Rural West Tennessee African-American Affairs Council, Inc. v. McWherter , 877 F. Supp. 1096 ( 1995 )

877 F. Supp. 1096 (1995)

RURAL WEST TENNESSEE AFRICAN-AMERICAN AFFAIRS COUNCIL, INC., et al., Plaintiffs,
v.
Ned McWHERTER, et al., Defendants.

No. 92-2407-TUBRO.

United States District Court, W.D. Tennessee, Western Division.

February 22, 1995.

Bruce S. Kramer (ACLU), Borod & Kramer, Memphis, TN, Richard Dinkins (ACLU), Williams and Dinkins, Nashville, TN, Kathleen Wilde (ACLU), American Civil Liberties Union, Atlanta, GA, Keenan R. Keller (ACLU), Davis, Polk & Wordwell, New York City, for plaintiffs.

Richard Fields (NAACP), Memphis, TN, for intervenor.

Charles W. Burson, Atty. Gen. and Reporter, Michael W. Catalano, Deputy Atty. Gen., John H. Reinhold, Asst. Atty. Gen., Nashville, TN, for defendants.

Before: MERRITT, Chief Circuit Judge; BAILEY BROWN, Circuit Judge; and TURNER, District Judge.

OPINION OF THREE-JUDGE COURT

MERRITT, Chief Circuit Judge.

In this legislative reapportionment case under § 2 of the Voting Rights Act, 42 U.S.C. § 1973, the plaintiffs seek additional majority-black seats in the state senate for west Tennessee. In a previous decision in this case, reported as Rural W. Tenn. African-American Affairs Council v. McWherter, 836 F. Supp. 453 (W.D.Tenn.1993), vacated, ___ U.S. ___, 114 S. Ct. 2775, 129 L. Ed. 2d 888 (1994) [hereinafter Rural West I], having found a violation of § 2, we required substantial statewide proportionality of majority-black districts to black voting-age population, and we awarded the plaintiffs one such additional seat in west Tennessee on proportionality grounds, as explained at length in our earlier reported opinion.

The State appealed our decision to the United States Supreme Court. The Court reversed and vacated our decision and remanded the case to us for reconsideration in light of Johnson v. De Grandy, ___ U.S. ___, 114 S. Ct. 2647, 129 L. Ed. 2d 775 (1994).

In our previous decision, we focused on statewide proportionality rather than regional proportionality in west Tennessee, and we stated one of the key questions at issue as follows:

When a choice is between two or more black ``influence' districts, most likely represented by whites, versus one majority black district represented by a black, what does § 2 require in geographical areas of racially polarized block voting with a history of prior racial discrimination?

Rural West I, 836 F.Supp. at 465. We construed Thornburg v. Gingles, 478 U.S. 30, 106 S. Ct. 2752, 92 L. Ed. 2d 25 (1986), to require the state to engage in affirmative action based on race in such situations by establishing majority-black districts in substantial proportion to the black voting-age population of the state. We held that "influence" districts would not be sufficient and could not be counted in the proportionality calculus.

We now conclude that the Supreme Court clarified its view in Johnson v. De Grandy, and we reverse our previous view requiring the State to establish an additional majority-black senate seat in west Tennessee. In De Grandy the Court appears to have held (1) that in reviewing a legislative apportionment plan, the proportion of majority-black districts to black population "is always relevant evidence in determining vote dilution, but is never itself dispositive," De Grandy, ___ U.S. at ___, 114 S.Ct. at 2664 (O'Connor, J., concurring and summarizing the holding of the Court), and (2) that a federal reapportionment court should focus on regional proportionality rather than statewide proportionality where that is the evidentiary focus of the parties in their proof in the case.

I. Previous Proceedings

Following the tabulation of the 1990 Census, the Tennessee legislature adopted 1992 Tenn.Pub. Acts 826 [hereinafter the 1992 Plan] which provided for reapportionment of Tennessee's thirty-three state senate districts.^[1] Under the 1992 Plan, three of these districts are so-called "majority-minority districts" — in other words, blacks comprise over 55% of the voting-age population in each of those districts. Two of these majority-minority districts are in Shelby County in southwest Tennessee; the third is in the middle of the state in Nashville. In addition, the 1992 Plan created three more districts in which blacks constitute over 25% of the voting-age population. Two of these districts are in west Tennessee and the third is in southeastern Tennessee and includes part of Chattanooga.

The Rural West Tennessee African-American Council filed suit in our court shortly after the passage of the 1992 Plan, claiming that the plan violated § 2 of the Voting Rights Act in that it caused vote dilution. The plaintiffs' central argument rested on the fact that blacks constitute 14.4% of Tennessee's voting-age population, but only 9.1% (3 of 33) of Tennessee's state senate districts are majority-minority districts under the 1992 Plan. In Rural West I we agreed with the plaintiffs and ordered the Tennessee legislature to draw a new districting plan with at least four majority-minority districts. Adding one more such district, we reasoned, would mean that 12.1% of the senate districts would be majority-minority districts. If we had required the legislature to add a fifth majority-minority district, as the plaintiffs requested, then 15.2% percent of the districts would have been majority-minority districts, a figure slightly higher than the 14.4% of the voting-age population that is black. We decided not to order the legislature to create a fifth district because to do so would "provide slightly more representation than the state's black voting-age population requires.... The issue of creating a fifth black Senate district should be left to the political judgment of the legislature. It is not required by federal law." Rural West I, 836 F.Supp. at 467. The implementation of our ruling was stayed pending an appeal to the Supreme Court.

The Supreme Court vacated and remanded our decision in Rural West I, ___ U.S. ___, 114 S. Ct. 2775, 129 L. Ed. 2d 888 (1994). The Court ordered us to reconsider our holding in light of its recently decided case, Johnson v. De Grandy, ___ U.S. ___, 114 S. Ct. 2647, 129 L. Ed. 2d 775 (1994). We should also note that since this case was first filed, Tennessee has held the 1992 and 1994 senate elections under the 1992 Plan. Given the Court's holding in De Grandy, we now believe that the 1992 Plan conforms to the Voting Rights Act.

