Mathis Kearse Wright, Jr. v. Sumter County Board of Elections and Registration

Opinion

[PUBLISH]

IN THE UNITED STATES COURT OF APPEALS

FOR THE ELEVENTH CIRCUIT

________________________

Nos. 18-11510; 18-13510; 20-10394

________________________

D.C. Docket No. 1:14-cv-00042-WLS

MATHIS KEARSE WRIGHT, JR., Plaintiff - Appellee,

versus SUMTER COUNTY BOARD OF ELECTIONS AND REGISTRATION, Defendant - Appellant.

________________________

Appeals from the United States District Court for the Middle District of Georgia

________________________

(October 27, 2020)

Before GRANT, MARCUS and JULIE CARNES, Circuit Judges. MARCUS, Circuit Judge:

On February 25, 2014, Georgia’s House Bill 836 changed the school board district map in Sumter County. The bill reduced the size of the board from nine

2

members to seven. And where previously, all nine members had come from single-member districts, now only five would, and two would be drawn from at- large seats. The plaintiff, a local reverend, Mathis Kearse Wright Jr., challenged this new map in the United States District Court for the Middle District of Georgia. He claimed that the electoral mechanism created by the new map would violate section 2 of the Voting Rights Act of 1965 by diluting the strength of black voters in Sumter County. The district court, following a four-day bench trial, agreed. After extensive litigation that went back and forth between the district court and this Court, the trial court entered a remedial order removing the at-large seats and drawing a new map with seven, single-member districts instead. School board elections under that map are set to go forward in November.

Over the course of these proceedings, the Sumter County Board of Elections and Registration (the “County Board”) has lodged three appeals with our Court -- from the district court’s March 17, 2018 order finding a section 2 violation; from its August 17, 2018 order enjoining the November 2018 school board elections; and from its January 29, 2020 order, on remand from this Court, drawing a new district map for Sumter County’s school board. On the extensive record before us, we can discern no clear error in the district court’s fact-finding or in its final judgment. We affirm.

3

I.

Sumter County is located in rural, southwest Georgia.1 It has a total population of 31,070. Of those, 13,095 (42.1%) are white and 16,159 (52.0%) are black. Most of the County -- 23,541 of its residents -- are of voting age. A plurality are black; 11,652 (49.5%) of the County’s voting-age residents are black, and 10,991 (46.7%) are white. The County has 15,683 total active registered voters. Here, too, blacks enjoy a slight numerical advantage; 7,604 (48.5%) of the County’s registered voters are black and 7,327 (46.7%) are white, a disparity of 277. From 1994 to 2014, the County’s school board included nine members. Voters elected each member from one of nine, single-member districts. That changed on February 25, 2014, when Georgia’s governor signed into law House Bill 836 (“H.B. 836”). 2 H.B. 836 reduced the size of Sumter County’s school board from nine members to seven. See H.B. 836 § 2(a), Gen. Assembly, Reg. Sess. (Ga. 2014). Five of those members would be elected from single-member districts. Id. The remaining two board members would “be elected from Sumter

1 We refer to Sumter County as “Sumter County” or the “County” throughout this opinion.

2 H.B. 836 is not codified in the Georgia Code. The official version is available on the Georgia General Assembly website at: http://www.legis.ga.gov/Legislation/20132014/140327.pdf (last visited Oct. 16, 2020).

4

County at large.” 3 Id. The County Board would oversee the first school board elections under the new map in May 2014. Id. § 4(c), 2014 Ga. Laws 3504–05.

But in March Reverend Wright sued the Sumter County Board of Elections and Registration, seeking to enjoin those upcoming elections. Wright, a registered voter, claimed that H.B. 836’s district map would “result in vote dilution” in violation of section 2 of the Voting Rights Act of 1965. The gist of his complaint was that “the method of electing two members of the Board of Education at-large and the unnecessary packing of blacks” into two of the five single-member districts demonstrably diluted the County’s “black voting strength.” Section 2 says that no

voting qualification or prerequisite to voting or standard, practice, or procedure shall be imposed or applied by any State or political subdivision in a manner which results in a denial or abridgement of the right of any citizen of the United States to vote on account of race or color.

52 U.S.C. § 10301(a). To establish a section 2 violation, a plaintiff must show, “based on the totality of circumstances,” that

the political processes leading to nomination or election in the State or political subdivision are not equally open to participation by members of a class of citizens protected by subsection (a) in that its members have less opportunity than other members of the electorate to participate in the political process and to elect representatives of their choice.

3 The legislature drew the five single-member districts to match the existing districts for the Sumter County Board of Commissioners. See H.B. 836 § 4(a), Gen. Assembly, Reg. Sess. (Ga. 2014).

5

Id. § 10301(b).

The statute emphasizes that no part of section 2 “establishes a right to have members of a protected class elected in numbers equal to their proportion in the population.” Id. Rather, in the words of the Supreme Court, the district court is required to determine, after reviewing the “totality of the circumstances” and, “based upon a searching practical evaluation of the past and present reality, whether the political process is equally open to minority voters. This determination is peculiarly dependent upon the facts of each case and requires an intensely local appraisal of the design and impact of the contested electoral mechanisms.” Thornburg v. Gingles, 478 U.S. 30, 79 (1986) (citations and quotations omitted). The statute allows a court to consider the “extent to which members of a protected class have been elected to office in the State or political subdivision” when evaluating a section 2 claim. 52 U.S.C. § 10301(b).

A plaintiff alleging vote dilution must satisfy “the three now-familiar Gingles factors: (1) that the minority group is ‘sufficiently large and geographically compact to constitute a majority in a single-member district;’ (2) that the minority group is ‘politically cohesive;’ and (3) that sufficient racial bloc voting exists such that the white majority usually defeats the minority’s preferred candidate.” Solomon v. Liberty Cty. Comm’rs, 221 F.3d 1218, 1225 (11th Cir. 2000) (en banc) (quoting Gingles, 478 U.S. at 50–51). Doing so, however, does

6

not end our inquiry. Instead, once “all three Gingles requirements are established, the statutory text directs us to consider the ‘totality of circumstances’ to determine whether members of a racial group have less opportunity than do other members of the electorate.” League of United Latin Am. Citizens v. Perry (LULAC), 548 U.S. 399, 425–26 (2006). As the Supreme Court has observed, the “general terms of the statutory standard ‘totality of circumstances’ require judicial interpretation.” Id. at 426. To that end, the Court has “provided some structure to the statute’s ‘totality of circumstances’ test” by adopting the Senate Report on the 1982 amendments to the Voting Rights Act (the “Senate Report”), which identifies the factors typically relevant to a section 2 claim. Johnson v. De Grandy, 512 U.S. 997, 1010 (1994).

These factors -- known as the “Senate Factors” -- include:

1. the extent of any history of official discrimination in the state or political subdivision that touched the right of the members of the minority group to register, to vote, or otherwise to participate in the democratic process;

2. the extent to which voting in the elections of the state or political subdivision is racially polarized;

3. the extent to which the state or political subdivision has used unusually large election districts, majority vote requirements, antisingle shot provisions, or other voting practices or procedures that may enhance the opportunity for discrimination against the minority group;

4. if there is a candidate slating process, whether the members of the minority group have been denied access to that process;

7

5. the extent to which members of the minority group in the state or political subdivision bear the effects of discrimination in such areas as education, employment and health, which hinder their ability to participate effectively in the political process;

6. whether political campaigns have been characterized by overt or subtle racial appeals; and

7. the extent to which members of the minority group have been elected to public office in the jurisdiction.

Solomon, 221 F.3d at 1225–26 (quotation omitted). Two additional factors may be probative in some cases:

8. whether there is a significant lack of responsiveness on the part of elected officials to the particularized needs of the members of the minority group; and

9. whether the policy underlying the state or political subdivision’s use 1

of such voting qualification, prerequisite to voting, or standard, practice or procedure is tenuous.

Id. at 1226 (quotation omitted). The Court has also explained that “proportionality is ‘a relevant fact in the totality of circumstances’” analysis. LULAC, 548 U.S. at 436 (quoting De Grandy, 512 U.S. at 1000). This factor asks whether “minority voters form effective voting majorities in a number of districts roughly proportional to the minority voters’ respective shares in the voting-age population.” De Grandy, 512 U.S. at 1000.

