Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 1 of 171 In the United States District Court for the Western District of Texas SHANNON PEREZ, ET AL. v. GREG ABBOTT, ET AL. § § § § § SA-11-CV-360 ORDER ON PLAN H283 Before Circuit Judge SMITH, Chief District Judge GARCIA, and District Judge RODRIGUEZ Circuit Judge Smith, dissenting XAVIER RODRIGUEZ, District Judge and ORLANDO L. GARCIA, District Judge: This Order addresses Plaintiffs’ claims concerning Plan H283, enacted by the 82nd Legislature in 2011. Plaintiffs assert “results” claims under § 2 of the Voting Rights Act (“VRA”), intentional vote dilution claims under § 2 of the VRA and the Fourteenth Amendment, and Shawtype racial gerrymandering claims under the Equal Protection Clause of the Fourteenth Amendment. The Perez, LULAC, and MALC Plaintiffs also assert one person, one vote claims under the Equal Protection Clause of the Fourteenth Amendment based on population deviations among the districts. This opinion is intended to be read in conjunction with the Court’s fact findings, which are issued separately, as well as the Court’s opinion on Plan C185 (docket no. 1339). 1 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 2 of 171 I. VRA § 2 results claims generally The Task Force, NAACP Plaintiffs, and MALC1 assert § 2 results claims on the basis that Texas could have enacted a plan with more minority opportunity districts (both single-minority and coalition) than were contained in Plan H283 and that enacting a plan with such additional districts was required by the § 2 results test. In their Fourth Amended Complaint, the Task Force Plaintiffs allege that “Plan H283 fails to create at least three additional Latino-majority House districts that afford Latinos the opportunity to elect their preferred candidate.” Docket no. 891 ¶ 37; see also id. ¶ 68 (“The Latino population of Texas is sufficiently geographically compact to comprise the majority of citizen voting age persons in at least 33 Texas House districts.”). They also assert socalled “nudge factor” claims against two HCVAP-majority districts in Plan H283, HD117 in Bexar County and HD78 in El Paso. Docket no. 1282 at 4-5. The Task Force Plaintiffs contend that “Latinos are sufficiently numerous and compact to comprise the citizen voting age majority in more districts than contained in Plan H283, including in Harris County, Nueces County, and the Rio Grande Valley.” Docket no. 1282 at 4. However, somewhat inconsistently, they offer Plan H292 as a demonstration plan, asserting that it has 34 Latino opportunity districts, created by restoring HD33 in Nueces County, “balancing” the Latino population in Bexar County to restore HD117, “balancing” the Latino population in El Paso to add HD78, and adding a Latino opportunity district in the Rio Grande Valley by combining population overages from Cameron and Hidalgo Counties to capture the “organic” district that grew in the 1 The United States asserts only intentional vote dilution claims under § 2, and does not pursue any § 2 results claims. As discussed below, the Court finds that the Perez Plaintiffs are not asserting § 2 results claims, though the pleadings are not entirely clear. 2 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 3 of 171 Valley. Docket no. 1282 at 6; docket no. 444 at 20.2 Defendants correctly note that “only 32 districts exceed 50% HCVAP or SSVR” in Plan H292. Docket no. 468 at 19. The NAACP’s Third Amended Complaint alleges that the Texas Legislative Black Caucus introduced a plan with four additional African-American opportunity districts (Plan H202), and alleges a § 2 results claim. Docket no. 900 ¶¶ 22, 58; see also docket no. 406 at 30 (arguing that four additional minority opportunity districts could have been drawn compared to the enacted plan).3 MALC also asserts § 2 results claims and contends that up to six additional minority opportunity districts were required. Docket no. 897 ¶¶ 74-75; docket no. 412 at 16-18. MALC has offered numerous demonstration maps, including statewide Plans H201, H205, H295, and H329, and various limited plans for certain geographic areas. “When applied to a claim that single-member districts dilute minority votes, the first Gingles condition requires the possibility of creating more than the existing number of reasonably compact districts with a sufficiently large minority population to elect candidates of its choice.” Johnson v. 2 The Task Force also points to this Court’s interim plan, Plan H309, arguing that it has 34 Latino opportunity districts (including HD35 in the Valley, HD78 in El Paso, HD117 in Bexar County, and HD144 in Harris County). Docket no. 1282 at 8. However, this Court did not reach the merits of the § 2 claims when fashioning the interim map, and did not make findings that any districts were required by § 2. Therefore, Plan H309 is not a valid demonstration map absent further evidence supporting the Gingles factors. 3 The Court notes that the NAACP Plaintiffs have not been consistent in their assertions concerning which districts they propose as additional minority districts in their demonstration plan, Plan H202. In their Third Amended Complaint, they note that the Texas Legislative Black Caucus tendered Plan H202 during the session and that “[t]he new districts created by the Legislative Black Caucus were State House District 26 in Fort Bend and Harris County, State House District 54 in Bell County, State House District 107 in Dallas County, and State House District 114 in Tarrant County.” Docket no. 900 at ¶ 22. In their 2011 Trial Brief, they asserted that Plan H202 had four new minority opportunity districts—HD72 in Hidalgo County as a Latino opportunity district; HD107 in Dallas County as a coalition district; HD114 in southeast Tarrant County as a coalition district; and HD132 in west Harris County as a coalition district. Docket no. 288 at 10; see also Tr813-14 (Turner) (discussing HD114). In their 2011 and 2014 Post-Trial Briefs and proposed Findings of Fact, the NAACP Plaintiffs focus only on HD26 in Fort Bend County, HD54 in Bell County, HD107 in Dallas County, and HD149 in Harris County as their proposed additional four districts. Docket nos. 406 at 30, 408 FF31-45, 1280 at 31-39, & 1281 FF121-134, 146-167. The Court will therefore likewise focus on these four districts. Further, although the NAACP Plaintiffs supported Plan H214 as an interim plan, see docket no. 441, they do not brief it as a demonstration Gingles plan, and thus the Court does not consider it. 3 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 4 of 171 De Grandy, 512 U.S. 997, 1008 (1994). Generally, to evaluate this claim, it must be determined how many “reasonably compact districts with a sufficiently large minority population to elect candidates of its choice” exist in Plan H283, and whether Plaintiffs have demonstrated that more were required, which is usually done through presentation of demonstration plans. LULAC v. Perry, 548 U.S. 399, 430 (2006) (“De Grandy requires a comparison between a challenger’s proposal and the ‘existing number of reasonably compact districts.’”). Because Plaintiffs are alleging § 2 claims based on the number of opportunity districts statewide, analysis of the claims should involve a comparison between the number of opportunity districts in the enacted plan (Plan H283) and a Gingles demonstration plan proposed by Plaintiffs. While normally this would be a straightforward task, it is not in this case. This task is made complicated by numerous factors, including that: (1) it remains unclear whether the Perez Plaintiffs are asserting § 2 results claims despite offering statewide demonstration Plan H232 and several limited-area demonstration plans4; (2) the parties (including the Plaintiffs among themselves) disagree concerning which districts in the enacted plan are opportunity districts; (3) even single Plaintiffs present conflicting assertions concerning which districts they are alleging to be additional opportunity districts; (4) some Plaintiffs fail to clearly set out specifically all the districts they contend are opportunity districts either in the enacted plan or their demonstration plan (or both), asserting only that more could have been drawn; (5) Plaintiffs and Defendants disagree on 4 The Perez Plaintiffs’ Sixth Amended Complaint (docket no. 960) is somewhat ambiguous concerning whether they assert § 2 results claims against Plan H283 or only intentional vote dilution claims. Although during this litigation they offered demonstration Plan H232, a plan offered by Rep. Coleman during the legislative session, they do not provide any briefing relevant to § 2 results claims. Further, their Complaint speaks only in terms of intentional discrimination and intentional vote dilution. Accordingly, the Court will treat the Perez Plaintiffs as not bringing any § 2 results claims, and will review their claims only as part of the intentional vote dilution claims analysis. Demonstration plans are relevant to determining dilutive effects for intentional vote dilution claims. 4 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 5 of 171 how to determine whether a district is a minority opportunity district and both incorrectly assert that districts with less than 50% minority CVAP are opportunity districts in some instances; (6) many of Plaintiffs’ experts reports and analyses focus on VAP instead of CVAP; and (7) certain demonstration plans contain 25 districts in Harris County, compared to 24 in Plan H283, making comparisons outside of drop-in counties more difficult.5 The Court finds that certain of the § 2 results claims are moot, given the Legislature’s adoption of the interim plan, with slight modifications, in 2013. Plan H309 and Plan H358 resolved some of the § 2 results claims, and no § 3(c) relief would be available based on any proven violations of the § 2 results claims. Given these facts, as well as the complicating factors listed above, the Court concludes that it would be a waste of the Court’s resources to delve into significant detail into all of the § 2 results claims with regard to Plan H283. Instead, because the Legislature adopted a new plan in 2013, the parties will be permitted to bring their § 2 results claims concerning the 2013 plan based on more recent ACS data. However, the Court will offer some analysis of the 2011 plan § 2 results claims to give preliminary guidance to the parties for the 2013 plan trial and as necessary where the claims relate to the Plaintiffs’ intentional vote dilution claims. A. How to measure opportunity districts For both the enacted plan and Plaintiffs’ demonstration districts, Defendants contend that opportunity districts are measured solely on demographics. For the enacted plan, Plaintiffs contend that they are determined by demographics coupled with a functional analysis including election 5 As will be discussed, the Court concludes that because the 2013 adopted plan currently in effect has 24 districts yet includes the additional opportunity districts Plaintiffs contend are required, results claims relying on a 25district map for Harris County are moot. Because HD149 could be maintained and an additional opportunity district could be drawn in Harris County with 24 districts, the Court finds that the 24-district plans are appropriate demonstration plans for the results claims, while the 25-district plans are not appropriate demonstration plans outside of the drop-in counties because the number of rural counties differs from Plan H283. 5 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 6 of 171 analysis. Some Plaintiffs (such as the Task Force Plaintiffs) also apply a functional analysis to their proposed demonstration districts, though some Plaintiffs (MALC and the NAACP) appear to rely more on meeting the demographic threshold for demonstration districts.6 Some Plaintiffs challenge specific majority-HCVAP districts as not providing real electoral opportunity and not being Latino opportunity districts—specifically, the Task Force Plaintiffs challenge HD78 in El Paso and HD117 in Bexar County under § 2 for both results and intent, and the United States contends that HD35 and HD41 are not opportunity districts as part of its intentional vote dilution claims.7 As this Court held in its congressional plan opinion, Plaintiffs may bring results claims against a particular district despite its majority-HCVAP status; the fact that a district is majority-HCVAP does not, standing alone, qualify it as a Latino opportunity district, and Plaintiffs may attempt to prove that it lacks “real electoral opportunity.” Similarly, as discussed in the congressional plan opinion, the Court views exogenous election indices in enacted districts as probative evidence of whether a district is an opportunity district, but declines to measure whether a district is an opportunity district based solely on a 50% win standard on such indices. Rather, the Court conducts a practical, searching inquiry based on the totality of circumstances and the particular 6 For example, the Task Force uses Dr. Engstrom’s election index (with the criteria of 50% wins) to determine whether a district, either in the enacted plan or their proposed plan, is an opportunity district. Using this metric, the Task Force includes some districts as “Latino opportunity districts” in Plan H292 that do not meet the demographic majorityHCVAP threshold. MALC, on the other hand, appears to focus more on reaching the majority-HCVAP threshold to classify districts in its proposed plans as opportunity districts, although it generally also offered some evidence to support a finding that such districts could perform for Latinos. 7 The United States’ Complaint in Intervention alleges that Plan H283 reduced the number of minority ability to elect districts (a § 5 standard) from 50 in the benchmark to 45. Docket no. 907 ¶ 50. Its post-trial briefing refers to a reduction in the number of minority opportunity districts (the § 2 standard) from 50 to 45 or 46; it views the status of HD41 as unclear due to the number of precinct splits in the district. See, e.g., docket no. 1279 at 36. The United States challenges HD35, HD41, and HD117 as no longer being ability to elect or opportunity districts. Docket no. 907, docket no. 1279. The Court notes that the United States’ expert Lisa Handley used the terms “ability to elect” and “opportunity to elect” interchangeably and applied the same standard to both. TrA599 (Handley). 6 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 7 of 171 facts of the case to determine whether any particular district is an opportunity district for § 2 purposes. In their 2011 proposed Fact Findings and Conclusions of Law, Defendants assert that Plan H283 contains 30 Latino opportunity districts and 12 African-American opportunity districts. Docket no. 413 FF93, 94. Defendants’ count of Latino opportunity districts is based solely on the number of districts with 50% or more HCVAP from the Red-106 Report. Docket no. 413 FF93.8 As discussed, a 50% HCVAP district is theoretically an opportunity district, but it may still be challenged by Plaintiffs as not providing real electoral opportunity. Although Defendants’ fact findings also cite the Red-106 report (the ACS Special Tabulation of CVAP) as support for the number of African-American opportunity districts, the Appendix to the post-trial brief states that Defendants’ criterion for African-American opportunity districts is 40% BVAP. Docket no. 411 Appendix at Table 5 n.1.9 Using the 40% BVAP criteria, Defendants list the following as “African-American districts”: HD22, HD27, HD95, HD100, HD109, HD110, HD111, HD131, HD139, HD141, HD142, and HD146.10 Docket no. 411 Appendix. Defendants claim that no proposed map creates more African-American districts using the 40% BVAP criterion. Defendants’ use of 40% BVAP appears to be based on the position taken by mapdrawers and redistricters, despite the fact that Defendants agree that Supreme Court case law has established a 8 Although Defendants rely on 50% HCVAP in their briefing, it is clear that Interiano, the primary mapdrawer of the House plan, and other redistricters used a 50% SSVR standard for Latino opportunity districts when drawing the House plan. This was partly due to the fact that HCVAP data was not available when they initially began drawing districts (though it was available before the plan went to the floor) and the fact that SSVR data was available in RedAppl. 9 That Defendants use 40% BVAP as the measure for African-American opportunity districts is confirmed by their proposed fact findings filed during the interim plan proceedings. See docket no. 627 FF135. 