II. Factual Findings

We note at the outset that we believe all of our factual findings in Rural West I to be correct, but we now reconsider our analysis of one of the plaintiffs' factual assertions. In order to assert a § 2 voting rights claim, a plaintiff must as a threshold matter demonstrate the existence of the three "Gingles preconditions." In Thornburg v. Gingles, the Supreme Court established a three-part test that a plaintiff must satisfy at a minimum before a court may inquire further into a voting rights claim. Those preconditions provide that:

1. The plaintiffs must demonstrate that the protected group is sufficiently large and geographically compact that it could constitute an effective majority in a single-member district.

2. The plaintiffs must show that the protected group is politically cohesive.

3. The plaintiffs must show that the majority votes sufficiently as a bloc to enable it usually to defeat the protected group's preferred candidate.

Gingles, 478 U.S. at 51, 106 S.Ct. at 2766-67. The presence of the first two preconditions was not contested by the parties, and after some analysis we agreed that the plaintiffs had demonstrated the presence of the third precondition. Rural West I, 836 F.Supp. at 456-60. We hold that today all three Gingles preconditions are still satisfied.

Nonetheless, once a court has determined that plaintiffs have demonstrated the presence of the Gingles preconditions, it must then proceed to examine the "totality of the circumstances." The "totality of the circumstances" language originates in the statutory text itself.^[2] In order to assist courts in evaluating the totality of the circumstances, the Senate Judiciary Committee listed several factors in the legislative history of the 1982 amendments to the Voting Rights Act.^[3] We concluded in Rural West I that the plaintiffs demonstrated a history of official discrimination in voting; racial polarization in voting; effects of discrimination in such areas as education, employment, and health; and that black candidates had not achieved sufficient electoral success. On the other hand, we found insufficient evidence to conclude that there were racial appeals in campaigns, or that elected officials were unresponsive to their black constituents. Finally, we found it difficult to evaluate the state's interests in adopting the 1992 plan in the context of a § 2 claim. Rural West I, 836 F.Supp. at 460-65. Significantly, however, we confined our evaluation of the "totality of the circumstances" to those factors listed in the Senate Report. Nonetheless, based on the Gingles preconditions and the factors listed in the Senate Report we found that the 1992 Plan violated § 2 of the Voting Rights Act. With one exception, our analysis of the considerations enumerated in the Senate Report remains valid. But, we amend our finding on whether black candidates have achieved sufficient electoral success.

In light of De Grandy, we now believe that we erred in two respects in considering the "totality of the circumstances" in Rural West I. First, we deliberately, but improperly, excluded consideration of "influence districts" when we weighed the "totality of the circumstances." Second, in evaluating the seventh factor listed in the Senate Report, "the extent to which members of the minority group have been elected to public office in the jurisdiction," as well as in fashioning a remedy, we did not correctly apply the concept of "proportionality." We now address each of these issues in turn.

III. Influence Districts

In our prior decision, we gave extensive consideration to the defendants' argument that we should consider as part of the totality of the circumstances the several influence districts which exist under the 1992 Plan. Ultimately we decided that the Voting Rights Act does not recognize influence districts. Rural West I, 836 F.Supp. at 465-66. Upon reconsideration, we now reach a different conclusion. The population of an influence district, as its name suggests, includes sufficient members of a minority group to influence substantially an election, but not enough to comprise a majority or super-majority as is necessary for a majority-minority district. While we have defined a majority-minority district as having at least a 55% minority voting-age population, 836 F.Supp. at 467, an influence district exists when members of a minority group compose 25% or more of the voting-age population of a district. In some cases, an influence district may also exist when a minority group consists of less than 25% of the voting-age population of a district.

The issue of influence districts may arise in voting-rights jurisprudence in two distinct contexts. It will be useful to distinguish these two situations before proceeding further.

Most commonly, plaintiffs have contended that § 2 may mandate the creation of influence districts in certain situations. Frequently this argument has been advanced in cases in which the minority group in question does not have sufficient population to comprise a majority in a majority-minority district, but nonetheless wants to be part of an influential minority in a single district. If the members of a minority group taken together can only comprise 35% of the voting-age population of one district, no matter how the district is drawn, then it is clear that they do not satisfy the first Gingles precondition that requires that "the plaintiffs must be able to demonstrate that the protected group is sufficiently large and geographically compact that it could constitute an effective majority in a single member district." Gingles, 478 U.S. at 50, 106 S.Ct. at 2766. This issue of whether § 2 may require the creation of an influence district can also arise when a minority group does have sufficient population to satisfy the first Gingles precondition but that population is already largely included within majority-minority districts. In such a case, plaintiffs may argue that the remainder population, not included in a majority-minority district, and not numerous enough to form an additional majority-minority district must be grouped together to form an influence district rather than be divided in small numbers among other districts. Again, members of this remainder population cannot satisfy the first Gingles precondition.

Most courts have rejected the contention that § 2 entitles minority groups to an influence district even if they cannot satisfy the first Gingles precondition. See Latino Political Action Comm., Inc. v. Boston, 784 F.2d 409 (1st Cir.1986); DeBaca v. County of San Diego, 794 F. Supp. 990 (S.D.Cal.1992), aff'd, 5 F.3d 535 (9th Cir.1993); Hastert v. State Bd. of Elections, 777 F. Supp. 634 (N.D.Ill. 1991); Skorepa v. Chula Vista, 723 F. Supp. 1384 (S.D.Cal.1989). Only one court has recognized that § 2 may require the creation of influence districts. Armour v. Ohio, 775 F. Supp. 1044 (N.D.Ohio 1991). Another court held that a plaintiff would not have to satisfy the Gingles "effective majority" precondition when he could demonstrate intentional discrimination. Garza v. County of Los Angeles, 918 F.2d 763 (9th Cir.1990), cert. denied, 498 U.S. 1028, 111 S. Ct. 681, 112 L. Ed. 2d 673 (1991). Despite this difference of opinion on when or if § 2 requires legislatures to create influence districts for minority groups, the Supreme Court on four separate occasions has refused to rule on the issue. De Grandy, ___ U.S. at ___, 114 S.Ct. at 2656; Voinovich v. Quilter, ___ U.S. ___, ___, 113 S. Ct. 1149, 1155, 122 L. Ed. 2d 500 (1993); Growe v. Emison, ___ U.S. ___, ___ n. 5, 113 S. Ct. 1075, 1085 n. 5, 122 L. Ed. 2d 388 (1993); Thornburg v. Gingles, 478 U.S. at 46 n. 12, 106 S.Ct. at 2764 n. 12. Thus the law remains unsettled on this question. Nonetheless, the problem of whether the Voting Rights Act requires the creation of influence districts does not arise in the present case.