Nor is the district court limited to considering only these factors; the list is “neither comprehensive nor exclusive.” Ga. State Conference of the NAACP v. Fayette Cty. Bd. of Comm’rs, 775 F.3d 1336, 1342 (11th Cir. 2015) (quoting

8

Gingles, 478 U.S. at 45). And there is no “requirement that ‘any particular number of factors be proved, or that a majority of them point one way or the other.’” Id. (quoting Gingles, 478 U.S. at 45). Moreover, while at-large districts are not “per se violative of § 2,” Gingles, 478 U.S. at 46, they have been characterized by the Supreme Court as posing “greater threats to minority-voter participation in the political process than do single-member districts,” Growe v. Emison, 507 U.S. 25, 40 (1993); see also De Grandy, 512 U.S. at 1012 (explaining that section 2 challenges to single-member districts are “more difficult to grasp”). For that reason, the Court has “strongly preferred single-member districts for federal-court- ordered reapportionment.” Growe, 507 U.S. at 40.

At the start of this lengthy litigation, the district court denied Wright’s motion to enjoin Sumter County’s May 2014 school board elections and, after discovery, granted the County Board’s motion for summary judgment. It found that Wright was unable to satisfy the third Gingles factor because he had “not established that non-black-preferred candidates” in Sumter County “normally defeat black-preferred candidates.” Wright appealed, and a panel of this Court reversed and remanded. See Wright v. Sumter Cty. Bd. of Elections & Registration, 657 F. App’x 871, 873 (11th Cir. 2016) (per curiam). The panel held that the district court had “erred by improperly weighing the evidence and making credibility determinations at the summary-judgment stage.” Id. at 872. The

9

district court, we said, had unfairly discounted the opinions of Wright’s expert political scientist and statistician. Id. It had also given undue weight to the results of older, exogenous elections at the expense of more recent, endogenous ones. Id. at 873. Writing separately, Judge Tjoflat suggested that, on remand, the district court focus on (1) how each district in the current plan voted in the most recent at- large elections; (2) how each district would be likely to perform under a plan with seven, single-member districts; and (3) how African-American candidates had previously performed in at-large elections. Id. at 874 (Tjoflat, J., concurring).

Back again in district court, and after further discovery, Wright moved for summary judgment. The district court granted the motion in part and denied it in part. The court found Wright had satisfied the first Gingles factor. The plaintiff had shown that “an additional district could be created in which African Americans would constitute a majority of voters.” But the court denied summary judgment in all other respects. It found that Wright had not yet established the second or third Gingles factors, nor had he satisfied the first, second, third, fifth, or seventh Senate Factors. The trial court also found that the following facts were undisputed: “(1) Sumter County uses a majority-vote requirement”; (2) no “African American has been elected to an at-large seat on the school board under the challenged plan”; and (3) no “African Americans have been elected in school board districts except where African Americans constitute a majority of the voting-age population.”

10

The case proceeded to a bench trial over the course of four days in December 2017. Wright relied principally on the testimony of his expert witness, Dr. Frederick McBride. Dr. McBride is a political scientist who specializes in race and politics. McBride, who was qualified as an expert in redistricting demographic analysis and the analysis of racial voting patterns, conducted what’s called a racial bloc voting analysis. He reviewed the election returns from twelve Sumter County elections and used a variety of generally accepted statistical techniques -- homogenous precinct analysis, ecological regression, and ecological inference -- to determine whether black voters in the County display political cohesion and whether white voters vote as a bloc to defeat the electoral preferences of their African-American neighbors.4 Three of the twelve elections McBride studied were endogenous -- the three at-large elections for the school board held under H.B. 836 -- and a fourth was a runoff race that followed one of those three. The remaining eight were exogenous elections: the May 2014 school board elections, under H.B. 836, for Districts One through Five; the March 2014 school board election for District Six; the November 2010 school board election for District Three; and the November 2004 election for Sumter County sheriff.

4 As do the parties, we use the terms “black” and “African American” interchangeably. See Sanchez v. Roden, 753 F.3d 279, 286 n.3 (1st Cir. 2014). We do the same with the terms “Hispanic” and “Latino.” See id.

11

McBride also used mapping software to determine whether there were enough black residents in Sumter County to constitute a majority in an additional, sufficiently compact district. As part of his report, Dr. McBride provided the district court with an illustrative map that “would,” in his opinion, “increase African American membership on the board.” Under H.B. 836, blacks comprised a majority of the voting-age population in two of the districts; Dr. McBride’s illustrative plan placed black majorities in three districts.

Dr. McBride opined that voting preferences in Sumter County were “highly polarized” along racial lines. He also found that white voters in Sumter County voted as a bloc “usually to defeat the black preferred candidate.” McBride concluded that H.B. 836’s district map did “not afford African American voters a realistic opportunity to elect candidates of choice.” For its part, the County Board qualified Dr. Karen Owen as an expert in political science and statistics. Dr. Owen disputed the reliability of Dr. McBride’s findings but conducted no analysis of her own. Nor did she provide any testimony analyzing the sufficiency of Dr. McBride’s illustrative map.

In March 2018, the district court agreed with Wright and concluded that H.B. 836’s districting scheme worked an impermissible vote dilution in violation of section 2. The district court set out its findings in a detailed opinion; we explicate those findings at some length, in order to more properly evaluate whether

12

the trial court clearly erred. As for the first Gingles factor -- numerosity and compactness -- the district court reminded that it had already granted summary judgment for Wright on that factor and had “no need to revisit it now.” The court then found that Wright had shown political cohesion among black voters in Sumter County, satisfying the second Gingles factor. It relied on Dr. McBride’s racial bloc voting analysis of the twelve elections. Here are the results in those elections:5

May 20, 2014 Elections

District One

Candidate Overall Support White Support Black Support Alice Green* 52.8% 15.0% 94.2% E. Lockhart 11.6% 21.2% 1.2% Allen Smith 35.4% 66.8% 1.1% District Two

Candidate Overall Support White Support Black Support Everette Byrd 28.9% 30.1% 23.3% Meda Krenson 48.7% 59.0% 0.0% Sarah Pride* 22.2% 5.8% 99.3% District Three

Candidate Overall Support White Support Black Support W. Fitzpatrick* 30.5% 4.6% 92.3% J.C. Reid 69.4% 95.0% 8.5% District Four

Candidate Overall Support White Support Black Support Rick Barnes 54.4% 54.7% 53.9% Gary Houston 45.5% 44.9% 46.6%

5 The eagle-eyed reader will notice that some of the percentages in the white and black support columns do not add up to 100%. That is not a typographical error. At trial, Dr. McBride explained that the principal statistical method he used to calculate these figures, ecological inference, “establishes bounds for individual candidates” but does not constrain the sum of those numbers to 100. Though this was a disputed point at trial -- Dr. Owen, for instance, said the court should discount Dr. McBride’s findings due to the aberrant sums -- the district court ultimately credited Dr. McBride’s explanation.

13

District Five

Candidate Overall Support White Support Black Support Edith Green* 55.4% 13.9% 85.3% Mark Griggs 44.5% 86.4% 14.3% At-Large Seat #1

Candidate Overall Support White Support Black Support Michael Coley* 36.7% 4.1% 89.1% David Kitchens 20.4% 32.7% 0.0% Sylvia Roland 36.4% 53.0% 9.7% Patricia Taft* 6.3% 6.5% 6.2% At-Large Seat #2

Candidate Overall Support White Support Black Support Michael Busman 59.8% 94.4% 3.0% Kelvin Pless* 40.1% 5.8% 96.7% July 22, 2014 Elections

At-Large Seat #1 Runoff

Candidate Overall Support White Support Black Support Michael Coley* 41.0% 7.5% 99.5% Sylvia Roland 58.9% 92.4% 0.0% May 24, 2016 Elections

At-Large Seat #1

Candidate Overall Support White Support Black Support Michael Coley* 44.5% 15.3% 93.6% Sylvia Roland 55.4% 84.7% 6.2% Other Elections

November 2, 2004 Sheriff Election Candidate Overall Support White Support Black Support Pete Smith 40.3% 54.6% 0.0% James Driver 32.3% 39.9% 6.3% Nelson Brown* (Write-In) 27.3% 4.4% 96.5% November 2, 2010 School Board District Three Election Candidate Overall Support White Support Black Support Donna Minich 44.4% 76.7% 6.3% Kelvin Pless* 55.3% 22.9% 94.0% March 18, 2014 School Board District Six Election Candidate Overall Support White Support Black Support Michael Mock 71.0% 85.1% 52.1% Sarah Pride* 28.9% 28.8% 68.0% * Asterisks indicate black candidates.

14

The district court found that, in ten of these twelve races, “the overwhelming majority of African Americans voted for the same candidate.” One of the remaining two -- the May 2014 District Four election -- pitted two white candidates against each other. The district court, relying on this Court’s decision in Davis v. Chiles, found that the results of that election were less probative. 139 F.3d 1414, 1418 n.5 (11th Cir. 1998) (explaining that “evidence drawn from elections involving black candidates is more probative in Section Two cases”). The trial court also paid particular attention to the results of the November 2004 race for Sumter County sheriff. There, a black candidate, Nelson Brown, won the overwhelming support of the County’s black voters as a write-in candidate. The court observed that write-in candidates “face obvious structural barriers that make their election in the American political system rare.” The court thought it “extraordinary” that black voters in Sumter County “demonstrate[d] that level of cohesion for a write-in campaign.” A white candidate won that election.