10 All of these districts are also above 40% BCVAP, but three (HD27, HD95, and HD100) are below 50% BCVAP, although HD95 is within the margin of error. In addition, HD147 has a BCVAP of 47.4% but its BVAP is only 38.2%, so Defendants did not include it as an “African-American district.” 7 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 8 of 171 majority (over 50%) requirement for minority opportunity districts under § 2. The Court thus finds Defendants’ reliance on a 40% BVAP threshold for African-American opportunity districts to be erroneous.11 Rather, for single-minority opportunity districts, the standard for a minority opportunity district is majority CVAP for the particular minority group. B. Whether § 2 can require coalition districts Defendants further contend that minority coalition districts are never required or protected by § 2, and thus they do not count them in any of their analyses. However, Plaintiffs assert that coalition districts can be required and propose certain coalition districts in their Gingles demonstration plans. The Court agrees with Plaintiffs that § 2 can require the creation of minority coalition districts. The Court in Gingles “ha[d] no occasion to consider whether § 2 permits, and if it does, what standards should pertain to, a claim brought by a minority group, that is not sufficiently large and compact to constitute a majority in a single-member district.” Thornburg v. Gingles, 478 U.S. 30, 46 n.12 (1986). Several cases presented this issue after Gingles, but the Supreme Court declined to address the issue and resolved the cases on other grounds. In Growe v. Emison, 507 U.S. 25 (1993), the district in question was composed of minority voters from three different minority groups, including Native Americans. The Supreme Court noted that “Gingles expressly declined to resolve whether, when a plaintiff alleges that a voting practice or procedure impairs a minority’s ability to influence, rather than alter, election results, a showing 11 Although it may be true that 40% BVAP districts usually perform for African-American voters, this has to do with the composition of the rest of the district’s population. Usually such districts are in mixed-race areas, where significant Hispanic population contributes to the performance for African-American voters. But the 40% BVAP criterion, standing alone, cannot make the district an African-American opportunity district. 8 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 9 of 171 of geographical compactness of a minority group not sufficiently large to constitute a majority will suffice. We do not reach that question in the present case either.” Id. at 41 n.5 (citations omittted). Again, in Johnson v. De Grandy, 512 U.S. 997, 1009 (1994), the Supreme Court stated, “As in the past, we will assume without deciding that even if Hispanics are not an absolute majority of the relevant population in the additional districts, the first Gingles condition has been satisfied in these cases.” In LULAC v. Perry, 548 U.S. 399, 443 (2006), the Court, as it had “done several times before,” assumed without deciding “that it is possible to state a § 2 claim for a racial group that makes up less than 50% of the population.” Id. at 443 (Kennedy, J., writing for a plurality). However, Justice Souter opined that it was time to recognize that “the integrity of the minority voting population in a coalition district should be protected much as a majority-minority bloc would be.” Id. at 485 (Souter, J., concurring in part and dissenting in part). The Supreme Court did address the issue of crossover districts in Bartlett v. Strickland, 556 U.S. 1 (2009) (plurality opinion). In Strickland, state house district 18 had been majority minority, but the African American voting age population had fallen below 50%, and it was no longer possible to draw a geographically compact majority-minority district. Id. at 8. The North Carolina legislature split a county in an effort to give “African-American voters the potential to join with majority voters to elect the minority group’s candidate of choice.” Id. Thus, the following question was posed to the Court: “In a district that is not a majority-minority district, if a racial minority could elect its candidate of choice with support from crossover majority voters, can § 2 require the district to be drawn to accommodate this potential?” Id. at 6. Justice Kennedy acknowledged that it had “declined to decide the minimum size minority group necessary to satisfy the first [Gingles] 9 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 10 of 171 requirement . . . [but] [w]e must consider the minimum-size question in this case.” Id. at 12 (Kennedy, J.). The plurality made clear at the outset that it was only addressing the issue as it pertains to crossover districts; it was not addressing coalition districts. As Justice Kennedy explained: In majority-minority districts, a minority group composes a numerical, working majority of the voting-age population. Under present doctrine, § 2 can require the creation of these districts. At the other end of the spectrum are influence districts, in which a minority group can influence the outcome of an election even if its preferred candidate cannot be elected. This Court has held that § 2 does not require the creation of influence districts. The present case involves an intermediate type of district—a so-called crossover district. Like an influence district, a crossover district is one in which minority voters make up less than a majority of the voting-age population. But in a crossover district, the minority population, at least potentially, is large enough to elect the candidate of its choice with help from voters who are members of the majority and who cross over to support the minority's preferred candidate. This Court has referred sometimes to crossover districts as “coalitional” districts, in recognition of the necessary coalition between minority and crossover majority voters. But that term risks confusion with coalition-district claims in which two minority groups form a coalition to elect the candidate of the coalition's choice. We do not address that type of coalition district here. Id. at 13-14 (citations omitted). After considering the parties’ arguments, the Court held that crossover districts are not protected under § 2. Id. at 14-15. However, the scope of the ruling was clarified in several ways. First, the ruling did not apply to coalition districts, as the Court was not addressing that issue. Id. at 13-14. The Court also made clear that “[o]ur holding does not apply to cases in which there is intentional discrimination against a racial minority,” and “if there were a showing that a State intentionally drew district lines in order to destroy otherwise effective crossover districts, that would raise serious questions under both the Fourteenth and Fifteenth Amendments.” Id. at 20, 24. The 10 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 11 of 171 Court stated, “Our holding that § 2 does not require crossover districts does not consider the permissibility of such districts as a matter of legislative choice or discretion.” Id. at 23. While legislatures have a choice to draw such districts, § 2 “does not mandate creating or preserving crossover districts.” Id. Although the Strickland decision may provide some guidance on crossover districts, the Supreme Court has not yet addressed coalition districts—those electoral districts in which two or more minority groups form a coalition to elect the candidate of their choice. While each minority group individually may not be able to meet the first Gingles precondition, they may be able to meet the criteria when combined as a coalition. Coalition districts are different from crossover districts, in which minorities require the help of white crossover voters to elect their candidate of choice. Minority voters in coalition districts do not rely on white crossover votes in order to elect the minorities’ candidate of choice. Again, the Court in Strickland clearly stated that it was not addressing coalition districts. Id. at 13-14. The Court very carefully distinguished crossover and coalition districts and cautioned against confusing the two. Id. at 13. However, the Court’s later decision in Perry v. Perez, 565 U.S. 388 (2012) (per curiam) may have created some of the confusion that the Court cautioned against in Strickland. In referring to a new congressional district in this Court’s interim redistricting plan, the Court in Perez stated, “If the District Court did set out to create a minority coalition district, rather than drawing a district that simply reflected population growth, it had no basis for doing so. Cf. Bartlett v. Strickland, 556 U.S. 1, 13-15, 129 S. Ct. 1231, 173 L.Ed.2d 173 (2009)(plurality opinion).” 565 U.S. at 399. The meaning of this statement, the use of the “Cf.” signal, and the reliance on Strickland is unclear, given that Strickland did not address coalition districts. If the 11 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 12 of 171 Court meant that there was no factual basis for drawing the district, there would be no need for a citation in support. If the Court meant that there was no legal basis for drawing the district, Strickland provides no support because the opinion expressly states that it does not apply to coalition districts. In any event, the Court’s statement in Perez provides no guidance on the issue of whether the majority-minority requirement under the first Gingles precondition mandates that a single racial or ethnic group (e.g., Hispanic or African American, but not a combination of both) constitute a majority of the citizen voting age population or whether the first Gingles precondition may be satisfied when minorities from more than one racial or ethnic group, when joined together, constitute a majority of the citizen voting age population. The Fifth Circuit addressed this issue more than twenty-five years ago and recognized that minority groups may be aggregated to meet the first Gingles precondition. In LULAC v. Midland ISD, 812 F.2d 1494, 1500 (5th Cir.), vacated on state law grounds, 829 F.2d 546 (5th Cir. 1987), the Fifth Circuit approved of the manner in which African Americans and Hispanics were joined together as a compact minority group “capable of carrying a district.” The Fifth Circuit reached the same result one year later in Campos v. City of Baytown, 840 F.2d 1240 (5th Cir. 1988)). In Campos, the district court found that a minority group composed of both African Americans and Hispanics was sufficiently large and geographically insular to form a majority in a single-member district. The district court also found that “Blacks were cohesive, Hispanics were cohesive, together the minority group was cohesive, and that Anglos voted sufficiently as a bloc to usually defeat the minority’s preferred candidate.” Id. at 1242. The district court reviewed the totality of the circumstances, considering the relevant factors, and concluded that vote dilution had occurred and a violation of § 2 had been established. Id. On appeal, the Fifth Circuit found that the district 12 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 13 of 171 court’s finding that the first Gingles requirement was satisfied was not clearly erroneous, and explained: There is nothing in the law that prevents the plaintiffs from identifying the protected aggrieved minority to include both Blacks and Hispanics. Section 1973(a) protects the right to vote of both racial and language minorities. . . . If, together, they are of such numbers residing geographically so as to constitute a majority in a single member district, they cross the Gingles threshold as potentially disadvantaged voters. Id. at 1244 (citations omitted). The Court in Campos also explained, however, that the plaintiffs must prove that the minorities “actually vote together” in a cohesive manner; otherwise, their claim will fail. “The key is the minority group as a whole. . . . If the evidence were to show that the Blacks vote against a Hispanic candidate, or vice versa, then the minority group could not be said to be cohesive. But if the statistical evidence is that Blacks and Hispanics together vote for the Black or Hispanic candidate, then cohesion is shown.” Id. at 1245. Because the Campos plaintiffs showed that African Americans and Hispanics, as a minority group, were politically cohesive, the district court's findings on the second Gingles factor was not clearly erroneous. Id. at 1248. The Eleventh Circuit agreed with the Fifth Circuit’s view on coalition districts in Concerned Citizens of Hardee County v. Hardee County Board of Commissioners, 906 F.2d 524, 526 (11th Cir. 1990), stating that “[t]wo minority groups (in this case blacks and hispanics) may be a single section 2 minority if they can establish that they behave in a politically cohesive manner.” The plaintiffs failed to prove, however, that Black and Hispanic voters in Hardee County were politically cohesive. Id. at 526-27. In Badillo v. City of Stockton, 956 F.2d 884, 886 (9th Cir. 1992), the Ninth Circuit assumed that a combined group of Black and Hispanic voters met the first Gingles precondition, but held that the minority plaintiffs failed to show political cohesion as required under the second Gingles 13 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 14 of 171 requirement. In Bridgeport Coalition for Fair Representation v. City of Bridgeport, 26 F.3d 271 (2nd Cir.), vacated on other grounds, 512 U.S. 1283 (1994), the Second Circuit also assumed that coalition districts are covered under § 2. The district court found that “[c]ombining minority groups to form [majority-minority] districts is a valid means of complying with § 2 if the combination is shown to be politically cohesive.” 26 F.3d at 275. The circuit court agreed that the first Gingles precondition had been met, and found “both testimonial and statistical evidence that African Americans and Hispanics in Bridgeport [were] politically cohesive and that voting in the City [was] remarkably racially polarized.” Id. at 275-76. In Nixon v. Kent County, 76 F.3d 1381 (6th Cir. 1996) (en banc), the Sixth Circuit expressly disagreed with the Fifth Circuit on the issue of coalition districts, thus creating a clear split among the circuit courts. The redistricting plan in Nixon included a district with both African-American and Hispanic voters in order to establish sufficient numbers and satisfactory geographical compactness. Id. at 1384. The district court granted the defendants’ motion to certify the question of whether two protected minority groups may aggregate to pursue a § 2 vote dilution cause of action. Id. at 1383. The circuit court decided that if Congress had wanted to protect a minority group that was composed of more than one race or ethnicity, it would have used more words in the plural form, such as “protected classes” rather than “protected class.” Id. at 1386-87. Thus, it refused to extend § 2 coverage to a minority group that includes more than one race or ethnicity. In 2012, the Second Circuit noted that “[t]he circuits are split as to whether different minority groups may be aggregated to establish a Section 2 claim.” Pope v. Cty. of Albany, 687 F.3d 565, 572 n.5 (2d Cir. 2012). The court ultimately resolved the case on the failure to show racially polarized voting but discussed the first Gingles precondition at length and noted that while plaintiffs may 14 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 15 of 171 choose a more expansive minority group to satisfy the first Gingles requirement, it may also add to their burden in demonstrating political cohesion required for the second precondition. Id. at 574-77 & n.11. To summarize, the Supreme Court has not addressed coalition districts, the Fifth Circuit has expressly permitted them, and the Second, Ninth, and Eleventh Circuits have tacitly recognized them. Only the Sixth Circuit has expressly denied § 2 protection to a combined group of minorities under the first Gingles requirement. As the Fifth Circuit has stated, “We are a strict stare decisis court.” Ballew v. Continental Airlines, Inc., 668 F.3d 777, 782 (5th Cir. 2012). Just as one panel of the circuit court “may not overrule the decision, right or wrong, of a prior panel in the absence of any intervening contrary or superseding decision by the court en banc or the Supreme Court,” a district court is bound by a circuit decision unless or until it is overturned by an en banc decision of the circuit court or a decision of the Supreme Court. See Society of Separationists, Inc. v. Herman, 939 F.2d 1207, 1211 (5th Cir. 1991). In the Fifth Circuit, Campos is binding precedent, and this Court must follow it in the absence of any authority to the contrary. Thus, if Plaintiffs can meet their burden of proof in all other