Influence districts may also arise in the context of § 2 cases when a court evaluates the "totality of the circumstances." In this posture, the question is not whether the Voting Rights Act requires the creation of influence districts, but whether the voluntary creation of influence districts by a legislature should be counted as a factor weighing against finding a § 2 violation. To date, other than our own previous decision, there have been no cases that explicitly address this issue. Because it arises in the case now before us, we extrapolate from Supreme Court precedent to determine whether influence districts should be counted in evaluating the totality of the circumstances.

As a threshold matter, we note that the De Grandy Court reemphasized that the inquiry into the "totality of the circumstances" must be based on "a searching practical evaluation of the ``past and present reality.'" De Grandy, ___ U.S. at ___, 114 S.Ct. at 2660 (quoting Senate Report at 208). Thus, the Court has instructed that the inquiry should be quite broad and take into account as many relevant factors as possible in evaluating the "totality of the circumstances." It should not be limited merely to the factors listed in the Senate Report. Therefore, the degree to which minority voters can claim strong or weak influence in districts other than majority-minority districts certainly should be included in the calculus.

In addition, the Supreme Court has suggested in several instances that considering influence districts is appropriate. The fundamental interpretive question at stake revolves around the meaning of the clause in § 2 which provides that a Voting Rights Act violation exists when members of a minority group have "less opportunity than other members of the electorate ... to elect representatives of their choice." 42 U.S.C. § 1973(b). The plaintiffs in the present case have argued that black voters do not have the opportunity to elect representatives of their choice if they are not part of a 55% black majority within a district. In other words, according to the plaintiffs, unless black voters have the opportunity to elect their first-choice candidate, a black candidate, they do not have the opportunity to elect a candidate of their choice at all. But the De Grandy Court rejected this interpretation of § 2 when it stated:

[F]or all the virtues of majority-minority districts as remedial devices, they rely on a quintessentially race-conscious calculus aptly described as the "politics of second best." ... If the lesson of Gingles is that society's racial and ethnic cleavages sometimes necessitate majority-minority districts to ensure equal political and electoral opportunity, that should not obscure the fact that there are communities in which minority citizens are able to form coalitions with voters from other racial and ethnic groups, having no need to be a majority within a single district in order to elect candidates of their choice. Those candidates may not represent perfection to every minority voter, but minority voters are not immune from the obligation to pull, haul and trade to find common political ground, the virtue of which is not to be slighted in applying a statute meant to hasten the waning of racism in American politics.

De Grandy, ___ U.S. at ___, 114 S.Ct. at 2661. When minority voters can influence elections, they have the opportunity to "pull, haul and trade to find common political ground." When such an opportunity exists, they are obliged to use it. We should not reject the existence of influence districts as irrelevant under § 2.

This statement in De Grandy accords with earlier decisions of the Court. In Voinovich v. Quilter, the Court also questioned whether the creation of majority-minority districts is the only means through which minority voters can be given the opportunity to vote for candidates of their choice:

[T]he creation of majority-minority districts[] does not invariably minimize or maximize minority voting strength. Instead, it can have either effect or neither. On the one hand, creating majority-black districts necessarily leaves fewer black voters and therefore diminishes black-voter influence in predominantly white districts. On the other hand, the creation of majority-black districts can enhance the influence of black voters.

Voinovich, ___ U.S. at ___, 113 S.Ct. at 1156. The Voinovich Court recognized, as we did in Rural West I, that in some cases the creation of majority-minority districts can actually weaken the influence a minority group has on the legislative process. See Rural West I, 836 F.Supp. at 465-66.

Furthermore, in Chisom v. Roemer, 501 U.S. 380, 111 S. Ct. 2354, 115 L. Ed. 2d 348 (1991), the Supreme Court rejected "the erroneous assumption that a small group of voters can never influence the outcome of an election." Id. at 397 n. 24, 111 S.Ct. at 2365 n. 24. This statement was made in response to an argument presented in dissent by Justice Scalia. The Court held that in order to demonstrate a § 2 violation, the plaintiffs would have to demonstrate they had been deprived of both their opportunity "to participate in the political process," and "the opportunity to elect representatives of their choice." Justice Scalia responded that such a reading would leave small minority groups who had been denied the opportunity to participate with no protection under the Voting Rights Act. Small groups would never be able to demonstrate that they had been denied the opportunity to elect representatives of their choice because their size would prevent them from ever having such an opportunity. Thus members of small minority groups who had been denied access at the polls would have no redress. The Court answered that small groups of voters can influence an election, and thus could meet the second requirement of the statute. In making this argument, the Court explicitly recognized that at least in the Chisom context the phrase "elect representatives of their choice" may be read as meaning having the ability to influence the outcome of an election.

Finally, in Gingles, Justice O'Connor's concurrence raised the issue of whether influence districts may be considered by courts together with majority-minority districts in assessing a voting rights claim. She asked, "Is the ``voting strength' of a racial group to be assessed solely with reference to its prospects for electoral success, or should courts look at other avenues of political influence open to the racial group?" Gingles, 478 U.S. at 87, 106 S.Ct. at 2785 (O'Connor, J., concurring). The Court's jurisprudence since Gingles has not answered this question explicitly, but, to remain consistent with recent cases, we believe that courts should consider the existence of "other avenues of political influence" besides majority-minority districts.

The plaintiffs in this case have presented only one argument as to why influence districts should be ignored under an analysis of the "totality of the circumstances." They point to the legislative history of the 1982 amendments to the Voting Rights Act and cite statements by opponents of the amendments who argued that § 2 would effectively require courts to create majority-minority districts under a quota system (a system of proportional representation). In particular, the plaintiffs point to the objections of Senator Orrin Hatch and the Subcommittee on the Constitution. Senate Report at 94-105, 107-87. The plaintiffs argue that the opponents of the bill presented arguments for a more flexible understanding of § 2 which would allow for consideration of influence districts. Because the opponents of the amendments lost, the plaintiffs contend that Congress rejected their arguments. Consequently, the plaintiffs would have us believe that Congress adopted a law that recognizes only the presence or absence of majority-minority districts, and does not permit courts to consider influence districts in the calculus of evaluating voting rights claims.