Next, the district court found that Wright had satisfied the third Gingles factor. Wright had shown that white voters in Sumter County “are usually able to defeat the candidate preferred by African Americans.” The district court grouped the twelve elections into three categories. The first one was comprised of seven races that pitted one white candidate against one black candidate. Black voters had a clear preference in six of those seven elections, but the black-preferred candidate

15

won only twice: in the May 2014 District Five race and in the November 2010 District Three race. The court discounted, for purposes of the third Gingles factor, the seventh election in this category -- the March 2014 District Six election. Black voters did not coalesce around a single candidate in that race, and the court reasoned that it could not “meaningfully consider whether white voters” had defeated a black-preferred candidate in an election that was without one. The district court also discounted the success of the black-preferred candidate in the May 2014 District Five election, since over 70% of the voting-age population in that district was black. That left only one “true success” among the first bucket of elections: the November 2010 District Three school board race.

The second category included elections with multiple candidates facing off against one black-preferred candidate. The court placed four elections in this bucket: the May 2014 school board elections for District One, District Two, and the first at-large seat; and the 2004 election for Sumter County sheriff. The court discounted, for purposes of the third Gingles factor, the results of the sheriff’s race. Though the black-preferred candidate in that race, Nelson Brown, earned “incredible” levels of support from black voters and only scant support from white ones, his write-in status left the court able only to “speculate as to whether white voters were” even “aware that Brown was running.” The district court also discounted the results of the May 2014 District One election. Though white and

16

black voters supported different candidates in that election, 62.7% of the district’s voting-age population was black, so the black-preferred candidate was able to prevail. White voters were able to defeat the black-preferred candidates in the remaining two elections in this category. In the first one, the May 2014 District Two election, white voters overwhelmingly voted against the black-preferred candidate, and the two white candidates proceeded to a runoff. In the at-large race, the black-preferred candidate was able to win a plurality of the votes cast. But in Sumter County, a candidate needs to earn a majority of the votes to win the election. The black-preferred candidate in that race, Michael Coley, thus proceeded to a runoff, where, as the district court found, “white voters then coalesced around a single candidate and defeated him.”

Finally, the district court considered as its own category the twelfth election -- the May 2014 District Four race. The court discounted that one because it included no clear preference among black voters, and indeed, no clear preference among white voters either. Stepping back, the district court found that there could be “no doubt black and white voters consistently prefer different candidates” and that, “[m]oreover, white voters are usually able to . . . defeat the candidate preferred by African Americans.” Thus, the district court determined that Wright had satisfied the third Gingles factor.

17

Having found each Gingles factor satisfied, the district court moved on --

just as it was obliged -- to consider whether, based on the totality of the circumstances, H.B. 836’s district map diluted Sumter County’s black vote in violation of section 2. The court’s analysis here was guided by the Senate Factors. The court found that the first, second, fifth, and seventh Senate Factors all weighed heavily in Wright’s favor. The first Senate Factor focuses on “the history of voting-related discrimination in the State or political subdivision.” LULAC, 548 U.S. at 426 (quotation omitted). Wright and the County Board jointly stipulated at trial that “Georgia and Sumter County have a long and extensive history of voting discrimination against African Americans.” Thus, the court concluded, this factor weighed heavily in Wright’s favor. Next, it evaluated the second Senate Factor, which focuses on the extent of racially polarized voting in the state or political subdivision. Id. (quotation omitted). The court concluded, largely based on Dr. McBride’s analysis, that Sumter County’s elections are “highly” racially polarized, and that this factor too weighed heavily in Wright’s favor.

The fifth Senate Factor directs the court to consider “the extent to which minority group members bear the effects of past discrimination in areas such as education, employment, and health, which hinder their ability to participate effectively in the political process.” Id. (quotation omitted). We have recognized in binding precedent that “disproportionate educational, employment, income

18

level, and living conditions arising from past discrimination tend to depress minority political participation.” United States v. Marengo Cty. Comm’n, 731 F.2d 1546, 1568 (11th Cir. 1984) (alterations adopted and quotation omitted). “Where these conditions are shown, and where the level of black participation is depressed, plaintiffs need not prove any further causal nexus between their disparate socio-economic status and the depressed level of political participation.” Id. at 1568–69 (quotation omitted); see also United States v. Dallas Cty. Comm’n, 739 F.2d 1529, 1537 (11th Cir. 1984) (“Once lower socio-economic status of blacks has been shown, there is no need to show the causal link of this lower status on political participation.”).

The district court found the socioeconomic disparities between the black and white residents of Sumter County “striking.” The court found, first, that only 13.6% of the County’s white residents had not graduated from high school. The figure was more than double for the area’s black residents -- 29.9%. The trend continued in higher education as well. The facts revealed that 30.9% of the County’s white residents have earned a bachelor’s degree or higher. Indeed, whites are more than three times as likely to have earned those degrees than the County’s black residents; only 8.8% of Sumter County’s black residents have earned a bachelor’s degree or higher. These educational trends spilled over into the workforce. The unemployment rate was 7.1% for white residents but 18.2%

19

for black ones. The same strong disparity was true for the County’s household economics. The court found that 15.3% of white residents live in poverty; that figure, however, is “an astonishing 46.2%” for African Americans. Three in four black households receive Supplemental Nutrition Assistance Program benefits; only one in five white households do. Moreover, the district court found a vast gulf between the median incomes of black and white households in Sumter County: it’s $48,672 for white households, and less than half of that -- $22,736 -- for black ones.

The court also found that these “disparities” had resulted “in decreased political participation” in the County’s black community. Although African Americans in fact outnumber Sumter County’s white residents in total population, voting-age population, and barely on the voter rolls, the court expressly found that “white voters have outnumbered black voters in school-general elections by an almost two-to-one margin since 1996.” Since Wright had established powerful disparities in the socioeconomic status of white and black residents “and a depressed level of political participation by African Americans,” the trial court concluded that the fifth Senate Factor weighed heavily in Wright’s favor too.

Finally, the district court determined that the seventh Senate Factor weighed heavily in Wright’s favor. This factor focuses on “the extent to which members of the minority group have been elected to public office in the jurisdiction.” LULAC,

20

548 U.S. at 426 (quotation omitted). “[N]o African American has ever been elected to an at-large seat on the School Board under the challenged plan,” the court found. 6 And no African American had been elected to the school board in any district under the plan except those in which “African Americans make up a majority of the voting-age population.” The court acknowledged that one black candidate, Cortisa Barthell, had won a countywide election to become clerk of the superior court. But it discounted that result. Barthell had run unopposed in the general election; in the contested Democratic primary she won, black voters comprised 81.9% of the electorate. Since there was “no evidence in the record of an African American in Sumter County winning a contested race for county-wide office,” the court found that “African Americans have lacked success in Sumter County elections.”

The district court also determined that the third Senate Factor weighed in Wright’s favor, although not heavily. This factor considers “the extent to which

6 The first time this Court heard Wright’s case, Judge Tjoflat wrote separately, offering three suggestions for the district court to consider on remand. The district court’s finding on this score satisfied the third of those suggestions -- to determine how “African Americans have previously performed in at-large elections.” Wright, 657 F. App’x at 874. But the court was unable to address the first two suggestions, through no fault of its own. Judge Tjoflat asked the district court to consider how each of H.B. 836’s single-member districts “voted in the May 2014 at- large elections,” and how each “would be likely to vote in the proposed seven-district plan.” Id. But, as Dr. McBride explained, election return data in Sumter County are kept at the precinct, not the district, level. And certain precincts are split among multiple districts. That meant McBride could not determine how the residents of any one district voted, or would vote, in a particular election.

21

the State or political subdivision has used voting practices or procedures that tend to enhance the opportunity for discrimination against the minority group, such as unusually large election districts, majority vote requirements, and prohibitions against bullet voting.” Gingles, 478 U.S. at 44–45. The district court concluded that three of Sumter County’s voting practices fit this bill. First, the County “uses staggered terms for the at-large seats, with one at-large seat filled at each regular election.” As the court noted, were “the County to instead seat the top two vote- getters for at-large seats every four years, African Americans would have an enhanced opportunity for election in those seats.” Second, Sumter County “uses a majority-vote requirement in elections for the at-large seats.” And third, the court found that the use of “at-large districts” itself “enhanced the opportunity for discrimination.” The court highlighted testimony from several lay witnesses at trial “with experience in local politics.” Those witnesses “testified that running at large in Sumter County is more expensive than running in a district and therefore presents a particular barrier for African-American candidates.”

