreasonable, and that this characteristic coupled with the long existence of the policy require that we assign the policy considerable weight in the calculus of dilution. We have also shown that the objectionable aspects of the Louisiana anti-single shot and majority vote requirements will be obviated or minimized under the mixed plan. Our examination of the record and of the probable practical effects of the Board's mixed election plan convinces us that the new scheme will not substantially dilute the voting rights of Ferriday's black citizens and that the plan is not unconstitutional under the rule of Chavis, Regester and Zimmer.

V

Since we have determined that both of the Board's alternative reapportionment plans — the mixed plan and the all-single-member plan — were constitutional, we must now consider whether a federal district court which is presented with two constitutional redistricting plans should choose the one preferred by the governmental unit involved or whether the court should be free to choose the "better" of the two plans. In other words, we must decide if the district court's choice of the all-single-member plan is reviewable only under the abuse of discretion standard, as plaintiffs contend.

At the very beginning of judicial scrutiny of legislative apportionment plans, the Reynolds Court cautioned that "legislative apportionment is primarily a matter for legislative consideration and determination, and judicial relief becomes appropriate only when a legislature fails to reapportion according to federal constitutional requisites in a timely fashion after having had an adequate opportunity to do so." 377 U.S. at 586, 84 S.Ct. at 1394, 12 L.Ed.2d at 541. Since the basis of federal jurisdiction over legal assaults on legislative apportionment is a plaintiff's claim that his right to vote has been abridged in an unconstitutional fashion,17 it follows that where there is nothing in a given scheme that is repugnant to the Constitution, a federal court ought not to substitute a plan which might seem to it only to be more efficient or more just than a plan preferred by the legislature concerned.

Even when an existing state or local election scheme is unconstitutional, the Supreme Court has consistently ruled that the legislative body involved ought, to be given a reasonable opportunity to devise a constitutional plan, and that if the legislature does so, its "freedom of choice to devise [constitutional] substitutes . . . should not be restricted beyond the clear commands of the Equal Protection Clause." Burns v. Richardson, 1966, 384 U.S. 73, 85, 86 S.Ct. 1286, 1293, 16 L.Ed.2d 376, 387. In accordance with this rule of deference to state or local legislative policies which are not unconstitutional, the Court has reversed or vacated several federal court decisions where the trial court declined to accept a constitutional state or local plan or failed to accommodate a cognizable legislative policy insofar as possible, in cases where the court was itself forced to draw a reapportionment plan. See Gaffney v. Cummings, 1973, 412 U.S. 735, 93 S.Ct. 2321, 37 L.Ed.2d 298; Mahan v. Howell, supra; Sixty-Seventh Minnesota State Senate v. Beens, 1972, 406 U.S. 187, 92 S.Ct. 1477, 32 L.Ed.2d 1; Whitcomb v. Chavis, supra; Burns v. Richardson, supra.

As this Court pointed out in Reese v. Dallas County, supra, one reason for judicial deference in dilution cases is a "respect for the institutional limitations on the courts' ability to gauge the ramifications of districting patterns. . . [In dilution cases,] the courts must evaluate evidence of the political alignments of allegedly disadvantaged factions and infer the intent of the legislature from actions that may have several plausible motives." 505 F.2d at 887. Our discussion of the political situation in Ferriday is a good illustration of the often ambiguous circumstantial evidence which is frequently the only evidence available to plaintiffs in dilution suits. We have joined with plaintiffs, defendants and the district court in speculation about the probable effects of particular facets of the traditional electoral mechanism in Ferriday, and everyone concerned has necessarily engaged in predictions of the results of the Board's proposed mixed election plan. Although courts must frequently attempt to define the undefinable and render judgments on the basis of uncertain or unknowable factors, the particular difficulty in dilution cases is that the factors to be weighed are often political in nature, and everything in our political learning teaches us that legislatures are better equipped than courts to balance all of the competing interests involved in legislative apportionment, unless the legislature's balancing is so fundamentally unfair as to be unconstitutional.18 See White v. Weiser, 1973, 412 U.S. 783, 795-96, 93 S.Ct. 2348, 2354-2355, 37 L.Ed.2d 335, 346; Lytle v. Commissioners of Election, 4 Cir. 1974, 509 F.2d 1049, 1051-52; The Supreme Court, 1966 Term, 81 Harv.L.Rev. 69, 154-55 (1967); Comment, Political Gerrymandering: A Statutory Compactness Standard as an Antidote for Judicial Impotence, 41 U.Chi.L.Rev. 398, 409-11 (1974).