Plaintiffs have misread the legislative history. The relevant sections of the legislative history concern a debate on a portion of § 2 that was eventually adopted called the Dole Proviso that states "nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population." Senator Hatch and the Subcommittee Report objected that the Dole Proviso simply would not be effective at achieving what its authors intended. Senator Hatch argued that despite the language expressly rejecting a requirement of proportional representation, the Dole Proviso would "fail to overcome the clear and inevitable mandate for proportional representation." Id. at 96. He believed that adopting the "results test" would require courts to impose proportional representation through the remedial phase, and thereby force courts to ignore the language disclaiming proportional representation.

The Subcommittee on the Constitution, chaired by Senator Hatch, considered an earlier version of the Dole Proviso proposed in the House of Representatives known as the "disclaimer provision." The Subcommittee disagreed with the conclusions of a House report which stated that the disclaimer provision would prevent § 2 from being construed as creating a right to proportional representation. The Subcommittee called the House conclusions "a misleading and irrelevant comment on the likely effect of the statutory reference to proportionality." Id. at 142. The Subcommittee Report went on to cite favorably the comments of Professor Irving Younger, who testified that "the disclaimer [would] likely be wholly ineffective because it is ``simply incoherent.' ... ``If enacted, that saving sentence will either be rewritten by the courts or ignored, in either event dishonoring Congress' responsibility to write the Nation's laws.'" Id. at 145. In short, Senator Hatch and the Subcommittee believed that the amendments to § 2 would require proportional representation irrespective of any express language to the contrary included in the statute.

As we read the legislative history, the majority rejected Senator Hatch's arguments that the Dole Proviso would be ineffective. Those who voted in the majority believed the Dole Proviso would have efficacy because its express command would be sufficient to prevent courts from imposing a right to proportional representation. Whether the majority or Senator Hatch was correct is not for us to decide. Instead, we simply note that the legislative history does not speak at all to what role influence districts should play in voting rights claims. In addition, both proponents and opponents of the 1982 amendments agreed that § 2 should not require proportional representation. They disagreed on how best to accomplish that purpose. Thus nothing in the legislative history even hints that consideration of influence districts would be improper.

Before analyzing the role that influence districts play in the 1992 reapportionment of the Tennessee Senate, it is necessary to define at what point a minority group gains sufficient influence in an electoral district for recognition under § 2. Some have argued for a highly particularized definition of an "influence district" that takes into account the degree of racial polarization in a specific district, the outcome of individual elections and the actual ability of minority voters to build coalitions within a district. See J. Morgan Kousser, Beyond Gingles: Influence Districts and the Pragmatic Tradition in Voting Rights Law, 27 U.S.F.L.Rev. 551 (1993). While such inquiries might provide the most precise measurements of the degree of influence exercised by a minority group within a particular district, they will vary from election to election. In a close election between two candidates in a predominantly white district in which each had the support of 49% of the population, a 2% minority population could decide the election by throwing its support to one of the two candidates. In such a situation, that minority group would have tremendous influence on the outcome of the election. But in the same district in another election, the same 2% of the population might be entirely irrelevant to the outcome if the contest were not as close. Thus courts would be obliged to engage in a new analysis every election. This would effectively make evaluation of influence districts unworkable.

Instead, we favor adoption of a standard rule that any district in which a minority group composes 25% to 55% of the voting-age population be considered an influence district.^[4] Standard rules enhance consistency, manageability, predictability and uniformity. Cases that involve elections and voting patterns often may require complex calculations which integrate many factors. Rather than continually wade through a morass of statistical information about electoral districts, in many instances courts have established standard rules. In the one-person, one-vote context, the Supreme Court has established a 10% de minimis rule that allows the deviation in population between state legislative electoral districts to vary up to 10% before an Equal Protection claim becomes cognizable. Brown v. Thomson, 462 U.S. 835, 842, 103 S. Ct. 2690, 2695-96, 77 L. Ed. 2d 214 (1983). Furthermore, in the § 2 context, courts already use a threshold minimum percentage of the voting-age population to determine whether a majority-minority district exists (even though they do not always agree on what that percentage should be). Rural West I, 836 F.Supp. at 467 (stating that majority-minority district requires minority group comprises at least 55% of voting-age population in district). But see Ketchum v. Byrne, 740 F.2d 1398 (7th Cir.1984) (establishing 65% as minimum minority population necessary for majority-minority district).^[5]

We are confident that a minority population that votes as a bloc (as required by the second Gingles precondition) and comprises at least 25% of the voting-age population in an electoral district will have significant influence on candidates in virtually every election. A serious candidate for office cannot ignore 25% of the population that tends to vote as a bloc. An analysis of the two Tennessee Senate elections conducted under the 1992 Plan supports this conclusion. In the 1992 and 1994 elections combined there were twenty-one contested senate races.^[6] In every contested race, if 25% of those who voted had changed their votes from the winning candidate to the losing candidate, they would have reversed the results of the election. Furthermore, in fifteen of the twenty-one contested races, if 25% of those who voted and also cast ballots for the winning candidate withdrew their support and did not vote at all, they would have changed the outcome of the election. Realities such as these are difficult for candidates to ignore. See Abigail Thernstrom, Whose Votes Count? Affirmative Action and Minority Voting Rights 109 (1987) (discussing influence of black voters on U.S. Senator Strom Thurmond).

Testimony from two senators who represent districts with black populations over 20% reinforces the conclusion that a minority group does not need a majority to have influence on an election or a particular senator. Senator John Wilder, who is white, represents District 26 which has a 21% black voting-age population. Senator Wilder testified convincingly that he has sought to further the interests of his black constituents by sponsoring and supporting particular projects of interest to them. (Trial Tr. at 851-54). He also observed that in one recent primary election, he defeated two black challengers, receiving substantial support even in predominantly black communities. (Trial Tr. at 842-44, 928-32). In addition, Senator Wilder recounted how as Speaker of the Senate he had used his authority to ensure that black senators became chairmen of committees. (Trial Tr. at 848-49). Regardless of whether he is opposed by a white candidate or a black candidate, Senator Wilder testified that he always works hard for the support of the black community and is strongly influenced by their interests. (Trial Tr. at 845-46). Even though the 21% black population in District 26 does not suffice to meet our definition of an influence district, it is clear that his black constituents have great influence on Senator Wilder and on the outcome of elections in his district nonetheless. Without the support of black voters he could not be assured of reelection. (Trial Tr. at 846).