respects, their § 2 claim will not fail simply because the minority group in question is composed of more than one race or ethnicity. This is consistent with the manner in which the Supreme Court has dealt with earlier cases, in which it assumed (without deciding) that it was permissible for the district court to combine distinct ethnic and language minority groups for purposes of assessing compliance with § 2. See, e.g., Growe, 507 U.S. at 41 (unanimous opinion). Rather than impose a strict prohibition, the Supreme Court simply cautioned that “when dilution of the power of such an agglomerated political bloc is the basis for an alleged violation, proof of minority political cohesion is all the more 15 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 16 of 171 essential.” Id. (citing Badillo v. Stockton, 956 F.2d 884, 891 (9th Cir. 1992); Concerned Citizens of Hardee County v. Hardee County Bd. of Comm'rs, 906 F.2d 524 (11th Cir. 1990); Campos, 840 F.2d at 1244). This is also consistent with the intent of Congress when it amended the Act in 1982. In amending § 2, Congress emphasized the need for courts to undertake a searching practical evaluation of the “past and present reality.” S. Rep. 97-417 at *30 (citing White v. Regester, 412 U.S. 755, 76077 (1973)). Coalitions of minority voters are a present reality, and affording them protection under § 2 is consistent with the Congressional goal of keeping political processes “equally open to minority voters.” S. Rep. 97-417 at *2. Minorities must still “pull, haul, and trade to find common political ground,” De Grandy, 512 U.S. at 1020, and they will not meet the second Gingles precondition if they do not. But once that has been achieved and the minority voters, when combined, constitute a majority, the first Gingles precondition is also satisfied. The bottom line is that every case is different, and “the Gingles factors cannot be applied mechanically and without regard to the nature of the claim.” Voinovich v. Quilter, 507 U.S. 156, 158 (1993). Thus, this Court follows the Fifth Circuit and holds that § 2 can require the creation of coalition districts, provided that the Gingles criteria (and totality of the circumstances) are satisfied as to the coalition. C. Whether the VRA can require the State to violate the County Line Rule12 It is undisputed that mapdrawers did not create certain opportunity districts even though the proposed districts met their population thresholds because they felt that doing so would have 12 The Texas County Line Rule is contained in Article III, § 26 of the Texas Constitution and provides that “whenever a single county has sufficient population to be entitled to a Representative, such county shall be formed into a separate Representative District, and when two or more counties are required to make up the ratio of representation, such counties shall be contiguous to each other; and when any one county has more than sufficient population to be entitled to one or more Representatives, such Representative or Representatives shall be apportioned to such county, and for any surplus of population it may be joined in a Representative District with any other contiguous county or counties.” 16 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 17 of 171 required a County Line Rule violation. In their briefing, Defendants argue that requiring the Legislature to violate the County Line Rule to comply with § 2 would violate the Equal Protection Clause and the Shaw v. Reno line of cases because they would be subordinating traditional redistricting principles to race. Docket no. 1295 at 54-55; docket no. 996 (Motion for Summary Judgment) at 28-31. The Court disagrees. The Supremacy Clause generally requires that state laws that are inconsistent with federal laws must yield to the federal law. As a result, it is well established that state constitutional principles such as the County Line Rule must yield to the Equal Protection Clause’s one-person, onevote requirement. Bartlett v. Strickland, 556 U.S. 1, 6 (2009) (“It is common ground that state election-law requirements like the Whole County Provision may be superseded by federal law—for instance, the one-person, one-vote principle of the Equal Protection Clause of the United States Constitution.”). Mapdrawers recognized this requirement by splitting one county in the map to maintain an overall deviation below 10%, but they asserted that this was the only county cut permitted by law. And although the TLC had advised that the County Line Rule would also have to yield to the VRA and Texas had itself taken this position in prior redistricting litigation,13 redistricting leadership flatly rejected that position. In Strickland, the Supreme Court considered the question “whether § 2 of the Voting Rights Act requires district lines to be drawn that otherwise would violate the Whole County Provision” in North Carolina’s Constitution. 556 U.S. at 7. The Court then stated, “That, in turn, depends on how 13 See, e.g., D-124 (Hanna’s County Line Rule presentation advising that the County Line Rule would have to yield to the VRA); D-128 at 142 (Hanna’s guidance book on redistricting for the Legislature, noting that the State argued in Clements v. Valles that it was required to violate the County Line Rule to comply with the VRA); Clements v. Valles, 620 S.W.2d 112 (Tex. 1981). 17 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 18 of 171 the statute is interpreted.” Id. The Court noted that § 2 “can require the creation of [majorityminority] districts” but concluded that the VRA did not require the particular district at issue to be drawn because the plaintiffs had failed to establish the first Gingles precondition under the Court’s newly established majority standard. Id. at 13-15. Thus, because the district was not required by § 2 in that case, the Supreme Court did not directly decide the initial issue of whether § 2, when satisfied, could have required the state to draw a district in violation of the Whole County Provision. Nevertheless, such a conclusion is implied, given that the Court could simply have held that § 2 could not require any district to be drawn in contravention of the Whole County Provision, regardless of how the statute is interpreted and regardless of whether § 2 requirements were satisfied, rather than determining whether § 2 required the district in the first instance. Nevertheless, Defendants argue that the “Legislature’s adherence to the Texas Constitution was not only a rational exercise of race-neutral policy; it was essential to avoid a violation of the Equal Protection Clause” and that Plaintiffs’ “suggesting that the [VRA] compels subordination of traditional redistricting principles to race, would simply redirect the strict scrutiny analysis to section 2 itself.” Docket no. 411 at 38-39. Defendants contend that Supreme Court jurisprudence holds “beyond question that traditional redistricting principles cannot be subordinated to race without running afoul of the Fourteenth Amendment.” Id. at 39-40. However, the Supreme Court’s Equal Protection cases debunk this argument. Justice O’Connor, writing for a plurality, held that, assuming compliance with § 2’s results test is a compelling state interest (this Court holds that it is), the “narrow tailoring” requirement of strict scrutiny allows states a limited degree of leeway in furthering such interests—“[i]f the State has a ‘strong basis in evidence’ for concluding that creation of a majority-minority district is 18 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 19 of 171 reasonably necessary to comply with § 2, and the districting that is based on race ‘substantially addresses the § 2 violation,’ it satisfies strict scrutiny,” meaning that it operates as a defense to a charge of racial gerrymandering. Bush v. Vera, 517 U.S. 952, 977 (1996) (plurality opinion). Further, the plurality noted, the “district drawn in order to satisfy § 2 must not subordinate traditional districting principles to race substantially more than is ‘reasonably necessary’ to avoid § 2 liability.” Id. at 979. On its face, this language makes clear that traditional districting principles may be subordinated to race when necessary to avoid § 2 liability (but no more than necessary). Similarly, the Fifth Circuit has held that a majority-minority district “is constitutional if the State has a ‘strong basis in evidence’ for concluding that the three Gingles preconditions are present and if the district drawn in order to satisfy § 2 does not ‘subordinate traditional districting principles to race substantially more than is “reasonably necessary” to avoid § 2 liability.’” Clark v. Calhoun Cty., Miss., 88 F.3d 1393, 1407 (5th Cir. 1996); see also id. at 1406 (“a tailored response to a found [§ 2] violation must use race at the expense of traditional political concerns no more than is reasonably necessary to remedy the wrong”). Thus, traditional districting principles such as the County Line Rule may be subordinated to race to remedy a § 2 violation, so long as it is no more than reasonably necessary to comply with § 2.14 Cf. Bethune-Hill v. Va. State Bd. of Elections, 137 14 There are also other cases in which the Supreme Court and Fifth Circuit have recognized that the County Line Rule (and similar rules from other states) must yield to the VRA. See Bartlett v. Stephenson, 535 U.S. 1301, 1302-04 (Rehnquist, C.J.) (denying North Carolina’s application for a stay of a state supreme court order that required, in part, “the new plan ‘must preserve county lines to the maximum extent possible, except to the extent counties must be divided . . . to comply with Section 2 of the Voting Rights Act’”); LULAC v. Clements, 884 F.2d 185, 188 (5th Cir. 1989) (finding that a county could not intervene in redistricting litigation for the purpose of “objecting to the proposed reduction of the district size in violation of Texas law” because “[i]f the present district lines are found to violate the Voting Rights Act and/or the United States Constitution, Texas’ constitutional and statutory provisions protecting district lines will be nullified under the Supremacy Clause”). In addition, numerous courts have recognized the broader possibility that state election laws of any sort should yield to the VRA by operation of the Supremacy Clause. See Abrams v. Johnson, 521 U.S. 74, 79 (1997) (“When faced with the necessity of drawing district lines by judicial order, a court, as a general rule, should be guided by the legislative policies underlying the existing plan, to the extent those policies do not lead to violations of the Constitution or the 19 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 20 of 171 S. Ct. 788 (2017) (affirming district court’s finding that although race predominated in the drawing of District 75, the district survived strict scrutiny because the legislature’s use of race was narrowly tailored to complying with § 5 of the VRA). Defendants’ position conflates the two steps of a Shaw-type Equal Protection analysis. When a party asserts a claim under Shaw that a districter has racially gerrymandered a district, the first inquiry is whether traditional redistricting principles are subordinated to race such that racial criteria predominated; if so, strict scrutiny is triggered. In Defendants’ view, the first step would be the end of the inquiry. However, as discussed above, strict scrutiny allows the use of race when narrowly tailored to further a compelling state interest, such as complying with the VRA. Accordingly, strict scrutiny permits the subordination of traditional redistricting principles to race insofar as reasonably necessary to avoid § 2 liability and/or remedy the wrong. While states choosing to draw districts may avoid strict scrutiny under the Equal Protection Clause by respecting their own traditional districting principles and the states retain “‘discretion to apply traditional districting principles,’ in majority-minority, as in other, districts,” Bush, 517 U.S. at 978, this does not mean that they may claim that a single traditional districting principle such as the County Line Rule allows them to avoid Voting Rights Act.” (emphasis added)); Katzenbach v. Morgan, 384 U.S. 641, 646-47 (1966) (“We hold that . . . by force of the Supremacy Clause, Article VI, the New York English literacy requirement [as a precondition to the right to vote] cannot be enforced to the extent that it is inconsistent with § 4(e) [of the VRA].”); Voinovich v. Quilter, 507 U.S. 146, 159 (1993) (noting a plan’s drafter’s “disregard[ing] the requirements of the Ohio Constitution where he believed that the [VRA] required a contrary result” demonstrated his “obedience to the Supremacy Clause”); State of S.C. v. Katzenbach, 383 U.S. 301, 333–34 (1966), abrogated by Shelby Cty., Ala. v. Holder, 133 S. Ct. 2612 (2013) (recognizing that the VRA could remedy a state constitution’s violation of the Fifteenth Amendment, such as improperly used literacy tests); Lopez v. Monterey Cty., Cal., 871 F. Supp. 1254, 1258 (N.D. Cal. 1994) (“Under the Supremacy Clause, the implementation of relief for a violation of the Voting Rights Act must take precedence over enforcement of state law that stands in the way of effective relief.”) (quoting Wise v. Lipscomb, 437 U.S. 535, 540 (1978)); Marylanders for Fair Representation, Inc. v. Schaefer, 849 F. Supp. 1022, 1047 n.21 (D. Md. 1994) (“Plaintiffs are correct in arguing that state statutes or constitutional provisions that result in violations of the U.S. Constitution or the Voting Rights Act are preempted by federal law. On the other hand, it is self-evident that a state plan should—indeed must—comply with state statutory and constitutional mandates if compliance with those mandates does not violate federal law” (internal citations omitted)). 20 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 21 of 171 drawing districts required by § 2 under the totality of circumstances.15 Thus, the Court rejects Defendants’ argument that the VRA cannot require the Legislature to violate the County Line Rule. The Court will thus consider Plaintiffs’ § 2 results claims and Gingles maps, including those that create purported coalition districts and that violate the County Line Rule. However, the Court emphasizes that the VRA requires the County Line Rule to be subordinated only to the extent reasonably necessary to remedy a § 2 violation. Accordingly, Plaintiffs’ demonstration maps should attempt to honor the County Line Rule to the extent doing so is consistent with the VRA. II. Intentional Vote Dilution and Select § 2 Results Claims by Geographic Area The Court applies the analytical framework for intentional vote dilution claims under § 2 of the VRA and the Fourteenth Amendment as discussed in the congressional plan opinion. For § 2 results claims, the Court applies the Gingles framework, including as discussed above. The United States asserts intentional vote dilution claims with regard to the entire plan, but points to specific districts as proof of intent, including districts in the Rio Grande Valley (including specifically HD41), HD35, Nueces County, El Paso (HD77 and HD78), Dallas County (HD103, HD104, and HD105), Bexar County (HD117 and HD118), and Harris County (HD137 and HD149). Docket no. 1304 at 2 n.1. The United States relies on circumstantial evidence as well as direct evidence, including: (1) the nudge factor emails and evidence that the nudge factor was implemented 15 Two district courts have recently rejected Defendants’ same argument. The court in Georgia State Conference of NAACP v. Fayette County Board of Commissioners, 950 F. Supp. 2d 1294 (N.D. Ga. 2013), vacated on procedural grounds, 775 F.3d 1336 (11th Cir. 2015), observed, as does this Court, that Defendants’ argument “ignores the applicable framework of an equal-protection claim” and that, upon a finding that a plan subordinates traditional districting principles to race, “the district is not simply rejected as a racial gerrymander” but instead “the court applies strict scrutiny.” Id. at 1304. The court further noted that Defendants’ position “would have the Court collapse an equalprotection inquiry into the first Gingles prong” but “the court must first determine whether Gingles is met before ensuring that the proposed remedy complies with the Equal Protection Clause.” Id. at 1305. The court in Montes v. City of Yakima, 40 F. Supp. 3d 1377, 1401 (E.D. Wash. 2014) also rejected Defendants’ position for “the reasons so cogently explained in Fayette County.” 