We might therefore conclude that the district court was in error when it failed to adopt the Board's constitutional mixed single-member-at-large-member electoral scheme. Plaintiffs contend, however, that even if the Board's mixed plan is constitutional, the district court was within the bounds of reasoned discretion when it ordered the implementation of the all-single-member plan as the more equitable alternative. They base this equitable remedy standard on our Turner decision and on the Supreme Court's recent decision in Chapman v. Meier, 1975, 420 U.S. 1, 95 S.Ct. 751, 42 L.Ed.2d 766. Plaintiffs argue that Turner and Chapman draw a distinction between apportionment plans already in operation and proposed schemes advanced by legislatures for adoption only in the event that the existing districting system is found to be unconstitutional. Even if the first variety of plan deserves judicial deference, plaintiffs contend that there is no reason for courts to defer to the legislative judgment for fear of dislocating the electoral mechanism where a new plan must be implemented in any event.

A major defect in plaintiffs' argument is that courts do not defer to legislative judgments regarding reapportionment out of a concern that judicial intervention will cost time, money or personal inconvenience to the affected parties. We have just shown that courts usually defer to legislative preferences because of the complex political judgments involved in reapportionment plans. In short, it does not matter whether the legislature's plan is operational or merely prospective. Nor does Turner offer any solace to plaintiffs. Although it is true that we affirmed the district court's choice of the black plaintiffs' reapportionment plan in that case, the fact is that the defendant police jury in Turner elected to stand on the existing all-at-large election scheme, so that when the trial court found that system to be unconstitutional, it had the choice of adopting the plaintiffs' admittedly constitutional plan or creating one of its own. There was no question of choosing one of two acceptable plans; the district court instead picked the only acceptable plan. Finally, Chapman was a case where a federal district court drew its own reapportionment plan for the North Dakota state legislature and utilized multi-member districts for the state senate. The unusual thing about this procedure was the fact that the North Dakota legislature had never used such districts in its own plans. In these circumstances, the Supreme Court applied the rule of Connor v. Johnson, supra, and held that since multi-member districts were disfavored electoral tools, federal courts genrally ought not to include them in plans of their own creation. The Court's determination was based not on any constitutional considerations but on its supervisory powers over federal courts. 420 U.S. at 17, 95 S.Ct. at 761, 42 L.Ed.2d at 779. There was no departure in Chapman from the rule of judicial deference with respect to reapportionment plans. On the contrary, the Court specifically stated that "the standards for evaluating the use of multimember districts . . . clearly differ depending on whether a federal court or a state legislature has initiated the use." 420 U.S. at 18, 95 S.Ct. at 762, 42 L.Ed.2d at 779-80, and warned that "reapportionment is primarily the duty and responsibility of the State through its legislature or other body, rather than of a federal court." 420 U.S. at 27, 95 S.Ct. at 766, 42 L.Ed.2d at 784-85.

There is no legal basis for plaintiffs' arguments in support of the district court's choice of redistricting plans. We conclude that the trial court should have adopted the mixed plan in deference to the Board of Aldermen's considered preference for a plan incorporating one at-large place into the aldermanic election scheme.