That these practices might affect the outcome of an election was not conjecture. As the district court observed, in the absence of at least one, if not two, of these election structures, a black-preferred candidate would have won an at- large seat on the school board. In July 2014, the County Board oversaw a runoff election for the first at-large seat on the school board. Back in May 2014, four

22

candidates had vied for that seat. The top vote-getter, Michael Coley, an African American, was the preferred candidate among Sumter County’s black voters. He earned 36.7% of the vote in May, enough for first place in the contested election, but not enough to avoid a runoff. In July, Coley faced off against the preferred candidate of Sumter County’s white voters, Sylvia Roland. In that runoff election, Roland defeated Coley handily, winning nearly 60% of the votes cast. The district court noted that had the County filled the two at-large seats on the school board with the top two performers in each election, rather than filling those seats by staggering the terms, “the African American candidate—Michael Coley—would be a school board member.” Similarly, Coley would have won under a plurality- win system.

The fourth Senate Factor carried no weight in this case. This factor asks whether minority candidates are excluded from any candidate slating process. Id. at 45. Here, there was “no evidence in the record of any slating process in Sumter County.” That meant the court could not “find whether a slating process would or would not exclude African Americans.” Finally, the court concluded that the sixth Senate Factor weighed in the County Board’s favor. The sixth factor asks whether political campaigns in the area are characterized by subtle or overt racial appeals. Id. (quotation omitted). The court noted that the record lacked any evidence -- indeed, Wright did not allege -- “that any political campaign employed overt or

23

subtle racist appeals.” The court determined, then, that this factor should weigh in the County Board’s favor, but not heavily so. See Marengo Cty. Comm’n, 731 F.2d at 1571 (“Evidence of racism can be very significant if it is present. But its absence should not weigh heavily against a plaintiff proceeding under the results test of section 2.” (citation omitted)).

In evaluating the totality of the circumstances, the court also considered “whether the policy underlying the state or political subdivision’s use of such voting qualification, prerequisite to voting, or standard, practice, or procedure is tenuous.” Gingles, 478 U.S. at 37 (quotation omitted). The court concluded it was not. Sumter County’s nine-member school board “was one of the largest in the state” -- surprising for a relatively small county like Sumter County -- and the board had voted to reduce its size at “the recommendation of its accreditation agency.” Moreover, the new map’s five single-member districts “mirror[ed] those of the County Board of Commissioners.” This factor, the district court determined, weighed in favor of the County Board. That the new map was justified on race- neutral grounds, however, did “not negate” Wright’s “showing through other factors that the challenged practice denies minorities fair access to the [electoral] process.”

Finally, the district court considered Dr. McBride’s illustrative remedial plan and held that it sufficed, but only “by the bare minimum,” and not without good

24

reason. Dr. McBride conceded that his approach involved “absolute guesswork.” He could not “say with certainty” that his illustrative plan would increase black membership on the school board, since there was “no way to predict the future outcome of elections in this manner.” Black voters were unable to elect candidates of their choosing to at-large seats on the school board, even as they comprised 49.5% of Sumter County’s total voting-age population. At the same time, Dr. McBride opined that black voters would be able to elect an additional candidate of their choice using his illustrative plan, relying on a proposed district in which black residents would comprise 54.5% of the voting-age population. While the district court found “no support for the idea that a five percentage point shift would have such a drastic impact on voting behaviors,” it acknowledged that 54.5% was a higher black voting-age population percentage than was necessary in some of the current single-member districts for black residents to elect their preferred candidates and that this increase, combined with an eight-percentage point drop in the number of white residents found in the at-large district, could have a significant impact on the election results.

Although the court viewed the question as a close one, it ultimately sided with Wright, relying on Dr. McBride’s expert opinion that “the illustrative plan provides black-preferred candidates an opportunity for victory in three districts, which is more than they have under HB 836.” This was, after all, the only

25

evidence before the district court on this point; the County Board “did not ask its expert to conduct any analysis of Wright’s illustrative plan.” Thus, “while far from perfect,” the court concluded that Wright’s illustrative plan would likely “give African Americans a more proportional representation on the Board of Education than does the current plan.”

Based on its review of the evidence, the district court concluded that, under the totality of the circumstances, “African Americans in Sumter County have less opportunity to elect candidates of their choice than do white citizens.” The court found five compelling facts that underlaid its conclusion:

(1) the incredibly high rates of polarized voting in races that pit an African American candidate against a white candidate;

(2) the glaring lack of success for African American candidates running for county-wide office, both historically and recently, despite their plurality in voting-age population;

(3) the undisputed history of discrimination in Sumter County and throughout Georgia;

(4) the lingering effects of that discrimination today, including the comparatively low income and education levels and high rates of poverty for African Americans in Sumter County; and

(5) the low rate of African American turnout in these elections which -

- in the absence of evidence to the contrary -- the Court attributes to the history of discrimination and the socioeconomic disparities.

“Because of these factors,” the court concluded, “the elections for at-large seats” to the Sumter County school board “do not give African Americans in Sumter County

26

a meaningful opportunity to elect the candidates of their choice.” The court ended this way:

The court finds that Plaintiff Mathis Kearse Wright, Jr. has established all three Gingles factors, that the majority of the Senate Factors weigh toward him, and that he has shown an illustrative plan which is likely to give African Americans a more proportional representation on the Board of Education than does the current plan. Accordingly, the Court finds, based on the totality of the circumstances, that the at-large districts of the Sumter County Board of Education dilute African-

American voting strength in violation of Section 2 of the Voting Rights Act of 1965, as amended, 52 U.S.C. § 10301.

The district court explained that the case would “move[] to a remedial stage.” It afforded the local elected officials “the first opportunity to remedy the unlawful election plan.” Since Georgia’s General Assembly would remain in session for at least another two weeks, the court ordered the County Board “to confer with Sumter County’s legislative delegation and inform” it “whether the General Assembly is inclined to enact a remedial plan before adjourning sine die or, if not, a timeline for when it believes a remedial plan could be adopted.”

Following the court’s instructions, and after consulting with Sumter County’s local representatives, the County Board provided notice that it would “not be possible” for the General Assembly “to pass legislation on the Sumter County School Board districting regime this session.” In response, Wright moved for a preliminary injunction. Without one, Wright said, Sumter County’s May 2018 school board elections would proceed under a district map the court had

27

found to be unlawful. The district court granted Wright’s motion later that day. Because the court had “already decided the merits of this action in Wright’s favor,” however, it concluded that a permanent injunction was the appropriate remedy and enjoined the May 2018 school board elections, resetting them for November of that year. The trial court also informed the parties that it would “enter an order no later than July 23, 2018 setting interim boundaries for the new Sumter County Board of Education districts.”

The County Board appealed, challenging the district court’s order permanently enjoining the May 2018 school board elections “and all orders forming the basis of or relating to that injunction,” including the district court’s order finding liability under section 2. With that appeal pending -- after the County Board had filed its opening brief, and while we awaited Wright’s response -- the district court sua sponte modified its injunction. Since the County Board’s notice of appeal had deprived it of jurisdiction, the court concluded it was without power to draw a new map. Therefore, school board elections would proceed under H.B. 836’s districting regime. The court denied Wright’s motion for reconsideration.

A panel of this Court then sua sponte remanded the case to the district court on a limited basis, for the second time. We construed its orders -- modifying the injunction and denying Wright’s motion for reconsideration -- “as equivalent to

28

indicative rulings under Fed. R. Civ. P. 62.1(a)(3).” On remand, we directed the district court to determine whether it was “still feasible to issue a new map with interim boundaries for the November election in a timely manner.” “If so,” we said, the court should draw a new district map “before returning the case and the record to this Court.” If not, however, the district court should enter an order saying so “and then return the case to this Court.”

On remand, the district court concluded that it was “not feasible for it to timely issue a new map.” And it enjoined the November 2018 school board elections, to be reset at a later date. With those elections off the calendar, the district court offered the General Assembly still another opportunity to draw a new district map. It ordered the parties to “confer expeditiously and as soon as feasible in good faith with the General Assembly to aid it in drawing new boundaries” for the County’s school board elections. The County Board appealed the district court’s injunction, and we consolidated the second appeal with the County Board’s first.

The action proceeded in this Court, and when the parties completed their briefing, we scheduled oral argument. But in April 2019, the General Assembly adjourned, again declining to take up the district court’s invitation to redraw the map. So Wright moved this Court for another limited remand to allow the district court to conduct remedial proceedings. He claimed the need for immediate

29

remedial proceedings was great because there had been no school board elections in Sumter County “under a lawful plan since 2010.” And without a remand, there would be “a high risk” that the district court “would not have time to devise a remedy before the 2020 election.” Thus, it “would be prudent,” Wright told us, “for this Court to authorize the district court to begin that process sooner rather than later.” We granted the motion and remanded the case to the district court a third time for further remedial action.