21 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 22 of 171 in certain Latino opportunity districts that elected an Anglo-preferred candidate in 2010; (2) racially focused statements by legislators and districting insiders showing they thought and spoke in terms of race; and (3) precinct splits, including the high number of splits in general (412 in Plan H283) and the indication that many splits were racial and indicate the use of race as a proxy for partisanship given that accurate political information is not available below the precinct level, while racial data is. Docket no. 1279 at 11-14. The Perez Plaintiffs also complain that Plan H283 “intentionally discriminated against minority voters.” Docket no. 960 ¶ 28. They (joined by the NAACP Plaintiffs) specifically allege such discrimination in Dallas County, Tarrant County, Harris County, Bell County, McLennan County, Fort Bend County, Nueces County, Bexar County, and HD41. Id.; ; docket no. 1303; docket no. 601. The NAACP alleges that Plan H283 was developed with the intent to disadvantage AfricanAmerican and other minority voters. Docket no. 900 ¶ 62. MALC challenges Plan H283 as a whole as intentionally diluting Latino voting strength by failing to draw new opportunity districts despite the population growth, using the County Line Rule to avoid drawing minority opportunity districts required by § 2 and diminishing Latino voting strength, packing and cracking politically cohesive Latino and minority communities, manipulating population, and racial gerrymandering. Docket no. 897 ¶¶ 2-10. MALC also specifically alleges that mapdrawers: pretextually used the County Line Rule to avoid drawing a new Latino opportunity district in the Midland/Odessa area of West Texas (¶ 38); eliminated a Latino opportunity district in Nueces County (¶ 42); failed to draw an additional Latino opportunity in Harris County (¶¶ 4748); failed to provide minority opportunity in Fort Bend County and unnecessarily fragmented the minority population there to minimize its political strength (¶ 51); failed to draw an additional 22 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 23 of 171 minority opportunity district in Bell County and intentionally fragmented the minority community in Killeen to minimize its political impact (¶¶ 54, 56); intentionally fragmented minority population while at the same time overpopulating Latino majority districts in Dallas County, leading to the failure to create at least one and possibly two additional minority opportunity house districts in Dallas County (¶ 57); impermissibly focused on race and targeted low-turnout Hispanic voters for inclusion in HD117 in Bexar County to protect incumbent Republican John Garza (¶ 58); and drew the border between HD77 and HD78 in El Paso with racially discriminatory intent (¶ 59). Defendants assert that the mapdrawers’ application of the County Line Rule is not evidence of intentional vote dilution because it was a consistently applied neutral, traditional districting principle. Docket no. 457 at 68. Defendants further generally deny any improper use of race, arguing that any consideration or use of race was only as needed to comply with the VRA. A. El Paso County Plaintiffs assert that the El Paso County configuration, particularly HD77 and HD78, is evidence of intentional vote dilution. Plaintiffs assert “nudge factor” claims against HD78 and contend that it intentionally does not provide Latinos equal opportunity in an effort to protect the Republican incumbent Dee Margo. Plaintiffs also assert that five compact Latino opportunity districts should have been drawn in El Paso County compared to only four opportunity districts in Plan H283. Although the HCVAP of HD78 is over 50% (55.2%), Defendants admit that “Plan H283 did not create an additional majority-Latino district in El Paso,” docket no. 413 FF71, presumably because it did not meet their 50% SSVR threshold even though it was (and all four other districts in El Paso County were) over 50% HCVAP. The United States asserts that mapdrawers’ intentional refusal to create a fifth Latino 23 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 24 of 171 opportunity district in El Paso, even though it could be done without violating the County Line Rule or other traditional districting principles, and instead making a “policy choice” to protect the Anglo incumbent was intentional vote dilution. Docket no. 1279 at 36, 39-40. To do this, the United States argues, Downton manipulated the SSVR in HD78 and analyzed election data to ensure that the Hispanic candidate of choice was unlikely to be elected. Docket no. 1279 at 40. The Task Force Plaintiffs argue that the Latino population “could be more evenly spread among all five districts in El Paso County,” giving Latinos the opportunity to elect their candidate of choice in all five districts, and that Plan H283 packs Latino voters in El Paso County into the other districts to prevent Latino voters in HD78 from electing their candidates of choice. Docket no. 634 FF458, 459; docket no. 1282 at 2. The Task Force Plaintiffs contend that HD78 in Plan H283 is not a Latino opportunity district but that it could be drawn as an additional Latino opportunity district as in their demonstration Plan H292. Docket no. 1282 at 6. MALC’s Third Amended Complaint notes that the border between HD77 and HD78 has a bizarre shape with deer antler protrusions that split multiple precincts between these two districts. Docket no. 897 ¶ 59. MALC notes that its Plan H205 creates five reasonably compact majoritySSVR and majority-HCVAP districts in El Paso. Docket no. 412 at 18.16 Defendants assert that the “possibility of creating an additional Latino-majority Democratic district does not obligate the State to create it” and that Plaintiffs have failed to show that the totality of circumstances required such a district, especially given that “Democratic voters are consistently 16 In the interim plan, the Court remedied the “not insubstantial” § 5 claims in El Paso, finding that this also provided “an appropriate remedy for the alleged Section 2 . . . violations.” Docket no. 690 at 11. The interim plan increased HD78's HCVAP to 58.3% “to provide minorities the ability to elect the candidate of their choice.” Id. at 12. The Court finds that the § 2 results claims against the El Paso County districts are moot given the changes in the current plan. 24 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 25 of 171 successful in electing their candidates of choice in El Paso County” and “the alleged number of Latino candidates of choice elected to the House from El Paso County—four out of five—is proportional to the Latino percentage of citizen voting age population.” Docket no. 457 at 33. Defendants further assert that the El Paso map was developed and agreed to by the fivemember delegation, four of whom were Democrats and members of MALC, and all of whom were elected from HCVAP-majority districts. Docket no. 1249-1 at 10. Defendants point out that under both Plan H100 and Plan H283, all five districts were over 50% HCVAP and all but HD78 were over 50% SSVR, and assert that the levels of both were “substantially similar across both plans.” Docket no. 1249-1 at 13. Defendants argue that “there is no evidence that the configuration of HD78 deprives any Latino voter in El Paso County of an ‘equal opportunity’ to participate in the political process or to elect candidates of their choice” and that, even if Plaintiffs could satisfy all three Gingles preconditions, they fail to “prove that the totality of circumstances requires the State to restructure HD78 to unseat the incumbent” because “Latinos are more than proportionally represented in El Paso County’s House delegation,”17 “[t]he only possible basis for Plaintiffs’ claim is the Legislature’s failure to maximize Latino representation,” and “[t]he Legislature’s decision to maintain existing Latino population levels in the district, presumably offering some protection to the incumbent, does not support a finding of vote dilution.” Docket no. 411 at 42; see also docket no. 17 As MALC points out, Defendants agree that § 2 proportionality should be evaluated on a statewide basis when Plaintiffs assert statewide vote dilution claims, yet argue that no fifth Latino opportunity district is required in El Paso County because Latinos already have proportional representation with four districts. Docket no. 459 at 3. The Court finds that overall statewide proportionality is the relevant inquiry for Plaintiffs’ § 2 claims. To achieve statewide proportionality, Latinos may be over- or under-represented in certain areas of the State. Exact proportionality in every area is not feasible, especially given the operation of the County Line Rule. Thus, while local proportionality is relevant in the totality of circumstances in determining whether an additional Latino opportunity district is required, § 2 proportionality in this case must ultimately be measured on a statewide basis, and over-representation in any given area cannot be the basis for rejecting an opportunity district. 25 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 26 of 171 457 at 31-32. The Court finds that Downton reconfigured the border between HD77 and HD78 to intentionally dilute the Latino vote in HD78. Although Pickett, the County delegation dean, submitted two County proposals to Solomons, Downton chose the one that was not preferred by Pickett,18 and he then made further changes to the border between HD77 and HD78 without input from the delegation. See TrJ2103 (Downton did not incorporate input from Margo or Marquez). The evidence appears undisputed that, after choosing to go forward with the Marquez map from the El Paso County delegation, Downton made further unilateral, race-based changes to the border of HD77 and HD78, ostensibly to address concerns raised by David Hanna about the fact that the SSVR of HD78 had dropped below benchmark levels. Hanna’s memo stated that this “risk of retrogression . . . could easily be remedied by swapping some precincts with an adjoining district,” especially in light of the disparity in SSVR between HD78 and the other El Paso districts. D-122. However, rather than follow Hannah’s advice to increase the SSVR of HD78 by swapping a few precincts, Plaintiffs contend that Downton surgically split precincts and moved individual census blocks between HD77 and HD78, while monitoring election data to ensure that he did not create a Latino opportunity district. Docket no. 1279 (United States Brief) at 39-40. The Court agrees with this assessment. Although Downton claimed to make these race-based splits with the goal of complying with § 5 by raising the SSVR of HD78 (TrJ2002, TrJ2102), his so-called “compliance” with § 5 was 18 The evidence also indicates that Rep. Marquez may have impermissibly focused on race in making changes to HD77 and HD78 from the initial delegation map supported by Pickett. She specifically moved precincts into her district to increase her SSVR numbers in order to increase her own election chances, but there is no evidence that this was necessary for any compelling state interest such as complying with the VRA. However, the evidence is clear that any race-based changes made by Marquez were not made with racially discriminatory intent. 26 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 27 of 171 intentionally superficial because he wanted to protect the Republican incumbent elected in 2010, who was not the Latino candidate of choice. Downton’s changes were designed to appear to comply with § 5 by increasing the SSVR of HD78, but without increasing or ensuring Latino ability to elect, which is the basis upon which § 5 permits such race-based districting actions. Downton’s precinct splits were deliberate and based on more than simply increasing SSVR; his purpose was to increase the SSVR while simultaneously ensuring that election success rates remained minimally improved for Latinos. Downton achieved his purpose by increasing the total SSVR by 1% (and non-suspense SSVR by .9%), while increasing election performance for Latinos by only .3 or .4%. PL-503; D-109. Further, Downton and mapdrawers knew from the OAG 10 that although HD78 remained at 2/10 wins for Latino-preferred candidates, the margin of victory for the two prevailing Latino-preferred candidates had been reduced from the benchmark. US-190; US-190A. This is not compliance with § 5, which looks to numerous factors, including Latino election success, in determining whether a district has “ability to elect,” and does not just focus on a single demographic such as SSVR. In fact, increasing the SSVR while at the same time monitoring Latino ability to elect to ensure that it is not correspondingly increased is antithetical to the purpose of § 5 and the VRA and is intentionally racially discriminatory. The changes cannot be excused as mere partisan gerrymandering because of their racially discriminatory nature. Given the existence of racially polarized voting in El Paso County and the totality of circumstances, this manipulation constitutes intentional vote dilution in violation of § 2 of the VRA and the Fourteenth Amendment. In sum, the § 2 results claims in El Paso County are moot, but the Court finds that mapdrawers intentionally diluted the Latino vote in violation of § 2 of the VRA and the Fourteenth 27 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 28 of 171 Amendment with regard to HD78. B. Bexar County Benchmark HD117 was a Latino opportunity district, electing the Hispanic candidate of choice in three out of the five last endogenous elections. In 2010, it elected Republican John Garza, who was not the Latino candidate of choice. Numerous Plaintiffs contend that the Legislature intentionally diluted Latino voting strength in the district to protect Garza. Plaintiffs19 argue that HD117 was no longer a Latino opportunity district in Plan H283 despite its majority-HCVAP and majority-SSVR status. The D.C. Court found that HD117 was an ability-to-elect district in the benchmark, but was no longer an ability district in Plan H283. Texas v. United States, 887 F. Supp. 2d 133, 170-71 (D.D.C. 2012), vacated on other grounds, 133 S. Ct. 2885 (2013). Based on a preliminary finding that mapdrawers were impermissibly focused on race in trying to make HD117 more Republican, this Court remedied the “not insubstantial” § 5 claim in the interim plan, reconfiguring HD117 to return it to benchmark performance levels. Docket no. 690 at 6. The Task Force Plaintiffs bring a “nudge factor” claim against HD117 in Bexar County. The United States also asserts a nudge factor claim, arguing that mapdrawers replaced high-turnout Hispanic voters with poor, low-turnout Hispanic precincts to increase Hispanic population levels but decrease SSVR, Hispanic turnout, and Hispanic electoral performance. Docket no. 1279 at 41. MALC also alleges that “[i]n creating a district to safely re-elect Rep. Garza the state impermissibly focused on race by targeting low-turnout Latino precincts.” Docket no. 897 ¶ 58. Defendants argue that the Bexar County map was developed by the ten-member delegation, 19 According to their proposed conclusions of law (docket no. 1265), the Perez Plaintiffs allege there is intentional vote dilution in HD117, but it is not briefed in their Post-Trial Brief (docket no. 1263). 