Senator Stephen Cohen represents District 30 which has a 33% black voting-age population. Senator Cohen, who is white, testified that without the support of black voters he most likely would have lost races in 1988 and 1992, both of which he won. (Trial Tr. at 888-89). He described how he has received support from prominent black leaders in his district and how he campaigns in black communities within his district. (Trial Tr. at 889-90, 892). Like Senator Wilder, it is clear that black voters choose to vote for Senator Cohen, actively support him, and in turn can expect that he will represent their interests in the Tennessee Senate.

The most probative aspect of Senator Cohen's testimony, however, concerned the role that senators elected from influence districts play in the political dynamics of the Senate. Specifically Senator Cohen persuasively testified that adding an additional majority-minority district in western Tennessee would actually reduce the influence of black voters in the Tennessee Senate. He cited as an example the legislative proposal to make the birthday of Martin Luther King a state holiday, a bill which passed the Senate by only one vote (17 to 16). Senator Cohen and another senator who represents a district with a substantial black population in west Tennessee both voted for the bill. Senator Cohen argued that by removing the black voters from the influence districts to create a new majority-minority district, at least one more conservative white senator would also be elected. This new conservative senator, elected from an overwhelmingly white district, would be uninfluenced by black voters and would have been inclined to vote against the Martin Luther King holiday, a measure strongly favored by black voters but opposed by many white voters in west Tennessee. As a consequence the measure would not have passed. (Trial Tr. at 920-23).

In addition, Senator Cohen named several other issues of interest to the black community which would not have passed if his influence district and another district with significant black population had been eliminated to form a majority-minority district. In particular, he cited a gambling bill and several funding bills of special concern to his district. (Trial Tr. at 922, 954-55).

Based on the statistical results of the 1992 and 1994 elections, as well as the testimony of two members of the senate, when a minority group comprises 25% or more of the voting-age population, members of that group will have sufficient political influence to deem that district an influence district. A serious political candidate cannot ignore such a sizable portion of the electorate. Furthermore, Senator Cohen's testimony reveals that black voters in west Tennessee will be able to achieve their political aims more effectively with a combination of majority-minority districts and influence districts as opposed to only majority-minority districts. The voting patterns within the influence districts also reveal cross-racial support for candidates, which means that racial polarization has been reduced somewhat in those districts, a result that comports with the goals of the Voting Rights Act.

The 1992 Plan creates three majority-minority districts in which blacks comprise at least 55% of the voting-age population.^[7] In addition, the Plan provides three influence districts in which the voting-age population is at least 30% black.^[8] Finally, there are another two districts in which the voting-age population is over 20% black, but less than 25% black.^[9] Considering only the three majority-minority districts and the three influence districts, we cannot say based on the totality of the circumstances that the 1992 Plan violates § 2 of the Voting Rights Act.^[10] Together these six districts comprise 18.2% of all districts. We reach this decision even though the plaintiffs have demonstrated that the 1992 Plan does not result in proportionality, that the Gingles preconditions have been met, and that several factors listed in the Senate Report exist.

Black voters in the influence districts have not necessarily been deprived of the opportunity to vote for candidates of their choice, even if they do not always have the opportunity to elect what some consider a perfect candidate, i.e., a candidate of their own race. Furthermore, the plaintiffs have not demonstrated that replacing two of the influence districts with one majority-minority district would reduce vote dilution. To the contrary, it appears as though black voters might have more influence on the legislative process with two strong influence districts than they would with one additional majority-minority district.^[11] Thus, when it is not demonstrated by the evidence that a reapportionment plan violates federal law, we must defer to the political decisions made by the elected legislature. See Upham v. Seamon, 456 U.S. 37, 41, 102 S. Ct. 1518, 1521, 71 L. Ed. 2d 725 (1982); Wise v. Lipscomb, 437 U.S. 535, 539, 98 S. Ct. 2493, 2496-97, 57 L. Ed. 2d 411 (1978); Connor v. Finch, 431 U.S. 407, 414-15, 97 S. Ct. 1828, 1833-34, 52 L. Ed. 2d 465 (1977); Chapman v. Meier, 420 U.S. 1, 27, 95 S. Ct. 751, 766, 42 L. Ed. 2d 766 (1975); Gaffney v. Cummings, 412 U.S. 735, 749 (1973); Burns v. Richardson, 384 U.S. 73, 84-85, 86 S. Ct. 1286, 1292-93, 16 L. Ed. 2d 376 (1966); Reynolds v. Sims, 377 U.S. 533, 586, 84 S. Ct. 1362, 1394, 12 L. Ed. 2d 506 (1964). Consequently, we hold that the 1992 Plan does not violate § 2 of the Voting Rights Act.

IV. Proportionality

Our decision on influence districts is reinforced by a reanalysis of the role of proportionality in § 2 cases. We conclude that § 2 of the Voting Rights Act does not require proportionality. In this case to require proportionality would be tantamount to violating the Dole Proviso because of the particular nature of racial polarization in Tennessee, as discussed below.

Voting rights doctrine has produced two distinct, but often confused concepts to measure the degree to which members of a minority group have been denied a fair opportunity to participate in electoral processes: "proportional representation" and "proportionality." The De Grandy Court clarified for us the difference between them:

"Proportionality" ... links the number of majority-minority districts to minority members' share of the relevant population. The concept is distinct from the subject of the proportional representation clause of § 2, which provides that "nothing in this section establishes a right to have members of a protected class elected in numbers equal to their proportion in the population." 42 U.S.C. § 1973(b). This proviso speaks to the success of the minority candidates, as distinct from the political or electoral power of minority voters.

De Grandy, ___ U.S. at ___ n. 11, 114 S.Ct. at 2658 n. 11.^[12] Thus were it not for the language of the Dole Proviso, a requirement of proportional representation would mean that because blacks comprise 14.4% of the voting-age population, 14.4% of the elected senators must also be black. On the other hand, an obligation to create proportionality would mean that the percentage of blacks in the voting-age population would require 14.4% of the districts to be majority-minority districts.

We recognized in Rural West I that the Voting Rights Act does not require proportional representation. Rural West I, 836 F.Supp. at 463. We believed that the statute did require proportionality when the Gingles preconditions have been met and dilution of the black vote has been proven under the totality of the circumstances. In considering the totality of the circumstances, we found the mere fact that one more majority-minority district could be created without exceeding proportionality central to our holding. But, the De Grandy Court made clear that the Voting Rights Act requires neither proportional representation nor proportionality.