On remand, the district court engaged Dr. Bernard Grofman as special master. None of the parties dispute that Dr. Grofman “is one of the world’s leading experts in the study of redistricting and voting rights,” and all agreed that he was well-qualified to serve as a special master. The court instructed Grofman “to submit a report and a proposed plan or plans to remedy” the “unlawful vote dilution” inflicted by H.B. 836. The court explained that any plan should include seven, “reasonably compact,” single-member districts that “[c]omply with the one- person, one-vote principle guaranteed by the Equal Protection Clause of the Fourteenth Amendment.” The remedial plan should “give[] African American voters a realistic opportunity to elect candidates of their choice to at least three seats on the school board.” Grofman, the court said, “should also consider whether it is possible to devise a remedial plan that gives African-American voters a

30

realistic opportunity to elect candidates of their choice to four seats on the school board.”

Dr. Grofman’s special master report produced “five different seven-district illustrative maps.” In each, at least three of the seven districts “contain[ed] a majority African American voting age population.” This result, Grofman said, was “essentially inevitable” in light of the County’s “demography and geography.” Grofman found that a majority (51.9%) of the County’s total population lives in the City of Americus, Georgia. Americus is large enough “to constitute 3.64 districts in a seven-district plan,” and black residents make up 58.4% of the city’s voting-age population. As Grofman explained, “since there is additional Black population located in areas immediately proximate to Americus,” maps “drawn on traditional districting principles” will “naturally generate at least three (and sometimes four) majority Black voting age population districts in Americus and its immediate surroundings.”

Dr. Grofman also found that, in Sumter County, a “competitive” electoral contest would be “one in which Black [voting-age population] is at least 60.2%.” That finding led Grofman to conclude that a Sumter County school board district requires “a roughly 60.2% Black voting age percentage in order to accurately characterize it as a ‘minority opportunity district.’” This conclusion undermined Dr. McBride’s analysis; Dr. McBride had testified that an illustrative district with a

31

black voting-age population of just 54.5% would give black voters a realistic opportunity to elect a third candidate of choice to the school board. Dr. Grofman opined that it was “implausible to expect a (Black) candidate of choice of the African American community to have a realistic chance of election in the non- partisan contests for the School Board in a district that is only barely above 50%” black voting-age population.

Four of the five maps that Dr. Grofman proffered included three “minority opportunity” districts.” One of those maps, Map Three, included a fourth district “which by some criterion would be considered a minority opportunity district.” Were the district court to move the school board elections to November, and were the white incumbent in that district not to seek re-election, Grofman would label that fourth district a minority opportunity to elect district. Map Three was the only one of the five that had “any real potential for electing four candidates of choice of the minority community in a seven-seat plan.”

The district court gave the parties the opportunity to object to Dr. Grofman’s submission; for its part, the County Board “lodge[d] no objection to the special master’s report.” The court also conducted a public hearing to allow interested persons the opportunity to comment. Ultimately, the district court selected Dr. Grofman’s Map Three and entered a remedial order and final judgment in favor of Wright. The court agreed with Dr. Grofman that Map Three was the only map that

32

would “give[] African Americans a realistic opportunity to elect board members to four of seven seats.” Map Three had the added benefit of being the most compact one; moreover, only Map Three included Sumter County public schools in four of the seven districts.

The district court found that a black voting-age population of 60.2% “is generally sufficient to give Black voters a realistic opportunity to elect their candidate of choice.” Although Map Three’s fourth minority opportunity district had a black voting-age population of 60.13% -- “slightly below this threshold” -- the court credited Dr. Grofman’s opinion that this minority opportunity district would be effective if the court moved the County’s school board elections to November, and if the white incumbent declined to run for reelection. Thus, the district court rescheduled Sumter County’s school board elections for November. The court concluded that, “to increase the efficacy of adopting Map 3,” it was “necessary to move the School Board elections to occur regularly in November.” The change would “increase the turnout of African-American voters in Sumter County and give them a meaningful opportunity to elect their candidates of choice.” The district court then entered final judgment in favor of Wright.7

7 Nonpartisan general elections are generally set to coincide with primary elections in the State of Georgia. See O.C.G.A. § 21-2-285.1. Because the County does not specifically contest the district court’s decision to realign the date for the School Board election with the general election instead of the primary election, we do not consider whether that was permissible.

33

The County Board appealed once again; we consolidated the third appeal with the first two for merits disposition.

II.

Our review of a district court’s finding of vote dilution under section 2 is only for clear error. See Gingles, 478 U.S. at 79. We may reverse a district court’s finding as clearly erroneous when, “after viewing all the evidence, we are left with the definite and firm conviction that a mistake has been committed.” Silva v. Pro Transp., Inc., 898 F.3d 1335, 1339 (11th Cir. 2018) (per curiam) (quotation omitted). “This standard” does not allow us “to reverse the finding of the trier of fact simply because” we are “convinced” we “would have decided the case differently.” Solomon, 221 F.3d at 1226 (quoting Anderson v. City of Bessemer City, 470 U.S. 564, 573 (1985)). When we review for clear error, it is not our task “to decide factual issues de novo.” Id. at 1227 (quoting Anderson, 470 U.S. at 573). Instead, we must give “due regard” to “the opportunity of the trial court to judge” the “credibility of witnesses.” Id. at 1226 (quoting FED. R. CIV. P. 52(a)). “Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.” Id. at 1227 (quoting Anderson, 470 U.S. at 574). And unless “we are compelled to conclude that the district court’s findings are not supported by substantial evidence, we must affirm.” Johnson v. Hamrick, 296 F.3d 1065, 1074 (11th Cir. 2002).

34

The Court emphasized in Gingles that “the application of the clearly-

erroneous standard to ultimate findings of vote dilution preserves the benefit of the trial court’s particular familiarity with the indigenous political reality without endangering the rule of law.” 478 U.S. at 79. Thus, we employ this deferential standard in vote dilution cases because “the section 2 inquiry ‘is particularly dependent upon the facts of each case and requires an intensely local appraisal of the design and impact of the contested electoral mechanisms.’” Solomon, 221 F.3d at 1226 (quoting Gingles, 478 U.S. at 79).

The clearly-erroneous standard extends not only to the district court’s ultimate conclusion of vote dilution, but also to its “finding that different pieces of evidence carry different probative values in the overall section 2 investigation.” Id. at 1227. We have recognized that the district court’s “findings made within the framework of the three Gingles preconditions, and the nine nonexhaustive ‘Senate factors,’ will be more or less probative depending upon the facts of the case.” Id. “In examining a jurisdiction’s electoral history,” we have affirmed, “a court may assign more probative value to elections that include minority candidates, than elections with only white candidates; it may consider endogenous elections more important than exogenous elections; or it may conclude that recent elections are more probative than those which occurred long before the litigation began.” Id.

35

(citations omitted). We review all “of the district court’s findings regarding the probative value assigned to each piece of evidence” for clear error. Id.

None of this is to suggest that clear-error review is a rubber stamp on the findings of a district court. It is not. We still must ensure that the district court’s findings are supported by “substantial evidence.” Hamrick, 296 F.3d at 1074. And our review for clear error does nothing to inhibit our “power to correct errors of law.” Solomon, 221 F.3d at 1227 (quoting Gingles, 478 U.S. at 79). But when “the district court’s understanding of the law is correct,” when “the record indicates that the court ‘engaged in a searching and meaningful evaluation of all the relevant evidence,’” and when “there is ‘ample evidence in the record to support the court’s conclusions,’ our review is at an end.” Id. at 1228 (alteration adopted) (quoting S. Christian Leadership Conference v. Sessions, 56 F.3d 1281, 1293 (11th Cir. 1995) (en banc)).

III.

We hold, based on all of the evidence presented and the governing legal standards, that the district court did not clearly err in finding that H.B. 836’s district map violated section 2. Before we say more, however, we address what we view to be an important, threshold question: what record do we review in this case? After the County Board appealed from the district court’s remedial order and its final judgment in favor of Wright, and after we consolidated that appeal with the

36

County Board’s first two, we sought supplemental briefing from the parties on this question. We asked the parties to tell us what relevance, if any, the district court’s remedial proceedings held for the proper resolution of this appeal. Wright argued that we ought to review the entire record, including the remedial proceedings the district court conducted on remand. The County Board disagreed. It claimed that the remedial proceedings were irrelevant to the question whether Wright had carried his burden at trial to establish a section 2 vote dilution case. At the same time, however, it sought to use those proceedings as a sword, claiming that Dr. Grofman’s findings underscored its argument that Dr. McBride had failed to establish an available remedy at the liability phase.