28 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 29 of 171 led by Democrats Villarreal and Ruth Jones McClendon. Docket no. 1249-1 at 15. Defendants contend that Garza wanted rural areas because he viewed them as more conservative and more likely to vote Republican, and that he wanted to create a rural, conservative district outside the City of San Antonio20 to bolster his re-election prospects while maintaining SSVR over 50%, and that this is not evidence of racial animus. Docket no. 1249-1 at 17, 1249-2 at 23 (citing TrJ399 (Garza), TrJ1518, TrJ1523 (Interiano)). Defendants point to Farias’s testimony that it was a partisan decision. Docket no. 1272 at 56 (citing Tr353-54). Defendants argue that Speaker Straus and his staff determined that in order to achieve these goals, HD117 needed to include rural areas of southern Bexar County that were previously represented by Farias under the benchmark. Docket no. 1249-2 at 23 (citing TrJ1559 (Interiano)). Defendants assert that ultimately HD117 was agreed to by Rep. Villarreal and 9 of 10 members of the Bexar County delegation, that even if Garza had an improper motive, there is no evidence other members of the delegation were aware of that motive, and that Plaintiffs have not shown that Garza’s individual motivations were a substantial or motivating factor behind the Legislature’s adoption of the plan. Docket no. 1249-2 at 23-24; docket no. 1272 at 54. It is undisputed that Garza’s initial ideal district was rejected by redistricting leadership because it fell below 50% SSVR and that mapdrawers and redistricting leadership felt that HD117 had to be maintained above 50% SSVR to avoid retrogression. Defendants’ argument that this was an agreed delegation map is demonstrably false with regard to the configuration of HD117. Rather, 20 As noted in the fact findings, this is inconsistent with Garza’s own testimony that other delegation members wanted some of the stronger areas within the City that had been in his benchmark district, and that he wanted to maintain some of this territory. TrJ372, TrJ378 (Garza). Further, Garza’s proposed “ideal district” included some portions of the City. 29 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 30 of 171 the evidence is clear that Garza’s and Larson’s staff worked with Interiano to draw the configuration of HD117 that was placed into Plan H283, and that Farias vigorously objected to the configuration of his district. The delegation members (except Garza and Straus) also voted against the motion to table Farias’s proposed amendment. D-190 at 2294-95. Interiano worked to draw a district with exactly 50.1% SSVR to maintain its appearance as a Latino opportunity district and avoid retrogression under the mapdrawers’ majority-SSVR criterion, while minimizing Hispanic turnout. He added in areas such as Somerset and Whispering Winds with higher Hispanic population but low Hispanic voter turnout and removed areas inside San Antonio, such as areas of South San Antonio ISD, that had highly mobilized Hispanic voters. Hispanic population was manipulated to maintain exactly 50.1% SSVR while minimizing Hispanic turnout and electoral performance. Although Interiano testified that they were simply trying to balance increasing SSVR with “keeping political numbers up,” his explanations for the district were not credible, as discussed in the fact findings. His testimony that Garza wanted to stay outside the City was not credible, as noted. His testimony that Garza was concerned with water issues also was not credible, given that Garza himself testified that the political issues surrounding Bexar Met were independent from redistricting, and he did not have any goals in redistricting relating to taking more or less of Bexar Met territory. TrJ405-08. Interiano’s denial of using turnout to shape the district is not credible. Rather, the Court finds that Interiano drew the district to increase SSVR while intentionally minimizing any gains in Latino electoral performance. This is consistent with Garza’s own testimony. Garza testified at his deposition that his ideal district went farther north because the area was “more Anglo and more conservative,” and he preferred more Anglo and more conservative areas 30 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 31 of 171 because “[t]hey would tend to vote Republican.” PL-454 (Garza 10-19-11 depo.) at 30-31; TrJ36769 (Garza). He also testified that he and his staff had looked at turnout, including the turnout between Anglos and Hispanics. TrJ373. He testified that rural Hispanic turnout tended to be low, and he thought Somerset Hispanic turnout would be low. TrJ374, TrJ403-04. Further, Interiano and Garza’s staff knew that Garza did not tend to win SSVR-majority precincts. PL-1664. The wishes of senior member Farias to maintain his relationship with Whispering Winds and Somerset were ignored and were not permitted to interfere with the goal of protecting Garza by suppressing Latino electoral performance. And Solomons’ objection to Farias’s proposed amendment, which would have kept the district at 50.1% SSVR—that it contained more split precincts—was obviously pretextual given the number of split precincts in the plan created by his own mapdrawers. The final configuration of HD117 includes a very large gap between HCVAP and SSVR of 13.7 points, well beyond any gap in any other House district.21 By minimizing Hispanic voter registration and turnout, Interiano successfully decreased the performance of HD117. According to Handley, HD117’s performance on the exogenous election index went from 60% to 20%. Considering all the evidence, that Court finds that mapdrawers intentionally drew HD117 with 50.1% SSVR but with lower performance for Latinos by manipulating Latino population and turnout. Interiano was given authority to draw the maps by redistricting leadership, and his knowledge and motives must be imputed to the Legislature as a whole, regardless of whether individual members of the Legislature were aware of the information when they voted for the map. The configuration of HD117 is evidence of intentional vote dilution under § 2 of the VRA and the 21 Although Garza indicated that he would be happy with the higher HCVAP because those Hispanics could be registered to vote, he admitted that he and his staff did not take any actions to increase registration, TrJ432, and this testimony is also undermined by the fact that Garza and redistricters refused to allow HD117 to exceed 50.1% SSVR. 31 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 32 of 171 Fourteenth Amendment. In sum, any § 2 results claims in Bexar County are moot, but the Court finds that mapdrawers intentionally diluted the Latino vote in HD117 in violation of § 2 and the Fourteenth Amendment. C. Nueces County It is undisputed that Nueces County had two benchmark Latino opportunity districts, though they both elected Republicans in 2010. It also contained part of a third district represented by Anglo Republican Todd Hunter. When faced with slower population growth in Nueces County such that it would only be entitled to two districts under the County Line Rule, Hanna advised mapdrawers that they had three options: (1) draw one “performing” Hispanic district and one not; (2) draw two equally Hispanic districts, which may not perform reliably; or (3) see if both Latino opportunity districts could be preserved by splitting county lines. D-122. Hanna advised that the County Line Rule would have to yield to the VRA if retrogression could be avoided by splitting the county and it would clearly contribute to total Hispanic voting strength statewide. As noted in the fact findings, mapdrawers chose the first option because they felt it allowed them to maintain more Republican seats and protect Republican Anglo incumbent Hunter. Mapdrawers did not look into whether two Hispanic districts could be maintained in Nueces County, either wholly within Nueces County or by breaking the County Line Rule.22 22 Mapdrawers and Hanna were concerned with whether districts were “performing” for § 5 purposes and used SSVR and election analysis (the OAG 10) to determine whether a district was “performing.” In deciding to create one “performing district” and one Anglo-Republican district favoring the incumbent Hunter, redistricters felt they could eliminate HD33 for § 5 retrogression purposes so long as they “offset” the loss somewhere else in the State. There is no indication that they seriously looked at whether § 2 might require two opportunity districts in Nueces County. Nueces County presents a strong case for requiring the State to follow § 2 of the VRA over the County Line Rule because the Legislature used the County Line Rule as a basis for destroying an existing Latino opportunity district (as opposed to simply refusing to draw a new one). HD33 was 60.4% HCVAP using 2005-2009 ACS data available to the Legislature and was completely within Corpus Christi, but was underpopulated by almost 19,000 persons. HD34 was 58.2% HCVAP using 2005-2009 ACS data and was 24,000 underpopulated. Instead of trying to maintain HD33 in addition to HD34, HD33 was eliminated. As noted, redistricters believed they could offset the elimination of HD33 32 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 33 of 171 Defendants contend that the reduction in population combined with the County Line Rule eliminated the minority district in Nueces County, and the delegation proposed the new districts. They contend that the configuration reflects only partisan and neutral districting principles and does not violate the VRA. The configuration of Nueces County remains unchanged in the current plan. The § 2 results claim in Nueces County presents two questions: (1) assuming only two districts are required within Nueces County, did plaintiffs prove a § 2 violation by the Legislature’s failure to draw both as Latino opportunity districts? and (2) have Plaintiffs proven that the VRA requires more than two districts to be drawn in Nueces County (i.e., a County Line rule violation)? Defendants argue that Plaintiffs fail to prove a § 2 violation because Nueces County was only entitled to two districts and, “[a]ccording to the only population data available at the time the Texas House map was drawn, Nueces County’s total SSVR percentage was 49.5%” such that creating two SSVR-majority districts within Nueces County was impossible. Docket no. 411 at 42-43. It is true that, at the time the first drafts of the map were created, the only current data mapdrawers had available was SSVR. Because Nueces County was less than 50% SSVR, it was mathematically impossible to draw two SSVR-majority districts within the County. However, mapdrawers knew that SSVR was typically lower than HCVAP, that Nueces County had been over 50% HCVAP in for § 5 purposes by increasing the SSVR of HD90 and HD148, both of which were already Latino ability districts. As noted by the D.C. Court in the § 5 preclearance litigation, “section 5 allows a state to dismantle an ability district as long as it offsets that loss by drawing a new ability district elsewhere.” Texas v. United States, 887 F. Supp. 2d 133, 144 (D.D.C. 2012). However, as will be discussed, this use of race was improper because HD90 and HD148 were already ability districts, and thus Texas did not draw any “new” ability districts. In addition, § 2 of the VRA does not permit offsets. As discussed in the congressional plan opinion, a state may “use one majority-minority district to compensate for the absence of another only when the racial group in each area had a § 2 right and both could not be accommodated.” Docket no. 1339 at 47 (citing LULAC v. Perry, 548 U.S. 399, 429 (2006). “Simply put, the State’s creation of an opportunity district for those without a § 2 right offers no excuse for its failure to provide an opportunity district for those with a § 2 right.” Id. (quoting LULAC, 548 U.S. at 430). Because voters in HD90 and HD148 were already successfully electing their candidates of choice, those were existing ability districts for § 5 purposes (as found by the D.C. Court) and those voters did not have a § 2 right to be placed into a majority-HCVAP or SSVR district that could “offset” the failure to create or maintain opportunity districts elsewhere in the State if required by § 2. 33 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 34 of 171 2000, and that Hispanic population had increased while Anglo population had declined, meaning they had every reason to believe that the HCVAP of Nueces County was over 50% even without more recent HCVAP data. Moreover, Defendants’ assertion that HCVAP data was not available is simply untrue. The HCVAP data was available to mapdrawers by April 21, 2011, before the map went to the floor, and would have shown that Nueces County HCVAP was over 50%. Interiano stated at trial that he knew that the HCVAP of Nueces County exceeded 50%, Tr1463, and he also would have been aware that HCVAP was the appropriate measure for determining whether § 2 required a Latino opportunity district. E.g., US-76 (Opiela November 20, 2010 email, with Interiano copied, discussing that they needed to evaluate districts in terms of VAP and CVAP and stating that Fifth Circuit law clearly established CVAP as the standard). Downton testified repeatedly that he used a 50% HCVAP standard for measuring Latino opportunity districts. Nevertheless, Interiano as primary mapdrawer and other redistricting leaders continued to focus solely on SSVR for Latino districts in the Texas House plan, and never explored whether two HCVAP-majority districts could have been drawn wholly within Nueces County. TrA64-65 (Interiano). In their briefing, the Task Force Plaintiffs assert that because Nueces County was over 50% HCVAP, two HCVAP-majority Latino opportunity districts could have been drawn wholly within the County.23 However, while it is certainly theoretically and mathematically possible to draw two HCVAP-majority districts wholly within Nueces County, no plaintiff has provided such a proposed map or shown that the Latino populations in such districts would be compact, taking into account 23 “The citizen voting age population of Nueces County is 54.6% Hispanic, thus allowing for two districts wholly contained within the county to be majority HCVAP.” Docket no. 1282 at 51 (citing docket no. 1274 FF742); docket no. 1296 at 32. 34 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 35 of 171 traditional redistricting principles. Thus, Plaintiffs fail to satisfy their burden under Gingles of showing that two Latino opportunity districts could be drawn wholly within Nueces County (i.e., without violating the County Line Rule). Plaintiffs have proffered demonstration maps that would maintain two Latino-majority districts in Nueces County by breaking the County Line Rule. Because Solomons decided never to break the County Line Rule to comply with the VRA, mapdrawers did not explore whether to include more than two districts in Nueces County or whether doing so would have been required by § 2 or § 5 of the VRA, despite Hanna’s advice to investigate this issue. This Court has found, and Defendants do not dispute, that both Nueces County-based benchmark districts HD33 and HD34 were majority-Hispanic, majority-HCVAP, and majority-SSVR Latino opportunity districts, even though they did not elect the Latino candidates of choice in 2010. Although Defendants have argued that this was an agreed-upon delegation map, see e.g., docket no. 1249-1 at 30, the evidence is clear that mapdrawers and redistricting leadership made the decision to eliminate HD33 and to create only one Latino opportunity district within Nueces County. Plaintiffs’ demonstration maps would purportedly maintain two Latino opportunity districts, but they subordinate the County Line Rule in order to do so. Defendants’ objections to the maps include the violation of the County Line Rule and racial gerrymandering. The Court finds that consideration of this issue is premature—the Court cannot hold that the Legislature was required to break the County Line Rule to maintain two Latino opportunity districts in Nueces County without Plaintiffs first showing that it was reasonably necessary to comply with § 2 (i.e., that two Latino opportunity districts could not be maintained consistent with the County Line Rule). Because, as 35 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 36 of 171 explained above, Plaintiffs have failed to make that showing, this issue is not ripe for consideration.24 Thus, with regard to Nueces County, Plaintiffs have failed to demonstrate a violation of § 2’s results test at this time. However, they may continue to pursue this claim with regard to Plan H358 in the 2013 plan trial because the Nueces County configuration remains unchanged from Plan H283. In addition to the § 2 results claims based on the failure to maintain two Latino opportunity districts in Nueces County, Plaintiffs contend that the decision to eliminate an existing Latino opportunity district in Nueces County was intentional vote dilution. Plaintiffs also argue that the district lines for those districts that remained in Nueces County are evidence of racial gerrymandering and intentional vote dilution. See, e.g., docket no. 1279 at 49-50. The United States notes that Plan H283 includes “a convoluted line that ensures Anglo control over HD32,” that HD34 “was packed with a greater share of Spanish surnamed registered voters than any Nueces County district prior to the 2011 redistricting, leaving HD32 with a substantial majority of the Anglo voters in Nueces County,” and HD34 was overpopulated compared to HD32. Docket no. 1279 at 50-51. The United States also contends that the “protuberance from HD34 . . . ensured that no potential Hispanic candidate with recent legislative experience—including the Republican incumbent of HD33, the former representative of HD33, and the former representative of HD34—would reside in the new HD32 to challenge the Anglo incumbent,” and the State’s witnesses provided no plausible alternative explanation for the convoluted boundary. Docket no. 1279 at 51. The Task Force also contends that the bizarre extensions along the border of HD32 and HD34 are evidence of racial 24 In the interim plan Order, this Court also focused on SSVR and the fact that Plaintiffs had not presented a map creating two Latino districts without a county