VI

The district court awarded attorneys' fees to plaintiffs on the basis of the "common benefit" and "private attorney general" rationales. The court reasoned that such an award was proper in that plaintiffs' action has rid Ferriday of a blatantly unconstitutional aldermanic election system, thereby rendering a signal service to Ferriday's black citizens — the full effectuation of their voting rights — and aiding the congressional intention embodied in 42 U.S.C. § 1983. 377 F.Supp. 1192, 1202-08. We believe that the common benefit or "common fund" rationale of Mills v. Electric Auto-Lite Co., 1970, 396 U.S. 375, 90 S.Ct. 616, 24 L.Ed.2d 593, and Sprague v. Ticonic National Bank, 1939, 307 U.S. 161, 59 S.Ct. 777, 83 L.Ed. 1184, is inapposite here because there was no fund created by the litigation, and no indication that an award of attorneys' fees against the Board of Aldermen will spread the costs of the lawsuit proportionately among the class that will benefit from this litigation. See Hall v. Cole, 1973, 412 U.S. 1, 93 S.Ct. 1943, 36 L.Ed.2d 702.

At the time of the district court's decision, its award of attorneys' fees on the private attorney general rationale was certainly justified by the general application of that rule by the lower federal courts. See, e. g., Fairley v. Patterson, 5 Cir. 1974, 493 F.2d 598; Lee v. Southern Home Sites Corp., 5 Cir. 1971, 444 F.2d 143; Yelverton v. Driggers, M.D.Ala. 1974, 370 F.Supp. 612, Sims v. Amos, M.D.Ala.1972, 340 F.Supp. 691, aff'd, 409 U.S. 942, 93 S.Ct. 290, 34 L.Ed.2d 215. In the very recent case of Alyeska Pipeline Service Co. v. Wilderness Society, 1975, ___ U.S. ___, 95 S.Ct. 1612, 44 L.Ed.2d 141, however, the Supreme Court has unequivocally declared that the private attorney general rule cannot support the award of attorneys' fees in federal cases, in the absence of a specific statutory mandate. After a thorough review of the general "American rule" against the award of attorneys' fees, the Alyeska Pipeline Court decided that the current trend towards invocation of the private attorney general exception was about to eradicate the general rule in many types of lawsuits. The Court expressed no opinion on the merits of the general rule, but rather based its decision on the consideration that Congress — and not the courts — has the responsibility to make specific exceptions to the rule against award of attorneys' fees:

___ U.S. at ___, 95 S.Ct. at 1627, 44 L.Ed.2d at 159-60.

The Alyeska Pipeline Court explicitly placed awards of attorneys' fees in section 1983 actions within the prohibition of its decision and specifically disapproved our Fairley and Lee cases which had espoused the theory upon which the district court relied. We thus have no doubt that Alyeska Pipeline precludes any award of attorneys' fees to plaintiffs on the theory that they have acted as private attorneys general.

However, the district court also found that "the Board of Aldermen's steadfast adherence to a constitutionally unacceptable plan during the course of this litigation would amount to bad faith," 377 F.Supp. at 1206, and that finding is supported by the record, although the able trial judge preferred to base his award on what at the time seemed to be the equally solid ground of the private attorney general rule. The Supreme Court recognized in Alyeska Pipeline that the "bad faith" rationale was an assertion of "inherent [equitable] power in the courts to allow attorneys' fees in particular situations," ___ U.S. at ___, 95 S.Ct. at 1622, 44 L.Ed.2d at 154, and found no defect in that exception to the general rule. See also F. D. Rich Co., Inc. v. United States for Use of Industrial Lumber Co., Inc., 1974, 417 U.S. 116, 129, 94 S.Ct. 2157, 2165, 40 L.Ed.2d 703, 713-14. Since I believe that the district court's award of attorneys' fees to plaintiffs is justified by its finding of defendants' bad faith in the defense of an obviously unconstitutional election scheme, I would affirm the trial court's order in that regard. My brothers Roney and Grooms, however, have concluded that the order awarding attorneys' fees should be vacated, for the reasons given in Judge Roney's concurring opinion.