De Grandy concerned a § 2 challenge to a reapportionment plan for Florida's state senate and house of representatives. Narrowly construed, De Grandy held that the Voting Rights Act does not require legislatures to maximize the number of majority-minority districts, particularly when a reapportionment plan already has achieved proportionality. De Grandy, ___ U.S. at ___ - ___, 114 S.Ct. at 2662-63. Yet, the De Grandy Court also addressed the role that proportionality plays generally in cases brought under § 2. The State of Florida argued that because under its reapportionment plan proportionality did exist, it should be immune from a § 2 attack. The Court did not accept this argument that the existence of proportionality creates a "safe harbor" for defendants. Id. at ___ - ___, 114 S.Ct. at 2660-61.

But the Court also rejected the idea that lack of proportionality should create a "safe harbor" for plaintiffs bringing a § 2 claim. In discussing the role that proportionality plays in a vote dilution claim, the Court observed "the degree of probative value assigned to disproportionality in a case where it is shown, will vary not only with the degree of disproportionality, but with other factors as well." Id. at ___ n. 17, 114 S.Ct. at 2661 n. 17. In case the Court left any doubt on this issue, Justice O'Connor, in addition to joining the Opinion of the Court, added a concurrence. She stated that "lack of proportionality can never by itself prove dilution, for courts must always carefully and searchingly review the totality of the circumstances, including the extent to which minority groups have access to the political process." Id. at ___, 114 S.Ct. at 2664 (O'Connor, J., concurring).

The Court's decision to interpret the Voting Rights Act as not requiring proportionality makes eminent sense. The Dole Proviso in the Voting Rights Act establishes that the Act does not create a right to proportional representation. But when extreme racial polarization exists, as in the case before us, there is no practical difference between proportional representation and proportionality. Technically, § 2 only requires that a court determine that white voters will refuse to vote for black-preferred candidates. Therefore, theoretically one could find a Voting Rights Act violation without holding that black voters will almost always choose to vote for a black candidate rather than a white candidate, if given the choice. As a practical matter, however, in most racially polarized districts where white voters prefer white candidates (as is effectively required by the third Gingles precondition to find a § 2 violation), black voters will choose to vote for black candidates. This is certainly true in Tennessee. Rural West I, 836 F.Supp. at 460 (finding 86.30% of black voters vote for black candidates in black-white elections). But see Thernstrom, supra at 218 (noting that some majority-black districts will continue to reelect preexisting white incumbents until they retire). Thus, in effect, a majority-minority district in which the voting-age population is at least 55% black virtually always will elect a black representative. Requiring a proportional number of majority-minority districts when racial polarization exists, therefore, essentially results in the election of a proportional number of black representatives. To mandate proportionality under the Voting Rights Act when the Act in its own terms clearly states that it does not require proportional representation would be nonsensical.

We agree with the conclusion reached by other federal courts that findings of racial polarization and historical inequality do not alone necessarily result in a violation of § 2 of the Voting Rights Act. This is true even when a reapportionment plan does not provide for substantial proportionality. See Marylanders for Fair Representation v. Schaefer, 849 F. Supp. 1022 (D.Md.1994); Jeffers v. Clinton, 730 F. Supp. 196, 215-17 (E.D.Ark. 1989), aff'd, 498 U.S. 1019, 111 S. Ct. 662, 112 L. Ed. 2d 656 (1991) (rejecting voting rights claim as to Pulaski County). The De Grandy Court appears to have come to the same conclusion when it noted that the district court in that case inflated "the relative importance of the Gingles factors and of historical discrimination, measured against evidence tending to show that in spite of these facts, [the reapportionment plan] would provide minority voters with an equal measure of political and electoral opportunity." De Grandy, ___ U.S. at ___, 114 S.Ct. at 2658.

The 1992 Plan does not provide sufficient majority-minority districts to create proportionality. To reach proportionality, it would have to provide four majority-minority districts, whereas it only establishes three. We earlier considered this disparity between the black percentage of the voting-age population and the percentage of majority-minority districts to be the dominant fact in our evaluation of the totality of the circumstances. In addition, once we found a violation, we assumed that requiring proportionality would be the only potential remedy. Rural West I, 836 F.Supp. at 466-67. We now recognize this to be an incorrect understanding of the law.

We addressed the proportionality issue in the context of evaluating the seventh Senate Report factor which calls for an evaluation of "the extent to which members of the minority group have been elected to public office." In making our analysis of this factor we used the concepts of proportionality and proportional representation interchangeably. But the inquiry into the "totality of the circumstances" permits consideration of both means of analysis. De Grandy, ___ U.S. at ___ - ___, ___, 114 S.Ct. at 2660-61, 2664. Consequently we believe any confusion in evaluating the seventh Senate Report factor in our earlier opinion to be harmless.^[13] Finally, once we found a violation, we assumed that requiring proportionality would be the only potential remedy. Rural West I, 836 F.Supp. at 466-67.

Despite the overemphasis on lack of proportionality, on balance, were we still unable to consider the existence of influence districts, we would reinstate our holding in Rural West I. Lack of proportionality certainly plays a role in evaluating the "totality of the circumstances." Were there no influence districts included in the 1992 Plan, the lack of proportionality in this case combined with other factors would be sufficient to find that § 2 had been violated. But once we include influence districts in the calculus, the lack of proportionality is not significant enough to compel a holding that the 1992 Plan violates the Voting Rights Act.

V. Regional Analysis

The defendants urge that De Grandy also stands for the proposition that a court should concentrate its analysis of whether a reapportionment plan violates § 2 on the geographic region presented in the proofs by the parties. In this case, the defendants correctly state that virtually all of the evidence presented in this case concerned west Tennessee. But, in Rural West I, we reached our holding, in part, based on statewide statistics. Rural West I, 836 F.Supp. at 462.

The De Grandy Court did not authoritatively state what frame of reference a court must use in evaluating a § 2 claim. Instead, the Court noted that the parties had "apparently agreed in the District Court on the appropriate geographical scope" — a regional rather than statewide scope. De Grandy, ___ U.S. at ___, 114 S.Ct. at 2662. Because these arguments based on a statewide perspective had not been raised in the District Court, the Supreme Court treated them as waived. Id. See also NAACP v. Austin, 857 F. Supp. 560, 568-69 (E.D.Mich.1994). Nothing in De Grandy, however, prohibits a district court from considering statewide statistics, even if the proofs largely concentrate on a particular region of a state. Id.