In light of the peculiar posture of this case, we review the entire record, both what McBride presented, and notably, the findings, report, and maps presented by Dr. Grofman in the subsequent remedial proceeding. For starters, a panel of this Court remanded this case to the district court, specifically directing it to conduct remedial proceedings. Only then did the district court open the record and adduce significant additional testimony about remedies from Dr. Grofman. We can see little reason now to constrict our review to only one portion of the record properly before us, particularly when Dr. Grofman’s special master report, and his maps, are squarely in the record and formed the basis for the district court’s remedy. Moreover, the County Board took an appeal from the district court’s remedial

37

order and from its final judgment; we consolidated that appeal for a merits disposition along with the prior two appeals. We could have denied Wright’s motion to remand for remedial proceedings and instead decided this case on the limited record before us at that time. A panel of this Court, however, chose not to do so.

Finally, we have often explained that our “inquiries into remedy and liability cannot be separated.” Burton v. City of Belle Glade, 178 F.3d 1175, 1199 (11th Cir. 1999) (quoting Nipper v. Smith, 39 F.3d 1494, 1530–31 (11th Cir. 1994) (en banc) (alterations adopted)). “We have repeatedly construed the first Gingles factor as requiring a plaintiff to demonstrate the existence of a proper remedy.” Id. A section 2 plaintiff cannot succeed without offering a satisfactory remedial plan. The first Gingles condition “dictates that the issue of remedy is part of the plaintiff’s prima facie case.” Nipper, 39 F.3d at 1530. That means that the “absence of an available remedy is not only relevant at the remedial stage of the litigation, but also precludes, under the totality of the circumstances inquiry, a finding of liability.” Id. at 1533.

Thus, a district court’s remedial proceedings bear directly on and are inextricably bound up in its liability findings. Dillard v. Baldwin County Commissioners is instructive. There, a county commission had conceded section 2 liability; the district court enjoined elections under the then-current plan and

38

entered a remedial order changing the commission’s size from four members to seven. 376 F.3d 1260, 1262–63 (11th Cir. 2004). Several years later, however, the Supreme Court and our Court made it clear that a section 2 remedy may not require a governing body to change its size. See Holder v. Hall, 512 U.S. 874, 881 (1994) (plurality opinion); Nipper, 39 F.3d at 1532. Following those decisions, the district court dissolved its injunction. We affirmed. “In light of the interplay between the inquiries into remedy and liability,” we said, the district court was “precluded from finding an ongoing section 2 violation in this case” when no appropriate remedy was available. Dillard, 376 F.3d at 1266.

The converse is true here. It was not inevitable that the district court, relying on Dr. Grofman’s analysis, would find a workable remedy after this Court’s remand. Had the court instead found that no viable remedy was available, Dillard would have instructed the trial court to dissolve its injunction. Instead, Dr. Grofman’s special master report found that a proper remedy -- indeed, more than one -- was readily available. See Mo. State Conference of the NAACP v. Ferguson-Florissant Sch. Dist., 894 F.3d 924, 935 (8th Cir. 2018) (“To the extent FFSD complains that the proper analysis was lacking in this case, we further note that the district court has since approved a remedial plan -- a plan that FFSD does not challenge.”). As we’ve said, the district court entered its remedial order based on the evidence presented following remand. The County Board would

39

nonetheless have us shut our eyes to the evidence that formed the foundation of that order. This we will not do. Given the unusual procedural posture of this case, it makes no sense for us to resolve the merits of this appeal, covering one of our eyes from the entire record now properly before us.

With that point cleared away, we turn to the merits. Wright adduced ample evidence supporting a finding of vote dilution. The district court’s determination was not clearly erroneous. We recap the most salient pieces of evidence. For one thing, the district court committed no clear error in concluding that Wright satisfied the three Gingles factors. To satisfy the first, a plaintiff must show that the minority group “is sufficiently large and geographically compact to constitute a majority in a single-member district.” Gingles, 478 U.S. at 50. “If it is not,” the Supreme Court has explained, “the multi-member form of the district cannot be responsible for minority voters’ inability to elect its candidates.” Id. (emphasis omitted). The Court has also observed that the first Gingles prong is “needed to establish that the minority has the potential to elect a representative of its own choice in some single-member district.” Growe, 507 U.S. at 40. This factor “asks whether the court can fashion a remedy for a demonstrated abridgement.” Nipper, 39 F.3d at 1511. Unless a plaintiff can satisfy this precondition, “there neither has been a wrong nor can be a remedy.” Growe, 507 U.S. at 41; see also Negron v. City of Miami Beach, 113 F.3d 1563, 1569 (11th Cir. 1997) (explaining “that the

40

foundational inquiry for the first Gingles precondition is whether the minority has the potential to elect a representative of its own choice in some single-member district” (emphasis and quotation omitted)).

The district court did not clearly err in finding the first Gingles condition (requiring numerosity and compactness) satisfied. The undisputed evidence presented showed that Sumter County’s black residents could form a majority in at least one additional, single-member district (and probably in two). H.B. 836’s district map featured two black-majority districts; Dr. McBride’s illustrative map featured three. Indeed, Dr. Grofman’s special master report included five maps, each containing at least three majority-black districts. Grofman opined that this result was “essentially inevitable” in light of the County’s “demography and geography.” And he concluded that a meaningful remedy was available. The map that the district court selected, he said, would give African American voters the opportunity to elect three, if not four, candidates of their choice to the Sumter County school board. The County Board has not disputed this finding.

Nor did the district court clearly err in finding the second Gingles precondition satisfied. The second factor asks whether the protected group is “politically cohesive.” Gingles, 478 U.S. at 51. If it is not, “it cannot be said that the selection of a multimember electoral structure thwarts distinctive minority group interests.” Id. The district court found that black voters in Sumter County

41

were “highly cohesive” in ten of the twelve elections the expert studied, including in all of the at-large elections held under the challenged plan. Indeed, in those ten elections, the district court found that “the overwhelming majority of African Americans voted for the same candidate.” Particularly noteworthy was the result of the November 2004 election for county sheriff. In that race, an African- American write-in candidate, Nelson Brown, “won the support of 96.5% of black voter[s].” “To see African American voters demonstrate that level of cohesion for a write-in campaign,” the district court said, was “extraordinary.”

Finally, the district court did not clearly err in finding that Wright had satisfied the third Gingles factor. The third precondition asks whether white residents vote “sufficiently as a bloc to enable” them “usually to defeat the minority’s preferred candidate.” Gingles, 478 U.S. at 51. This factor requires a plaintiff to prove “not only that whites vote as a bloc, but also that white bloc voting regularly causes the candidate preferred by black voters to lose; in addition, plaintiffs must show not only that blacks and whites sometimes prefer different candidates, but that blacks and whites consistently prefer different candidates.” Hamrick, 296 F.3d at 1074 (emphases in original) (quotation omitted). It is the “usual predictability of the majority’s success” that “distinguishes structural dilution from the mere loss of an occasional election.” Gingles, 478 U.S. at 51. In reaching this conclusion, the district court considered again the twelve elections

42

studied by Dr. McBride and focused on the seven it thought most probative. The district court found that, in all but one of those elections, white residents voted as a bloc to defeat the black-preferred candidate. This was the “legally significant bloc voting” that section 2 addresses. Id. at 57.

Our conclusion that the district court did not clearly err in finding that Wright satisfied the Gingles preconditions gets him much of the way there. After all, “it will be only the very unusual case in which” a plaintiff “can establish the existence of the three Gingles factors but still have failed to establish a violation of § 2 under the totality of circumstances.” Ga. State Conference of the NAACP, 775 F.3d at 1342 (quotation omitted). But just as the Supreme Court has identified those three factors as “generally necessary to prove a § 2 claim,” so too has it “declined to hold them sufficient in combination, either in the sense that a court’s examination of relevant circumstances was complete once the three factors were found to exist, or in the sense that the three in combination necessarily and in all circumstances demonstrated dilution.” De Grandy, 512 U.S. at 1011. Instead, “to establish liability, § 2 plaintiffs must still demonstrate -- after satisfying the three Gingles preconditions -- that the totality of the circumstances results in an unequal opportunity for minority voters to participate in the political process and to elect representatives of their choosing.” Ga. State Conference of NAACP, 775 F.3d at 1342.