VII

We hold today that pursuant to the rule of rational empiricism announced by Chavis and Regester, Ferriday's traditional all-at-large aldermanic election system has operated to dilute the voting rights of the Town's black citizens by depriving them of meaningful representation on the Board of Aldermen and thus denying them any hope of consideration in the affairs of the community in which they live. On the other hand, close scrutiny of the record convinces us that the fact that the all-at-large plan is unconstitutional does not mean that no at-large places at all can be tolerated in the circumstances of this case. We find that the Board's mixed plan will have no unconstitutional impact upon the voting rights of anyone in Ferriday. Due deference to the long-established Louisiana policy favoring at least some at-large positions in aldermanic elections leads us to conclude that the Board's preference for the mixed plan must override the district court's preference for the all-single-member plan.

We do not write today concerning every election plan in every city and town in this Circuit for all time to come; we deal only with a particular proposed aldermanic election scheme in Ferriday, Louisiana. We sympathize with the position of the able trial judge who, when faced with overwhelming evidence of longstanding political injustice, perhaps leaned over backwards to ensure that the wrongs of many generations were righted by his order in this case. The Constitution, however, demands not racial representation by ratio but racial equity in the political process, and although the traditional aldermanic election plan in Ferriday has illegally abridged the voting rights of local blacks, there is nothing here to show that the Board's new and very different plan will yield similarly unconstitutional results. See Dallas County v. Reese, 1975, ___ U.S. ___, 95 S.Ct. 1706, 44 L.Ed.2d 312; Dusch v. Davis, 1967, 387 U.S. 112, 117, 87 S.Ct. 1554, 1556, 18 L.Ed.2d 656, 660. We are not prepared to presume, as plaintiffs do, that no white person in Ferriday can fairly represent a black person. If our faith in the future be mistaken, if the one at-large aldermanic position should perpetuate white officeholder underrepresentation of black interests, the courthouse door will not be locked against anyone who may be effectively disenfranchised by the electoral scheme we approve today.

Affirmed in part; reversed in part; vacated in part and remanded.

GROOMS, District Judge.

I concur in Judge Goldberg's opinion, except as to his position on the award of attorney's fees. I concur in and adopt Judge Roney's opinion on that issue.

RONEY, Circuit Judge (specially concurring).

The appellants conceded before this Court that the all-at-large election scheme was unconstitutional. It seems to me that we should start with that concession and that we are relieved of any judicial necessity to decide the correctness of the district court's decision in this regard. I would not voice an opinion one way or the other, then, as to the validity of the all-at-large election scheme in this town of 5,200 people.

With that concession, the only issue for us to decide is whether the appellant town's mixed single-member-at-large-member electoral system is constitutional such to make erroneous the district court's ordered implementation of appellee's all-single-member plan. I fully concur in the decision that the Board of Aldermen's plan is constitutional and in Judge Goldberg's thorough and well-reasoned opinion dealing with this issue.

With respect to the district court's award of attorney's fees, I would vacate and remand for reconsideration by the district court in light of two factors: the demise of the private attorney general rationale in Alyeska Pipeline Service Co. v. Wilderness Society, ___ U.S. ___, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975), and our decision here that the Board of Aldermen's plan submitted to the district court is constitutional.

First, it is impossible to tell to what extent the district court's reliance on the attorney general rationale may have permeated its finding that bad faith likewise supported the attorney's fee award. The award of attorney's fees should be reconsidered by the district court on the issue of bad faith alone.

Second, the district court had decided that the plan proffered by the town was unconstitutional when it found that "the Board of Aldermen's steadfast adherence to a constitutionally unacceptable plan during the course of this litigation would amount to bad faith." 377 F.Supp. at 1206. The issue of bad faith should be reconsidered in light of the fact that, having lost its defense of the all-at-large plan, the Board did submit a constitutionally acceptable plan to the district court. It is for the district court to decide whether, in light of this fact, there was "steadfast adherence to a constitutionally unacceptable plan."

The town briefed on appeal the issue as to the district court's finding that the all-at-large plan was unconstitutional, but then conceded the point on oral argument. All of these facts should be remanded to the district court to reconsider anew the issue of bad faith. I would not indicate one way or the other which way that issue should be resolved.