Nonetheless, we believe we should have evaluated more closely regional statistics in our earlier decision in addition to examining statewide evidence. The difficulty in using regional statistics in this case arises because there are several equally valid ways to decide precisely which districts should be included in a regional analysis. Proofs in this case focused on seven counties in south-west Tennessee: Shelby, Tipton, Lauderdale, Fayette, Haywood, Madison, and Hardemann. Under the 1992 Plan, six senate districts fall entirely within the seven counties, and parts of two other districts are located within this region. Of the six districts which fall entirely within the seven-county region, two are majority-minority districts and two are influence districts. Both of the districts which are partially within the region have a 21% black voting-age population.^[14] Blacks comprise 37.9% of the voting-age population within these seven counties, and 41.6% of the total population.

The first means of analysis posits that this region as a whole should be entitled to either six or seven complete senate districts. If we use the total population living in this region, it would be entitled to 6.99 senate seats; when we use voting-age population it would be entitled to 6.76 seats. If we accept the defendants' claim that the region should only be entitled to the six complete seats it now has, only two of those seats must be majority-minority districts to achieve substantial proportionality in the region.^[15] On the other hand if we accept the plaintiffs' argument that the region is entitled to seven complete seats, the two existing majority-minority districts still suffice to create substantial proportionality within the seven counties.^[16]

This line of reasoning could have led us in Rural West I to conclude that no § 2 violation exists. We recognize that De Grandy has stated that the existence of proportionality does not create a "safe harbor" for defendants. De Grandy, ___ U.S. at ___ - ___, 114 S.Ct. at 2660-61. Nonetheless, when substantial proportionality exists, we are reluctant to interfere with the political judgments made by the legislature, absent some special circumstances not shown here. Rural West I, 836 F.Supp. at 467. See also NAACP, 857 F.Supp. at 571 (stating that "showing of persistent, proportional representation is sufficient to defeat the plaintiffs' Section 2 claims, absent proof of a racially discriminatory motivation in the drawing of the district lines.").

But another means of analyzing the region, if accepted, could lead us to conclude that substantial proportionality does not exist in those seven counties. The plaintiffs note that the seven-county region includes all or some portion of eight senate districts. The two majority-minority districts in that region comprise only 25% of the eight districts. If a third district were added, 37.5% of the districts would be majority-minority districts, slightly below the 37.9% figure necessary for exact proportionality. Using this analysis we would conclude that substantial proportionality does not exist.

When we include the influence districts in our analysis of the "totality of the circumstances" we find that both methods of regional analysis yield findings that § 2 has not been violated. Four of the eight districts which fall at least partially within the seven-county region are either majority-minority districts or influence districts. Six of the eight districts have black voting-age populations over 20%. We conclude, as we do in our statewide analysis, that creating an additional majority-minority district at the expense of two or more districts with significant black population would not necessarily enhance black influence in the state senate and reduce vote-dilution. Instead, amending the 1992 Plan to add another black district might have the opposite effect. In sum, a consideration of the seven-county region alone does not change the outcome of our decision today.

VI. Conclusion

In conclusion, we find that following the clarifications to § 2 presented in De Grandy, we erred in our first opinion in this case. We now recognize that § 2 does not necessarily require proportionality even when the Gingles preconditions have been met and when the plaintiffs have demonstrated the presence of racial polarization and a history of racial inequality. Nonetheless, were we still prevented from recognizing the existence of influence districts, we would reinstate our previous holding. But, an inquiry into "the totality of the circumstances" should be broad and exhaustive, and, therefore, does allow for consideration of influence districts. When we now include the three influence districts in the 1992 Plan in our evaluation of the "totality of the circumstances," we find that the 1992 Plan does not violate § 2 of the Voting Rights Act.

For the foregoing reasons, it is ordered.

APPENDIX A[1]
1992 Tennessee Senate General Election Results
========================================================================================
District                                     Candidates                    % Votes
========================================================================================
2[*][†][‡]          Jim Holcomb (R)                 55.1%
Wallace Ketron, Jr. (D)         44.9%
----------------------------------------------------------------------------------------
4[*][†][‡]          Frank S. Niceley (R)            50.9%
Danny Wallace (D)               49.1%
----------------------------------------------------------------------------------------
6[*][†]                       Ben Atchley (R)                 67.9%
Steve McGill (D)                32.1%
----------------------------------------------------------------------------------------
8[*][†][‡]          Carl O. Koella (R)              53.4%
Vernon Scarbrough (D)           35.8%
Hubert Patty (I)                10.8%
----------------------------------------------------------------------------------------
10                                         Ward Crutchfield (D)           100.0%
----------------------------------------------------------------------------------------
12[*][†][‡]         Anna Belle O'Brien (D)          51.2%
W. Van Hilleary (R)             48.8%
----------------------------------------------------------------------------------------
14                                         Jerry W. Cooper (D)             100.0%
----------------------------------------------------------------------------------------
16                                         Andy Womack (D)                 100.0%
----------------------------------------------------------------------------------------
18[*][†][‡]         Don Wright (R)                   50.3%
Charles Hill Beatty (D)          49.7%
----------------------------------------------------------------------------------------
20[*][†][‡]         Joe Haynes (D)                   59.4%
Ben Jones (R)                    40.6%
----------------------------------------------------------------------------------------
22[*][†][‡]         Carol Rice (R)                   50.2%
Riley Darnell (D)                49.8%
----------------------------------------------------------------------------------------
24                                         Milton Hamilton (D)             100.0%
----------------------------------------------------------------------------------------
26                                         John S. Wilder, Sr. (D)         100.0%
----------------------------------------------------------------------------------------
28[*][†]                      Jim Kyle (D)                     67.8%
James H. Brown, Jr. (R)          22.7%
Jesse E. Neely (1)                9.5%
----------------------------------------------------------------------------------------