line cut, finding that “the only way to maintain two Latino districts in Nueces County is to cut a county line.” Docket no. 690 at 8. The interim plan order was based only on preliminary findings and analysis, and Plaintiffs at that time had not argued that two HCVAP-majority districts could be drawn wholly within Nueces County. Further, the interim plan order did not rule out a conclusion that the VRA could require a county line cut. Id. 36 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 37 of 171 gerrymandering. Docket no. 1282 at 64. The Court agrees that the facts and evidence concerning the decision to eliminate HD33 in Nueces County demonstrate intentional vote dilution. Specifically, the Court relies on (1) the fact that redistricting leadership eliminated an existing Latino opportunity district without considering whether two Latino opportunity districts were required by § 2 of the VRA under the appropriate standard (HCVAP), (2) the fact that they decided to eliminate ability district HD33 without replacing it elsewhere but instead raised the SSVR of two existing ability districts to claim an “offset” under § 5 of the VRA and attempt to ward off § 2 challenges in bad faith,25 and (3) the ultimate configuration of the remaining Nueces County districts as demonstrating racial gerrymandering. Mapdrawers intentionally did not consider whether § 2 required two districts in Nueces County, and used a false reliance on SSVR and the County Line Rule to justify their refusal to maintain two Latino opportunity districts in Nueces County. Further, they knew that HCVAP was the proper measure for a Latino opportunity district and had HCVAP data available before the map went to the floor, but insisted on using SSVR as the measure because SSVR was lower (and lower than 50%), and it allowed them to argue that mathematically, two SSVR-majority districts could not be drawn. Although Nueces County was majority HCVAP and had been since at least 2000, they refused to consider whether two HCVAP-majority districts could be drawn wholly within the County. In addition, mapdrawers and redistricting leadership steadfastly refused to consider whether the County Line Rule might have to yield to maintain two Latino opportunity districts in the area, 25 As noted, there is no “offset” under § 2 of the VRA, but a traditional § 2 results claim requires a plaintiff to show that more HCVAP-majority districts could have been drawn, and thus redistricters attempted in bad faith to try to preclude such a showing by artificially increasing the number of HCVAP-majority districts in the plan (e.g., they used the creation of HCVAP-majority HD148 to argue that no additional Latino opportunity districts were required in Harris County). 37 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 38 of 171 despite Hanna’s advice to the contrary. The Court finds that their refusal to consider the proper requirements of § 2 and to attempt to comply with it was in bad faith and intended to limit Latino opportunity. Rather than attempting to comply with § 2 by maintaining two Latino opportunity districts, mapdrawers and redistricting leadership attempted to use § 5 as a shield for their failure to attempt compliance with § 2 and for eliminating an existing Latino opportunity district. Mapdrawers and redistricting leadership asserted that, under § 5, they could “offset” the loss of HD33 as an ability district by creating a new 50% SSVR district elsewhere in the State. Mapdrawers chose two districts that were already performing reliably for Latino voters despite being under 50% SSVR (HD90 and HD148) and raised their SSVR over 50%, claiming that these “new” Latino districts could offset the loss of HD33 and avoid statewide retrogression for preclearance purposes. However, as the D.C. Court found (and as redistricters knew), these were already ability districts, and thus increasing their SSVR did not create any “new” ability district to offset the loss of HD33. Mapdrawers and redistricting leadership relied in bad faith on a 50% SSVR standard as the definition of “ability district” despite clear DOJ guidance to the contrary so that they could claim compliance with § 5 while simultaneously ensuring that no new Latino districts were created to offset the loss of HD33. At the same time, increasing the SSVR in HD90 and HD148 increased the number of HCVAP-majority districts in the plan by at least one (HD148) and arguably two (using 2005-2009 ACS data, the HCVAP was of HD90 was 49.7 +/- 2%). Because § 2 results claims require plaintiffs to show that more HCVAP-majority districts could have been drawn, mapdrawers acted in bad faith to try to thwart such claims by artificially inflating the number of HCVAP-majority districts—creating them where there was no § 2 right because the districts were already performing 38 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 39 of 171 for Latinos. Redistricting leadership ignored warnings from Nina Perales in her April 27 letter that the HD33 loss was not offset by artificially increasing the SSVR in HD90 and HD148 and that additional Latino opportunity districts were required. They chose to increase the SSVR in districts that were already performing for Latinos so that they could claim VRA compliance and minimize the number of Latino opportunity districts without losing or jeopardizing any Republican incumbent seats—that is why they did not raise the SSVR in El Paso County’s HD78, which elected a Republican in 2010. Mapdrawers’ use of race to increase the SSVR in HD90 and HD148 in the name of VRA compliance turned the VRA on its head—instead of using race to provide equal electoral opportunity, they intentionally used it to undermine Latino voting opportunity. Further, there is evidence that the mapdrawers (including specifically Rep. Hunter) racially gerrymandered the districts that remained in Nueces County to further undermine Latino voting strength. There are ten precinct splits along the HD32/HD34 border, see US-387, indicating that mapdrawers were likely using race to assign population since accurate political data is not available below the precinct level. Although mapdrawers claimed that they needed to make HD32 “performing,” there is no evidence that they needed to put as many Hispanic voters in HD32 as they did, either to comply with § 2 or § 5. Thus, mapdrawers intentionally packed Hispanic voters into HD32 to minimize their number and influence in HD34 and protect Hunter. As discussed in the fact findings, there is also evidence that they targeted low turnout minority areas for inclusion in HD34 and intentionally drew out potential Hispanic rivals (both Republican and Democrat), again to protect the Anglo incumbent. And, as discussed below, there is evidence that Hunter intentionally overpopulated HD32 (and underpopulated his own HD34), without a legitimate justification for 39 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 40 of 171 doing so. In sum, Plaintiffs’ § 2 results claims are not moot, but Plaintiffs failed to prove them with regard to Plan H283. Because this configuration remains unchanged in the current plan, Plaintiffs may bring their results claims with regard to the 2013 plan. However, the Court finds that redistricters intentionally diluted Latino voting strength by eliminating HD33 in Nueces County. Mapdrawers’ use of race in HD90 and HD148 to “offset” this loss was not to comply with the VRA but to intentionally dilute Latino voting strength by (1) allowing redistricting leadership to claim that they were complying with the VRAdespite eliminating HD33 and creating no new opportunity or ability districts; (2) thwarting arguments that § 2 might require additional opportunity districts by artificially inflating the number of Latino-majority districts in the plan; and (3) racially gerrymandering and utilizing population deviations to further dilute the Latino vote in the remaining two Nueces County districts. As discussed in the congressional plan opinion, there is legally significant racially polarized voting in Nueces County. Given the existence of racially polarized voting and the history of discrimination and its legacy in Nueces County, and considering the totality of circumstances, the elimination of an existing Latino opportunity district and ensuing racial gerrymandering in Nueces County is intentional vote dilution in violation of § 2 of the VRA and the Fourteenth Amendment. D. HD35 HD35 was and remains a district composed of whole counties, though the county configuration was altered in Plan H283 and all Hispanic population metrics were decreased from the benchmark configuration. The United States contends that HD35 was a Latino opportunity district in Plan H100 and is no longer in Plan H283, and this is the result of intentional vote dilution. 40 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 41 of 171 Docket no. 1279 at 44-45. The Court agrees that benchmark HD35 was a Latino opportunity district, though it elected a Hispanic Republican in 2010, who was not the Hispanic candidate of choice. As drawn pursuant to a court remedy in 2001, it had 51.5% SSVR. Over the decade, it elected the Hispanic-preferred candidate four out of five times, for an endogenous election index score of 80%. Its exogenous election index score was 40%. By the time of redistricting, the SSVR of the district had risen to 55.4%, yet Plan H283 reduced it to 52.7% through its new configuration of whole counties. Its exogenous election score dropped to 20%. Because of the SSVR reduction, Hanna advised mapdrawers to perform election analysis to measure its performance, and Interiano and mapdrawers knew from the OAG election analysis that the performance for the Hispanic candidate of choice decreased. Dr. Arrington testified that benchmark HD35 was an effective Latino opportunity district, but that it was no longer effective in Plan H283. TrJ119, TrJ134-35.26 During the session, MALDEF offered Plan H115 that increased the SSVR of HD35 to 58.2%, but it broke the County Line Rule. In addition, Rep. King (an Anglo Democrat) proposed an amendment (Plan H161) that would have changed the county configuration, but the incumbent Aliseda objected. He stated that the amendment increased his SSVR “a little bit more” (he stated it was less than .02%), but asserted that was not the reason he objected; rather, he asserted, it was because it made his district “into a more democrat district.” D-13 at S183-84. Rep. Coleman offered Plan H232, which he stated “strengthened existing Latino opportunity district HD35” and made it 26 Dr. Engstrom, the Task Force’s expert, found that enacted HD35 did not provide a reasonable opportunity for Latinos because they did not win a majority of votes cast in 7 general and 6 Democrat primary elections with Latino candidates more often than not. Engstrom Corr. Rebuttal Report (docket no. 307-1) at 28. However, the Task Force also takes the position that benchmark HD35 was not a Latino opportunity district. Defendants’ expert Dr. Alford found that Democrats went from 28/48 (58%) wins to 23/48 (48%) wins in HD35. US-350 at 7 & Table 2. 41 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 42 of 171 more likely to perform. Martin opined that Plan H232 strengthened HD35 while preserving other Latino districts in the region and complying with the County Line Rule. Joint Expert Ex. E-5 (Martin Report) at 13-14. According to the D.C. Court’s opinion in the § 5 preclearance litigation, the parties agreed that HD35 in Plan H283 was not an ability district, but they disagreed on whether it was an ability district in the benchmark and thus whether there was retrogression. Texas v. United States, 887 F. Supp. 2d 133, 167 (D.D.C. 2012) (vacated on other grounds). The D.C. Court found that it had been an ability district because it had a high endogenous success rate (80%) despite lower exogenous index scores. The D.C. Court concluded that it no longer was an ability district in Plan H283 based on the “low exogenous election results for the enacted district combined with HCVAP changes that push the district even closer to the majority line.” Id. at 168. It noted that this particular district presented “a close and very difficult case” and that it might have reached a different conclusion if presented “with more or different evidence,” but that the State had not met its burden of proof. Id. at 168 n.37. This Court finds that HD35 remains an opportunity district in Plan H283. The evidence shows that benchmark HD35 performed well for Latinos despite weak exogenous election scores. Some changes to the district were required by the fact that it was underpopulated by 9.4%. Although Handley’s index and the OAG 10 showed a one election decrease in Latino performance on the exogenous index between the benchmark and enacted districts, Dr. Engstrom’s analysis actually showed an increase in Latino performance from 2-5 to 3-4. Plaintiffs have failed to meet their burden of proof of demonstrating that HD35 is no longer an opportunity district. In addition, as discussed in the fact findings, there is no evidence to support the United States’ contention that 42 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 43 of 171 mapdrawers manipulated turnout in this district. Accordingly, the Court finds that this district does not support a claim of intentional vote dilution or racial gerrymandering. E. HD41 Several Plaintiffs, including MALC, Perez Plaintiffs,27 and the United States, challenge HD41 as the product of racial discrimination and intentional vote dilution. Defendants contend it is merely the product of partisan gerrymandering to protect newly-Republican Aaron Peña. The United States argues that HD41 was changed so significantly and contains so many split precincts that its continued status as an opportunity district is in question. Although its HCVAP and SSVR percentages decreased by around 5%, they remained high at 72.1% HCVAP and 63% total SSVR. The D.C. Court concluded that HD41 remained a Hispanic ability district in Plan H283 because the court applied a 65% HCVAP ability presumption that was not disproved. Texas, 887 F. Supp. 2d at 169-70. Further, Hidalgo County was over 85% HCVAP. Dr. Engstrom concluded that HD41 in Plan H283 remained a Latino opportunity district. The Court finds that HD41 remains an opportunity district in Plan H283. Nevertheless, the Court finds that HD41 was drawn in part with racially discriminatory (dilutive) motive. The United States argues that the precinct splits in the district cannot be explained by partisan motivation and were racially discriminatory. Docket no. 1279 at 45. Dr. Arrington looked at all splits in HD41 and found that they made the district more Anglo. However, the most 27 Perez Plaintiffs’ proposed conclusions of law (docket no. 1265) assert intentional vote dilution with regard to HD41, but the district is not mentioned in their 2014 Post-Trial Brief (docket no. 1263). Their 2011 Post-Trial Brief argues that the population deviations in the district violate the Equal Protection Clause and that the Valley configuration is less consistent with the County Line Rule as construed by Clements v. Valles than MALC’s and other Plaintiffs’ proposals creating a Cameron+Hidalgo County district. Docket no. 401 at 14-15. 43 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 44 of 171 relevant splits are those made after Plan H11328 that actually involved some population. As discussed in the fact findings, some of Downton’s precincts splits after Plan H113 were consistent with Anglo maximization to increase the Anglo VAP of the district, and they were not consistent with Republican maximization. Although mapdrawers (including Peña) testified that the changes were made to “maximize” the district for Republican performance, election analysis results showed that the changes resulted in a decline in Republican performance, with corresponding increases in Democratic performance. See US-517 (Red-225 Report for Plan H113); D-109 (Plan H283 reports). Thus, the explanation that the numerous precinct splits were made to “maximize” the district for Republican performance is not credible. Nor was other testimony offered by mapdrawers that the numerous precinct splits were made simply “to follow roads.” As noted in the fact findings, the district HVAP and HCVAP population did not change despite the substantial manipulation and precinct changes that occurred on April 17. However, although the net effect of all the changes made between Plan H113 and Plan H283 was not to make the district more Anglo and less Hispanic, that does not negate the fact that some particular precinct splits were made with a racially discriminatory motive. In Plan H113, Peña’s district was significantly (4.92%) below ideal population, while the neighboring districts were all overpopulated. Downton recognized this as a potential problem and worked with Peña to increase the population of the district while “maximizing” it for Peña’s benefit.29 However, because the area is heavily Hispanic and the original configuration of the district already contained the most Republican areas, 28 Peña’s district had four precinct splits in Plan H113, but they were adequately explained by non-racial reasons. 