43

The district court found that Wright had done just that. The court’s analysis on this score was guided, as it ought to have been, by the Senate Factors. The court considered each of the Senate Factors and found that the first, second, fifth, and seventh weighed heavily in Wright’s favor. The first factor considers “the history of voting-related discrimination in the State or political subdivision.” LULAC, 548 U.S. at 426 (quotation omitted). The district court here rested on the joint stipulation of the parties. At trial, Wright and the County Board agreed “that Georgia and Sumter County have a long and extensive history of voting discrimination against African Americans.”

The second Senate Factor focuses on “the extent to which voting in the elections of the State or political subdivision is racially polarized.” Id. (quotation omitted). This “factor will ordinarily be the keystone of a dilution case.” Marengo Cty. Comm’n, 731 F.2d at 1566. The district court found, as we’ve discussed, that elections in Sumter County were “highly polarized.” “In ten of the twelve elections analyzed,” the court wrote, “over 85% of African American voters voted for the same candidate.” And the average level of white cross-over vote in those races was under 10%. There was no clear error in the district court’s finding that the second factor weighed heavily in Wright’s favor.

Next, the fifth Senate Factor considers “the extent to which minority group members bear the effects of past discrimination in areas such as education,

44

employment, and health, which hinder their ability to participate effectively in the political process.” LULAC, 548 U.S. at 426 (quotation omitted). The district court found “striking” socioeconomic disparities and depressed levels of political participation among blacks in Sumter County. We highlight them again because the disparities are substantial. Only 13.6% of white residents of Sumter County lack a high school diploma; the figure is more than double, 29.9%, for the County’s black residents. White residents are more than three times as likely as their black neighbors to have earned a bachelor’s degree or higher. The unemployment rate is more than twice as high for blacks as it is for whites. More than three times as many black residents live in poverty. The median African American household earns less than half of the median white household. And even though blacks outnumber whites, albeit barely, in total population, in voting-age population, and on the voter rolls, “white voters have outnumbered black voters in school-general elections by an almost two-to-one margin since 1996.” The court also found that this depressed political participation among blacks in Sumter County was the “result” of these “disparities.”

The seventh Senate Factor was also found to weigh heavily in Wright’s favor. This one asks whether “members of the minority group have been elected to public office in the jurisdiction.” Id. (quotation omitted). No black candidate “has ever been elected to an at-large seat on the School Board under the challenged

45

plan.” In fact, the only black members to have been elected to the school board under H.B. 836 represented single-member districts where African Americans make up a majority of the voting-age population.

Wright adduced substantial evidence in support of his section 2 claim. His expert witness, Dr. McBride, performed a racial bloc voting analysis that found high levels of racially polarized voting. He found that white and black voters in Sumter County consistently prefer different candidates, and that white residents vote as a bloc to usually defeat the candidate preferred by black residents. The court took note of Georgia’s, and Sumter County’s, painful history of discrimination against its black residents. It emphasized the high levels of racially polarized voting and observed the lack of success enjoyed by black candidates in Sumter County. And Dr. Grofman’s special master report expressly found an easily achievable remedy available. Based on all of this, the district court found, under the totality of the circumstances, that H.B. 836’s district map denied Sumter County’s black voters the equal opportunity “to participate in the political process and to elect representatives of their choice.” 52 U.S.C. § 10301(b). We cannot say, on this record, that the district court clearly erred.

The County Board broadly advances three arguments challenging the district court’s findings. None of them carries the day. First, the County Board claims that a heightened burden ought to apply in this case because blacks outnumber

46

whites in Sumter County. It relies on a decision from one of our sister circuits, Salas v. Southwest Texas Junior College District, 964 F.2d 1542 (5th Cir. 1992). There, the Fifth Circuit considered a vote dilution claim brought by Hispanic voters challenging an at-large district. Salas, 964 F.2d at 1543. Notably, Hispanic voters constituted a registered voter majority in the challenged district. Id. at 1544. Some 63% of the district’s total population was Hispanic, and 57% of its voting- age population was Hispanic. Id. The court first held that, as a matter of law, even members of a registered voter minority-majority are not precluded from bringing a section 2 vote dilution claim.8 See id. at 1547. It then proceeded to consider how those plaintiffs might succeed. The Fifth Circuit panel concluded that the same Gingles framework, and the totality of the circumstances analysis, still governed. See id. at 1554 (“[W]e stay fixed on, and follow, the controlling totality of circumstances path and do not tarry long, or wander off, in pursuit of trying to fashion some alternative third (white bloc voting) precondition for instances where the protected class is, in fact, the majority.”). The plaintiffs bore the burden of proving “that the use of a multimember electoral structure operates to

8 At least two other courts have reached the same answer. See, e.g., Mo. State Conference of the NAACP, 894 F.3d at 933–34 (explaining that section 2’s use of the term “‘minority’ does not refer to a purely numerical fact”); Pope v. County of Albany, 687 F.3d 565, 575 n.8 (2d Cir. 2012) (explaining that “the law allows plaintiffs to challenge legislatively created bare majorityminority districts”).

47

minimize or cancel out their ability to elect their preferred candidates.” Id. (quoting Gingles, 478 U.S. at 48).

The district court in Salas had found that the plaintiffs failed to meet this burden. It concluded that “the true cause for lack of Hispanic electoral success was not unequal electoral opportunity, but rather the failure of Hispanic voters to take advantage of that opportunity.” Id. at 1555. The Fifth Circuit held that the district court’s conclusion was not clearly erroneous. It said that, when members of a registered voter minority-majority bring a section 2 claim, they must prove “that their inability to elect results from white bloc voting.” Id. In doing so, they face “an obvious, difficult burden.” Id.

Nothing in Salas would yield a different conclusion here. 9 For one thing, and critically, Salas addressed a different factual scenario than the one we face. Hispanic voters there comprised a majority of the challenged district’s voting-age population. Not so here. To the contrary, blacks are not a majority of Sumter County’s voting-age population at all, but a plurality, and a bare one at that. There are 11,652 blacks of voting age in Sumter County, making up 49.5% of the voting- age population. That’s scarcely more -- a numerical advantage of only 661 people, or 2.8 percentage points -- than the white voting-age population, which comes in at

9 We are bound only by the decisions of the former Fifth Circuit handed down prior to the close of business on September 30, 1981. See Bonner v. City of Prichard, 661 F.2d 1206, 1209 (11th Cir. 1981) (en banc). We do, however, look at Salas for its persuasive value.

48

10,991, or 46.7%. Whatever gap there is between the two populations winnows even further when we count only registered voters. Though it appears that black residents form a plurality here too, it’s by an even thinner margin. There are 7,604 (48.5%) blacks on Sumter County’s voter rolls, and 7,327 (46.7%) whites. That means that black voters enjoy the scantest of pluralities, a difference of all of 277 people, just 1.8 percentage points. And even that faint advantage may be an illusory one; the record reflects that there are 456 registered voters in Sumter County of unknown race. If these voters are largely white, then black voters would enjoy no plurality at all. Salas does not purport to address the bare plurality advantage that Sumter County’s blacks have over whites, if there is any advantage at all.

We also stress that, just like in Salas, this Court requires a plaintiff to prove the existence of legally significant white bloc voting, which Wright did here. Indeed, as the district court expressly recognized, it is the plaintiff’s burden to establish each of the Gingles preconditions and to show, under the totality of the circumstances, that members of a protected class suffer unequal access to the political process. The district court did nothing to relieve Wright of his burden. The district court assumed a causal nexus between present socioeconomic disparities resulting from past discrimination, on the one hand, and depressed political participation, on the other. In doing so, the court relied on our

49

unambiguous precedent and on the Senate Report; that was not an error of law. In United States v. Marengo County Commission, we wrote that “when there is clear evidence of present socioeconomic or political disadvantage resulting from past discrimination, as there was in this case, the burden is not on the plaintiffs to prove that this disadvantage is causing reduced political participation, but rather is on those who deny the causal nexus to show that the cause is something else.” 731 F.2d at 1569. The Senate Report said the same thing; plaintiffs need not “prove any further causal nexus between their disparate socio-economic status and the depressed level of political participation” when they show present socioeconomic disparities “arising from past discrimination.” Id. at 1568–69 (quotation omitted). And it is notable here that the County Board did not offer any evidence or advance any argument showing what the low African-American voting rate could otherwise be attributable to or to negate the nexus. We can find no other obvious explanation in this record.

This is not to say that a plaintiff can make out a successful vote dilution claim under section 2 by doing no more than pointing to depressed voter turnout in the present and discrimination in the past. But the district court did not clearly err in concluding, based on the substantial body of evidence presented, that Wright had shown significant socioeconomic disparities arising out of past discrimination that “result in decreased political participation.” The parties stipulated to the long,

50

painful history of discrimination in Georgia in general and in Sumter County in particular. The court found “striking” disparities, calling them the “lingering effects” of past discrimination. The court found depressed political participation among Sumter County’s black residents and that the socioeconomic disparities contributed to blacks’ “decreased political participation.” Indeed, Dr. McBride opined that the connection between the disparities suffered by Sumter County’s black residents and their depressed participation was “well known” and “established.”