========================================================================================
District                                     Candidates                    % Votes
========================================================================================
30[*][†]                      Steve Cohen (D)                 62.8%
Bill Davis (R)                  37.2%
----------------------------------------------------------------------------------------
32                                         Tom Leatherwood (R)            100.0%
========================================================================================
1994 Tennessee Senate General Election Results
========================================================================================
District                                     Candidates                    % Votes
========================================================================================
1[*][†]                       Tommy G. Haun (R)               64.9%
William O. (Bill) Shults (D)    35.1%
----------------------------------------------------------------------------------------
3[*][†]                       Rusty Crowe (D)                 66.5%
Helen Stafford Wilson (R)       33.5%
----------------------------------------------------------------------------------------
5[*][†][‡]          Randy McNally (R)               52.7%
William (Bear) Stephenson (D)   47.3%
----------------------------------------------------------------------------------------
7                                          Bud Gilbert (R)                100.0%
----------------------------------------------------------------------------------------
9[*][†]                       Jeff Miller (R)                 67.2%
Larry Guy (D)                   32.8%
----------------------------------------------------------------------------------------
11                                         David Fowler (R)               100.0%
----------------------------------------------------------------------------------------
13[*][†][‡]         Gene Elsea (R)                  60.3%
Jane W. Dawkins (D)             39.7%
----------------------------------------------------------------------------------------
15[*][†][‡]         Tommy Burks (D)                 59.0%
James H. Threet (R)             41.0%
----------------------------------------------------------------------------------------
17[*][†][‡]         Robert Rochelle (D)             54.0%
Jack Bowers (R)                 45.4%
----------------------------------------------------------------------------------------
19                                         Thelma Harper (D)               89.6%
Kwame Leo Lillard (I)           10.4%
----------------------------------------------------------------------------------------
21[*][†][‡]         Douglas Henry, Jr. (D)          56.4%
Sharon Bell (R)                 43.6%
----------------------------------------------------------------------------------------
23[*][†][‡]         Keith Jordan (R)                58.8%
Gary C. Ledbetter (D)           41.2%
----------------------------------------------------------------------------------------
25[*][†][‡]         Kenneth N. (Pete) Springer (D)  57.3%
----------------------------------------------------------------------------------------
27[*][†][‡]         Bobby Carter (R)                51.4%
Joe Nip McKnight                48.6%
----------------------------------------------------------------------------------------
29                                         John N. Ford (D)                84.9%
Fred Hooks (I)                  15.1%
----------------------------------------------------------------------------------------
31                                         Curtis Person (R)              100.0%
----------------------------------------------------------------------------------------
33                                         Roscoe Dixon (D)               100.0%
========================================================================================

APPENDIX B

NOTES

[*] This is a "contested" race, in which both a Democrat and a Republican ran.

[†] In this race, had 25% of those who voted and cast votes for the winning candidate switched their vote to the second place candidate, the result of the election would have changed.

[‡] In this race, if 25% of those who voted and cast votes for the winning candidate abstained from voting, the result of the election would have changed.

[4] We do not address whether a district in which minorities comprise less than 25% of the voting-age population can never be considered an influence district, even when proof presented by the parties leads to the opposite conclusion. Although two districts that might be called "weak influence districts" exist under the 1992 Plan, we have no need to include them in our calculation of the totality of the circumstances.

[5] In Rural West I we noted that courts have not agreed on what the minimum percentage of minorities necessary for majority-minority district should be. No bright-line rule exists common to all courts. Yet every court has established a standard rule to avoid engaging in a district-by-district analysis of whether a minority group constitutes an "effective majority." Rural West I, 836 F.Supp. at 467.

[6] For the purposes of this analysis, contested races include only races with a Democratic and Republican candidate in the general election. Races in which either a Democrat or a Republican faced opposition only from Independent or write-in candidates we did not consider "contested" because no Independent or write-in candidate received a significant portion of the vote.

[7] For the reasons stated in Rural West I, 836 F.Supp. at 467, we continue to use 55% as the minimum voting-age population necessary for a majority-minority district. District 19 has a 58% black voting-age population; District 29 has an 83% black voting-age population; and District 33 has a 69% black voting-age population.

[8] District 10 has a 32% black voting-age population; District 28 has a 31% black voting-age population; and District 30 has a 33% black voting-age population.

[9] Districts 26 and 27 each have a 21% black voting-age population.

[10] If we were to consider all five districts in which blacks comprise over 20% but less than 55% of the voting-age population, our conclusion would be even stronger. Although we do not think the two districts with black voting-age populations between 20% and 25% are irrelevant to a consideration of the totality of the circumstances, we do not need to include them to reach the conclusion that we do.

[11] We do not imply that the legislature could adopt a plan under current circumstances with no majority-minority districts at all if it created a sufficient number of influence districts. Under current doctrine, the facts in this case require some majority-minority districts.

[12] The De Grandy Court also employed the phrase "substantial proportionality." De Grandy, ___ U.S. at ___, 114 S.Ct. at 2658. Substantial proportionality exists when a legislature has created the maximum number of majority-minority districts possible without exceeding proportionality. For example, in the current case, the reapportionment plan would have to include four majority-minority districts (12.1% of the total districts) for substantial proportionality to exist because blacks are 14.4% of the total voting-age population of Tennessee. Four is the greatest number of districts that can be created without exceeding the 14.4% figure.

[13] We also note that in this case there are three majority-minority districts and three black senators (all elected from the majority-minority districts). Therefore, the analyses of proportionality and proportional representation under the 1992 Plan yield the same results.

[14] Districts 28, 29, 30, 31, 32, and 33 are entirely within the seven-county region. Districts 26 and 27 fall partially within the region.

[15] The two majority-minority seats would comprise 33% of the six-seat total. Adding a third majority-minority district would raise this percentage to 50%, well above the 37.9% figure necessary for proportionality.

[16] The two majority-minority districts would comprise 28.6% of the seven seats. While this is less than 37.9% (the percentage of the voting-age population that is black) adding a third seat would raise this figure to 42.9%, significantly above proportionality.

[1] These tables rely on the official election results published by the office of the Tennessee Secretary of State. The information here does not include the number of votes received by write-in candidates, who in no case received more than .0006% of the votes combined in any one race.

[*] This is a "contested" race, in which both a Democrat and a Republican ran.

[†] In this race, had 25% of those who voted and cast votes for the winning candidate switched their vote to the second place candidate, the result of the election would have changed.

[‡] In this race, if 25% of those who voted and cast votes for the winning candidate abstained from voting, the result of the election would have changed.