29 As discussed in the section on the one person, one vote claims, HD41 was nevertheless left intentionally underpopulated because adding more persons would have meant adding more Hispanics. 44 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 45 of 171 Downton had to work to increase the population of the district while simultaneously maintaining the Anglo and Hispanic populations in an attempt to protect Peña. Downton increased the population of the district by 857 persons (he admitted to not increasing the population any more because it would make the district less favorable to Peña), but kept its racial population metrics basically the same, in part by using precinct splits in a racially discriminatory manner. Had he not utilized the precinct splits in such a manner, the effect of increasing the population would have been an increase in Hispanic population in the district, making it more likely to perform for Hispanic voters (and redistricters presumed, less likely to re-elect Peña). Thus, although all of the precinct splits together did not negatively impact the Hispanic population of the district, the racially discriminatory precinct splits were used to offset the racially favorable splits. In addition to manipulating total population, the SSVR of HD41 is significantly lower (63%/64.6%) than the other Hidalgo County districts, which are 85.8%/86.4% (HD40), 82.2%/83.1% (HD39), 85.1%/85.8% (HD36), and 90.8%/91.3% (HD31), and is lower than any of those benchmark district’s SSVR levels, again demonstrating an intent to draw the district with fewer Hispanic voters. The SSVR of Peña’s district was reduced, while the SSVR of three of the other districts was increased, and at the same time mapdrawers knew that the performance of HD41 for Latino voters decreased from 7/10 to 5/10 on the OAG 10. Although Solomons repeatedly touted the map as “member driven,” the Hidalgo County configuration was not drawn with input from the other members of the Valley delegation and disregarded member-constituent relations in order to gerrymander the district, both politically and racially. The Court thus finds that mapdrawers intentionally used race to draw the district to perform 45 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 46 of 171 less favorably for Latinos. They also intentionally manipulated the SSVR and total population to dilute the Latino vote in HD41 in order to protect an incumbent who they believed would no longer be the Latino candidate of choice given his decision to switch parties. Given the existence of racially polarized voting in the Valley (as found by Dr. Engstrom), the lingering effects of past discrimination in terms of socio-economic factors, and the totality of circumstances, the configuration of HD41 is racially discriminatory and constitutes intentional vote dilution in violation of § 2 and the Fourteenth Amendment. Although the individual members of the Legislature may not have been aware of this when they voted to pass Plan H283, Downton was given the authority to make these changes to the map and his intent must be imputed to the Legislature as a whole. F. Cameron+Hidalgo/Rio Grande Valley configuration The benchmark plan contained six Latino opportunity districts wholly within Hidalgo and Cameron Counties—four wholly within Hidalgo County and two wholly within Cameron County. The excess population in Cameron County was joined with counties to the north in HD43, also a Latino opportunity district. Plan H283 also creates six HCVAP-majority districts30 wholly within Cameron and Hidalgo Counties (four within Hidalgo County and two within Cameron County); the surplus of Cameron County is joined into HD43 to the north, and the surplus of Hidalgo County is joined into HD31 to the west; both HD31 and HD43 are Latino opportunity districts. The Task Force and MALC assert that § 2 required the State to create an additional district in the Rio Grande Valley by combining the surplus population of Cameron and Hidalgo Counties, creating seven Valley-based districts, and that this would actually be more consistent with the 30 As discussed, the continued status of HD35 and HD41 as Latino opportunity districts is challenged by the United States. The Task Force regards HD41 in Plan H283 as a Latino opportunity district. Docket no. 1282 at 6 n.4. The Court has found that HD41 remains a Latino opportunity district in Plan H283. 46 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 47 of 171 County Line Rule than the State’s configuration. This was done in the Court’s interim map (Plan H309) such that seven districts are wholly contained within Hidalgo and Cameron Counties, with no spillover. This is also done in MALC’s Plan H201 through the creation of HD144 and MALC’s Plan H295 through the creation of HD35. Additional proposed plans create a seventh Valley district but also include spillover in another district, including Task Force Plan H292 (HD32) and MALC Plan H205 (HD72).31 In their 2011 briefing, Defendants contended that the “undisputed evidence at trial showed . . . that creating a new district in Hidalgo and Cameron would have required the Legislature to violate the county-line rule in other parts of the State.” Docket no. 411 at 44; docket no. 457 at 2930; see also TrA78 (Interiano) (“I tried to do it, and I told David Hanna that I had not found a way to put the population together without causing a county line split further up the road.”). This assertion has since been disproved, as shown in the Court’s interim plan H309. However, it is true that all maps presented to the mapdrawers during the session that created this district also had a county line break around Nueces County. Downton testified that it was permissible to create the Cameron+Hidalgo County district, but it was not legally required, and redistricters made a policy decision not to create it. TrJ2094-95. Defendants argue that no voter in that area suffered a § 2 injury from the State’s chosen configuration because all Cameron and Hidalgo County voters are in Latino opportunity districts 31 Both of these plans include 25 districts within Harris County, which affects districts outside of drop-in counties. Further, as noted, the NAACP also included such a district (HD72) in their demonstration Plan H202, but did not brief it as a required § 2 district. The Hidalgo County configuration of Plan H202 is identical to that of MALC’s H205, but Cameron County differs. 47 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 48 of 171 under either configuration. Docket no. 1272 at 102.32 The Court agrees that if the Legislature created the required number of § 2 districts and substantially addressed the § 2 violation, they had discretion in determining the configuration of the districts. However, Plaintiffs are arguing that creation of the Cameron+Hidalgo district was more consistent with the County Line Rule, recognized the Latino growth in the Valley, and allowed for the creation of an additional Latino opportunity district over the number created in Plan H283, which could establish a § 2 violation. Thus, for Plaintiffs to succeed on their § 2 results claim, they must demonstrate that creating this additional Hidalgo+Cameron district creates more Latino opportunity districts in South Texas (or statewide) than does Plan H283. Although Plan H201 creates an additional Valley Latino opportunity district over Plan H283 by adding HD144 and retaining HD33 by utilizing a county line split, HD35 is no longer a Latino opportunity district, whereas it is in Plan H283. Plan H283 has 13 South and West Texas Latino opportunity districts (not including El Paso but including Nueces County). Plan H201 has 14, but only because it maintains HD33 in Nueces County district. Thus, it is really the retention of HD33 as a Latino opportunity district that results in an additional HCVAP-majority district over the enacted plan. This issue therefore boils down to the Nueces County issue, and the Court will not consider proposed maps that split Nueces County to create an additional district until it is shown that two Latino opportunity districts cannot be drawn wholly 32 The Task Force responds that Defendants miss “the critical difference between the voters of Cameron and Hidalgo Counties living in six Latino opportunity districts versus living in seven Latino opportunity districts.” Docket no. 1296 at 36 (emphasis in original). To the extent the Task Force is thus arguing that § 2 required the State to draw seven districts wholly within Cameron and Hidalgo Counties instead of six wholly within and two partially within simply so that the voters of Hidalgo and Cameron Counties could have seven districts to themselves, the Task Force fails to support this assertion legally. 48 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 49 of 171 within Nueces County.33 The Task Force contends that Plan H292 adds HD33 in Nueces County and HD35 in the Valley compared to Plan H283. But it does not count HD35 in Plan H283 as an opportunity district. Counting HD35 in Plan H283 as a Latino opportunity district, the only additional district created is in Nueces County.34 Again that simply boils down to a Nueces County issue. Plan H292 also has 25 districts in Harris County, which has ripple effects in all the non-drop-in county districts statewide because there is one less district outside the drop-in counties. MALC Plan H295 creates seven districts wholly within Hidalgo and Cameron Counties by moving HD35 to the Valley. Again, however, it does not appear to create more Latino opportunity districts other than the Nueces County district. Plan H205 purports to create an additional district over Plan H283. It also splits Nueces County to create a Latino opportunity district, and it includes 25 districts in Harris County. Further, Plan H205 substantially disregards the County Line Rule—Defendants argue that it violates the County Line Rule at least 25 times. Docket no. 457 at 29-30; docket no. 645 at 14-15. Plaintiffs must show that these County Line Rule violations were reasonably necessary to comply with § 2 and, given the Court’s conclusion below that the proposed West Texas districts are not compelled by the 33 Similarly, eleven-district “plug in” plan H294 does not create an additional HCVAP-majority district; it just reconfigures the districts. 34 The Task Force uses a combination of demographics and Dr. Engstrom’s 13-election analysis to classify a district as a Latino opportunity district or not. Using the Task Force’s classifications, Plan H283 has 12 districts in South/West Texas (excluding El Paso) and Plan H292 has 14. Plan H283 includes: 31, 34, 36, 37, 38, 39, 40, 41, 42, 43, 74, & 80, while Plan H292 has 14, including those same districts plus 32, 33, 35, but not 80, for an increase of two districts. However, this Court does not agree with the Task Force’s classification of all opportunity districts, and for the reasons explained in the Congressional map opinion, has rejected a 50% success test for determining opportunity. The Court has concluded that HD35 in Plan H283 is a Latino opportunity district, and as such the increase in Plan H292 is only one district—HD33 in Nueces County. It may be that HD80 (with a HCVAP of 55.9%) in Plan H292 is also a Latino opportunity district, but the Task Force has not made such a claim (in fact it claims it is not one) nor provided evidence or briefing to support it. 49 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 50 of 171 VRA, resulting changes to that portion of the map will also affect non drop-in county areas of the State, making an analysis of the number of Latino opportunity districts created in South Texas unreliable. Given these issues and the fact that the current plan includes the Hidalgo+Cameron district but not HD35, the Court finds that this results claim is best resolved with regard to the 2013 plan in effect. However, certain Plaintiffs (including the United States and the Task Force Plaintiffs), argue that the State’s “deliberate decision not to recognize explosive Hispanic growth in the Rio Grande Valley through the creation of an additional Hispanic opportunity district constitutes intentional vote dilution.” Docket no. 1279 at 39. The Court finds that Plaintiffs have failed to prove an intentional vote dilution claim based on the failure to draw the Cameron+Hidalgo district. As noted, mapdrawers were not presented with a plan that clearly created an additional Latino opportunity district compared to what they had drawn, and they had discretion in choosing how to configure the Hidalgo and Cameron County districts to address § 2. G. Harris County In addition to their one person, one vote claims based on population deviations (addressed below), Plaintiffs assert §2 results and intentional vote dilution claims in Harris County. Plaintiffs primarily complain that Defendants did not create an additional minority opportunity district in Harris County despite the minority population growth, but also contend that the Legislature’s decision to reduce the number of districts in Harris County to 24 and to eliminate HD149, which Plaintiffs contend was a tri-ethnic coalition district, while protecting Anglo districts with slower population growth, were racially discriminatory actions designed to dilute the minority vote. The Task Force Plaintiffs argue that an additional Latino opportunity district could have been 50 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 51 of 171 drawn in Harris County. MALC offers demonstration district HD144 in Plan H205 and Plan H295 (though Plan H205 includes 25 districts in Harris County).35 The NAACP also complains about the reduction of minority voting strength in HD144, which it argues had naturally become viable for minorities given the population growth, and argues that an additional minority coalition district could have been drawn (by maintaining both HD137 and HD149), as shown in Plan H202 (which has 24 districts in Harris County). Docket no. 625 at 5-7; docket no. 1280 at 37. Defendants admit that “Plan H283 did not create additional majority-Latino or majority African American districts in Harris County,” though they assert that the SSVR and HCVAP of HD148 were increased over 50%. Docket no. 413 FF62.36 They argue that the reduction of districts in Harris County was not discriminatory but “resulted from basic arithmetic and the combined legal requirements of the Texas and United States Constitutions.” Docket no. 457 at 69. They also assert that the Democrat members never offered a proposed map such that “Downton worked from a proposal supplied by the Republican members of the delegation” and “then accommodated Democratic changes to the extent possible.” Id. Defendants argue that the decision to apportion 24 seats in Harris County “did not prevent the creation of an additional minority-majority district” and 35 Defendants argue that Plan H205 creates the same number of Latino opportunity districts (4) in Harris County (HD140, HD143, HD144, and HD145) as Plan H283 (HD140, HD143, HD145, and HD148). Docket no. 457 at 30-31. Defendants contend that Plan H205 also “retrogresses” HD148 by reducing its SSVR from 40% to 35.7%, whereas Plan H283 “strengthens Latino make-up of this district by increasing the SSVR to 50.0%.” Id. at 31 n.16. 36 Defendants assert that, “at the request of MALDEF, H283 increased SSVR and HCVAP to over 50% in Harris County House District 148.” Docket no. 413 FF62. Using their demographic criteria for opportunity districts, Defendants argue that Plan H283 creates nine minority opportunity districts, and that none of the proposed 25-district plans creates more than nine (they say Plan H292 creates 9), while other plans create fewer. Docket no. 411 at 45; docket no. 413 FF85-88. However, to the extent the Legislature increased the SSVR of HD148 in order to count it as an ability or opportunity district for purposes of the VRA, it created a § 2 remedy where there was no § 2 liability because it is undisputed that HD148 routinely elected the Latino candidate of choice. Thus, the Legislature’s choice to artificially inflate the SSVR (and with it the HCVAP) of HD148 does not preclude Plaintiffs from arguing that a different Latino opportunity district was required in Harris County. 