The result we reach today might have been different had the evidence presented told a different story. Thus, for example, if the evidence had shown that Sumter County’s black voters regularly turned out in other, exogenous, non-school board elections at levels sufficient to elect the candidates of their choice, then the district court might have found that blacks’ inability to succeed in school board elections was on account of some reason other than the lack of electoral opportunity. But this kind of evidence is missing from the record. No statistical analysis was presented to the district court concerning the results of the 2012, 2014, 2016, or 2018 November general elections for presidential, congressional, or gubernatorial candidates. And we can hardly fault the district court for basing its decision on the evidence presented.

51

We do not suggest by the holding we reach today that section 2 allows a protected group to bring a vote dilution claim in perpetuity and irrespective of its numerical advantage. But we have no occasion to consider today -- like the problem the Fifth Circuit faced in Salas -- the vote dilution claim of a registered- voter minority-majority.

We also caution that section 2 immunizes no group “from the obligation to pull, haul, and trade to find common political ground”; the right guaranteed by section 2 “is equality of opportunity, not a guarantee of electoral success for minority-preferred candidates of whatever race.” De Grandy, 512 U.S. at 1014 n.11, 1020. And we agree with Salas that, when members of a protected class form a majority of a challenged district’s voting-age population, their section 2 vote dilution claims will face an obvious and difficult burden. But this added difficulty is derived from the Gingles analysis itself. After all, a minority-majority group that regularly loses elections will struggle to demonstrate the political cohesion that Gingles requires. Cf. Bartlett v. Strickland, 556 U.S. 1, 16 (2009) (plurality opinion) (rejecting “crossover-district claims” -- “in which success for a minority depends upon crossover majority voters” -- because it “is difficult to see how the majority-bloc-voting requirement could be met in a district where, by definition, white voters join in sufficient numbers with minority voters to elect the minority’s preferred candidate”). On this record, however, we cannot say the district court

52

erred by declining to impose any additional burden onto the well-tread Gingles framework.

The County Board argues in the alternative that the district court clearly erred by giving no weight at all “to the black community’s registration, population, and voting-age population advantages.” We disagree. Although the district court did not give these facts dispositive weight, it plainly made reference to all of them. And it was not clear error -- especially in the face of all the other evidence -- not to give these facts dispositive weight.

Second, the County Board claims the district court clearly erred by finding an available remedy. We think the question of finding a realistic remedy would have been a much closer one in the absence of Dr. Grofman’s special master report and his proposed plans. The problem for the County Board is that it finds itself challenging a different record than the one before us. The County Board complains that Dr. McBride’s illustrative plan, presented to the district court at trial, did not suffice. But we have before us a more complete record -- a record that includes a remedial order from which the County Board took an appeal. The full record answers without any real doubt that a practicable and reasonable remedy is available in this case. The district court’s remedial order and its final judgment expressly relied on Dr. Grofman’s special master report. That report included five proposed remedial district maps; Grofman opined that four of them

53

included three minority opportunity districts (one more than the two afforded by H.B. 836). The district court selected one of those maps -- Map Three -- on the ground that it was the only option that gave black voters in Sumter County an opportunity to elect four candidates of their choice to the school board. The court reasoned, too, that Map Three was the most compact one, and that it was the only map to place public schools in four districts. On this complete record, we discern no clear error.

The Supreme Court’s recent decision in Abbott v. Perez is not to the contrary. There, the Supreme Court considered a claim that the boundary between two state legislative districts in Texas violated section 2 by diluting the votes of the county’s Latino voters. 138 S. Ct. 2305 (2018). Though Latinos made up 56% of the county’s voting-age population, only one of the two districts was a “Latino opportunity district.” Id. The plaintiff’s own expert, however, “determined that it was not possible to divide” the county “into more than one performing Latino district.” Id. (emphasis omitted). Nevertheless, the three-judge panel found in favor of the plaintiff on its section 2 claim, reasoning that the expert had not shown that the illustrative second district would not perform. Id. at 2333. This approach, the Court sensibly observed, “twisted the burden of proof beyond recognition.” Id. But that’s not what happened here. Throughout these proceedings, the district court heard without contradiction that an additional majority-black district easily

54

could be drawn and an effective remedy implemented. Dr. McBride testified to that effect. As far as the district court was concerned, nothing in Dr. Owen’s testimony undermined that expert opinion. And, most significantly, Dr. Grofman reached the same result in his report and in Map Three.

This case is different from the archetypal claim that a plaintiff cannot show an available remedy. Thus, this is different from a case like Bartlett v. Strickland, where the plaintiffs relied on a crossover district to remediate a section 2 violation. See 556 U.S. at 14–19. It is likewise different from Holder v. Hall or Nipper v. Smith, where an effective remedy would have required changing the size of a governing body. See 512 U.S. at 881; 39 F.3d at 1532. And different too from Burton v. City of Belle Glade, where an effective remedy would have required the extraordinary step of adopting a court-ordered annexation. See 178 F.3d at 1199– 200. The remedy adopted by this district court is simple, straightforward, and indisputably available: changing a seven-member board with two at-large seats to one with only single-member districts.

Finally, the County Board challenges the weight the district court afforded to some pieces of evidence. These arguments are unpersuasive as well. In the first one, the County Board complains that the court erred by crediting election results in school board districts where blacks comprised “between 28% and 36%” of the voting-age population. The district court then “compounded this error,” the

55

argument goes, by discounting the results from districts where blacks formed a majority of the voting-age population. But this claim cannot overcome our deferential, clear-error review. The district court made no clear error in its “findings regarding the probative value assigned to each piece of evidence” presented. Solomon, 221 F.3d at 1227. Moreover, as Wright points out, these two exogenous elections are almost “beside the point here.” As the district court recognized, the most probative elections are the three endogenous at-large elections, and in those three races, Sumter County’s black voters are 0 for 3. See Solomon, 221 F.3d at 1227 (a district court “may consider endogenous elections more important than exogenous elections”).

Next, the County Board says the district court relied on an insufficient quantity of data. It’s true, of course, that section 2 focuses on “larger trends” and on “pattern[s] of racial bloc voting that extend[] over a period of time.” Hamrick, 296 F.3d at 1074 (quotation and emphasis omitted). But the County Board cites no support for the proposition that an expert racial bloc voting analysis of twelve elections cannot, as a matter of law, support a vote dilution claim. To the contrary, Dr. McBride opined at trial that his sample size was not too small “in terms of inferential statistics” and that his analysis rested on a “sufficient” number of “endogenous elections.” The district court did not clearly err in relying on the quantity of data in Dr. McBride’s report.

56

Third, the County Board claims the district court “turned a blind eye to dozens of races where Democratic Party candidates (black and white) won the Sumter County total vote.” This the district court did not do. Instead, the court explained that those election results were “of diminished relevance” precisely because “[n]either side has presented a statistical analysis of these races.” As we have already noted, the district court could not have clearly erred by failing to heavily weigh elections that no party analyzed, and exogenous elections at that.

Finally, the County Board claims that the district court erred by ignoring the electoral success of black members who have served on the school board, including in districts without a black majority voting-age population. Specifically, the County Board objects to the court’s observation that no “African American has been elected in a School Board district except in districts where African Americans make up a majority of the voting-age population.” Here, though, the County Board has misread the record. The County Board points us to the success of one former school board member, an African American, Carolyn Whitehead, and says that she prevailed in a 2008 election in a district with a black voting-age population of 49.5%. Yet the district court’s order made clear reference to election results “under the challenged plan,” which was not enacted until 2014. Indeed, the district court took its support for this proposition straight from the County Board’s own statement of undisputed facts. The County Board had not disputed -- nor could it

57

have -- in response to Wright’s motion for summary judgment that no “African- American candidate has been elected under the current school-board plan except in districts where African Americans constitute a majority of the voting-age population.” Again, we see no basis for finding clear error.

The long and short of it is that the district court, after reviewing an extensive evidential foundation, concluded that H.B. 836’s district map impermissibly diluted black voting strength in violation of section 2 of the Voting Rights of 1965. It then conducted remedial proceedings and, with the help of a well-qualified special master, drew new district boundaries that plainly remedied the violation. On this record, we can find no clear error.10 AFFIRMED.

10 The County Board filed a second opening brief following its appeal of the district court’s order enjoining the November 2018 school board elections. Wright moved this Court to strike facts from that brief that were, in his view, outside the record. A motions panel of this Court carried Wright’s motion with the case; we now DENY Wright’s motion.