51 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 52 of 171 that “Plaintiffs have not demonstrated that the addition of another district in Harris County would have resulted in an additional minority opportunity district.” Docket no. 411 at 45. Defendants also contend that neither HD137 nor HD149 was a protected minority coalition district and that the decision to eliminate them was partisan, not racial. Id. at 45-46. In this Court’s interim plan H309, the Court found that Plaintiffs had demonstrated a likelihood of success on the merits of the § 2 claim in eastern Harris County, specifically that Plaintiffs had presented numerous demonstration plans illustrating that an additional compact majority-HCVAP district was possible in eastern Harris County, and made a preliminary finding that creation of a new Latino opportunity district was justified by the totality of circumstances. Docket no. 690 at 8-9. The Court’s interim plan thus reconfigured HD144 in the manner requested by Plaintiffs to have a majority HCVAP. Id. The Court also maintained both HD137 and HD149, finding that the § 5 claims were not insubstantial, and making no other claims determinations. Id. This configuration remains in the plan adopted by the Legislature and currently in effect (Plan H358). Given that the current plan includes a district similar to MALC’s proposed HD144 and maintains both HD137 and HD149 with only 24 districts, the results claim is moot. However, Plaintiffs claim that the decision to reduce Harris County to 24 districts, the elimination of HD149, and the deliberate failure to draw any new opportunity districts in Harris County are evidence of intentional discrimination.37 MALC contends that redistricting leadership had discretion to choose either 24 or 25 districts, yet chose 24 to avoid drawing an additional Latino opportunity district in Harris County. 37 The Perez Plaintiffs cite Harris County as an example of intentional fragmentation (docket no. 1263 at 5), but there is no briefing on this issue in their 2014 Post-Trial Brief. Their proposed conclusions of law (docket no. 1265) refer to HD132, HD137, HD145, HD148, and HD149 as districts that evidence discrimination. 52 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 53 of 171 Docket no. 459 at 17.38 Other Plaintiffs also complain that the decision to decrease the delegation to 24 was intentionally racially discriminatory because it would eliminate a minority district. The Court finds that the mapdrawers’ decision to include only 24 districts in Harris County was reasonable, given the language of the Texas Constitution and simple math, and was not intentional vote dilution. The Court notes that Hanna recommended rounding down to 24 districts initially, even when he believed this would mean the loss of a Republican seat. D-192. He noted that 24 more closely tracked the constitutional language, and that the risk of choosing incorrectly was “catastrophic” in terms of affecting all districts outside of the drop-in counties. D-135. The Court is not convinced that past practice of rounding up indicates pretext in this case, given that different decisionmakers were involved in those decisions (different legislature and the LRB), and that Democratic and minority members had voted to round down in the prior legislative session when presented with similar numbers. The Court does not find any evidence to support the assertion that the decision to round down to 24 was racially discriminatory. The NAACP Plaintiffs argue that Texas ran afoul of the Supreme Court’s warning in Bartlett v. Strickland by intentionally eliminating a performing minority district (HD149)39 in violation of the Fourteenth Amendment. Docket no. 1294 at 14. They contend that Interiano admitted dissolving HD149 despite knowing it was a district in which a diverse group of minority voters elected the candidate of their choice, Hubert Vo, because they did not think the VRA required them to maintain 38 MALC points to its Plan H205, which it asserts “maintains the current Latino opportunity districts, but also enhances the Latino population in district 144 so that it goes from an emerging Latino district to a true Latino opportunity district.” Id. at 17-18; docket no. 1185 at 13 n.29. 39 The NAACP argues that HD149 “was a compact, naturally-occurring multi-ethnic coalition district whose voters had a proven track record of being politically cohesive and electing their candidate of choice, Rep. Vo.” Docket no. 1280 at 37. They note that HD149 could have been preserved even with a 24-district configuration, as shown in their Plan H202, and without diminishing HD137 Docket no. 1280 at 37-38 (Joint Map Ex. J-25). 53 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 54 of 171 it. Docket no. 1280 at 37 (citing Tr1482 (Interiano)). They argue that “[t]his callous disregard for proven voting rights gains from an extant cohesive minority population is certainly evidence of an intent to discriminate . . . .” Id. The Court need not decide whether HD149 was a protected coalition district because the evidence does not sufficiently support Plaintiffs’ claim that it was targeted for removal because it was a minority opportunity district or because its elected representative was a minority candidate of choice, rather than simply because it was electing a Democrat. Mapdrawers and redistricting leadership did not believe it to be a protected minority opportunity district and it was chosen for removal because one district was being eliminated and Republicans saw an opportunity to pair two Democrats, thereby increasing Republican seats and protecting all other incumbents.40 The pairing of two Democrats was consistent with Solomons’ decision to always pair two members of the same political party to theoretically give each incumbent a chance to be re-elected in the general election. Plaintiffs further contend that the enacted 24-district configuration in Harris County intentionally diluted minority voting strength. The Task Force complains that no new Latino opportunity district was created and that instead Latino voters were packed into other districts to minimize their electoral influence. The Task Force Plaintiffs note that Hanna’s April 12 memo advised redistricting leadership to consider whether a fifth Hispanic-majority district could be drawn and whether it would be required by § 2, yet mapdrawers ignored this advice. Docket no. 1274 FF472, 483, 749; docket no. 1282 at 26, 33. Hanna testified that he was able to draw a fifth Latino 40 Defendants note that the United States contended in its interrogatory responses (D-120) that redistricting leadership ignored advice from David Hanna of the TLC regarding whether the minority coalition districts in Harris County were protected. Docket no. 1249-1 at 23. However, Hanna’s initial assessment of HD137 as a Latino opportunity district changed, and he later advised that it was not a protected district. D-123. Hanna never advised that HD149 was a protected district, and in fact characterized that as a “novel” theory. Id. 54 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 55 of 171 district. TrJ1207. Downton claimed he did not attempt to draw one and does now know if anyone else did. TrJ2052-53. MALC asserts that mapdrawers avoided creating a new Latino opportunity district in Harris County by artificially increasing the SSVR in HD148 (incorrectly cited as HD145) even though it was already performing for Latino voters. Docket no. 1185 at 13 n.29. MALC argues that this increase of SSVR in a “district that the State itself has characterized as a performing Latino opportunity district, to avoid the development of a real new Latino opportunity district in Harris County is not evidence of compliance with Section 2, but rather evidence of intentional discrimination.” Docket no. 412 at 20. The Perez Plaintiffs note that Anglo population in Harris County declined such that Anglos are now only 33% of the population, yet Plan H283 gives them control of 54% of seats. Docket no. 401 at 11.41 They argue that “creative line drawing” and population deviations were necessary tools in this accomplishment. Docket no. 401 at 13 (citing Tr341-42 (Martin)).42 They further note that, although Interiano testified that they increased the SSVR in HD148 at MALDEF’s urging, Nina 41 They note that the plan was drawn by the Republican delegation, with Democrat and minority members marginalized, and “every Anglo member in Harris County is protected.” Id. at 12-13 (citing Tr349 (Martin)). Defendants take issue with this statement, noting that Hochberg, an Anglo, was paired and therefore not protected. However, Plaintiffs have argued that Hochberg was favored in this pairing, and thus it was racially discriminatory. The Court concludes that Plaintiffs have failed to prove that the ultimate configuration favored Hochberg for racially discriminatory reasons. Plaintiffs further point to various racial references made by the Republican delegation members, who drew the initial Harris County configuration. These include references in Rep. Smith’s plan log (US-441) to making districts “more Hispanic” or “more Anglo” and statements made by Rep. Woolley to Rep. Coleman. The Court agrees that these references are evidence that the Republican delegation members were thinking in terms of race when drawing the Harris County districts, but it does not rely on these statements to find intentional vote dilution. 42 As discussed in the fact findings, some of the initially suspect lines in Harris County turned out to be the result of the minority Democrats’ changes, not racially discriminatory gerrymandering. 55 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 56 of 171 Perales sent a letter to Solomons “specifically repudiating any such request.” Id. at 13-14.43 The Court finds that there is persuasive evidence of intentional vote dilution in Harris County. Harris County is yet another example of where the member-driven process was at odds with § 2 compliance. Redistricting leadership assumed that county delegations would consider VRA compliance because the delegations contained minority members, yet in Harris County the minority members were essentially shut out of the initial map-drawing process.44 Accordingly, although Hanna had recommended considering whether an additional Latino opportunity district could be drawn (as he was able to), this recommendation was ignored. Anglo Republican districts in eastern Harris County, which had grown more slowly than the more heavily minority areas, were protected, and no new minority district was considered, even though the evidence shows that a new Latino opportunity district could have been drawn that reflected the Hispanic population growth in Harris County.45 Faced with a map that failed to create any new Latino opportunity districts, mapdrawers then artificially inflated the SSVR and HCVAP of existing Latino ability district HD148 to claim VRA compliance. Redistricting leadership feigned VRA compliance but used it to undermine minority voting strength instead of truly complying with the Act. They claimed that newly majority- 43 The Court notes that Perales did not “repudiate” MALDEF’s assertion that the SSVR/HCVAP of HD148 could or should be increased above 50%. Rather, she asserted that increasing the SSVR/HCVAP of existing ability districts HD90 and HD148 could not offset the loss of HD33 because they were existing ability districts, and that additional Latino opportunity districts were still required. 44 As detailed in the fact findings, when minority members were eventually given an opportunity to give input, they were severely restricted, and unable to affect the Republican districts without agreement. 45 In the interim map proceedings, Defendants noted that their proposed Plan H303 created an additional HCVAP-majority district (HD144) while maintaining the four existing Latino ability districts (HD140, HD143, HD145, and HD148), and that HD144 “reflects the strong growth of the Latino population in Harris County,” though they refused to concede that the Gingles factors or the totality of circumstances required a new district. Docket no. 668 at 10-11. 56 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 57 of 171 SSVR/HCVAP HD148 could offset the loss of existing ability district HD33 in Nueces County and used it to thwart claims that an additional Latino opportunity district was required by § 2 based on the number of HCVAP-majority districts in Harris County. Yet the evidence is clear that mapdrawers already viewed HD148 as one of the performing Latino districts in Harris County. And as the Task Force Plaintiffs note, Downton agreed that raising the SSVR of HD148 to 50% did not enhance the ability of minority voters to elect their candidate of choice. Docket no. 1282 at 60; docket no. 1274 FF822; TrJ2050 (Downton). Thus, mapdrawers were using this superficial compliance as a tool to avoid creating any new minority opportunity districts (more specifically, to defend their failure to create any) and to mask the loss of HD33 as an existing opportunity district.46 Given the existence of racially polarized voting, the lingering effects of past discrimination, and the totality of circumstances, the Court finds that the Harris County configuration is the result of intentional vote dilution in violation of §2 and the Fourteenth Amendment. H. Fort Bend County The NAACP and MALC contend that an additional minority coalition district was required in Fort Bend County. The NAACP Plaintiffs argue that this was an area experiencing substantial population growth among a diverse group of voters, mostly minority. Docket no. 1280 at 38 (citing 46 Plaintiffs have argued that increasing the SSVR of HD148 was packing, made it more difficult to create an additional Latino opportunity district in Harris County, and weakened minority influence in other districts. See, e.g., Tr345-46 (Martin); Joint Expert Ex. E-5 (Martin Report) at 6-7; Farrar Decl. (docket no. 331-3) at 4 (noting that fewer Latinos were available to buttress emerging minority districts like HD138 and HD144). However, when the SSVR of HD148 was increased between Plan H113 and Plan H153, population was merely shifted among several minority districts, meaning that no additional minority votes were placed in a minority district and “wasted.” Thus, the act of increasing the SSVR in HD148 was not itself packing. However, mapdrawers used the fact that HD148 was now over 50% SSVR and HCVAP to ward off claims under § 2 that additional Latino districts were required. And, as noted, they used it in bad faith to claim offset of the loss of HD33. In addition to intentional vote dilution, the Court notes that this racial manipulation of the district for the sole purpose of inflating the SSVR of HD148 would seem to present a clear Shaw-type racial gerrymandering case, but no party has asserted such a claim. 57 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 58 of 171 TrJ1411 (Korbel)). They assert that, “[i]nstead of drawing compact districts that would recognize the naturally occurring minority district in Fort Bend—that is, the 150,000 more minority voters than Anglo added over the decade—the enacted plan drew HD 26 as an incredibly non-compact district, intended to be one that could be maintained as an Anglo district over the decade.” Docket no. 1280 at 38 (citing TrJ1412-14 (Korbel) and TrJ1607 (Interiano)). MALC asserts that minority population contributed 80% of the growth for Fort Bend, Wharton, and Jackson Counties, and new HD85 was placed there to account for growth, but it is not a minority opportunity district. Docket no. 1275 FF90, 91 (citing Tr1412 (Korbel)). MALC further contends that a “heavy concentration of minority Texans in Fort Bend County is divided into 4 districts, which has the effect [of] diminishing the electoral opportunities of minorities.” Docket no. 1275 FF92 (citing Tr1416-17 (Korbel)). The NAACP offers Plan H202, with HD26 as a proposed tri-ethnic coalition district that would operate as an Asian-American opportunity district. Docket no. 406 at 33. The NAACP argues that African-American, Latino, and Asian-American voters in the larger Houston area are politically cohesive, that this district had become viable for minority voters in the benchmark,47 and that the voters in this area are similar to those who act in a tri-ethnic coalition to elect Vo in HD149, just across the line in Harris County. Docket no. 625 at 7-8; docket no. 1280 at 38 (citing TrJ1422 47 Certain parties argued in the D.C. Court preclearance litigation that HD26 had become an ability district, but the D.C. Court disagreed because “the only evidence presented” showed it was not a majority-minority district and it was currently represented by an Anglo Republican. Texas v. United States, 887 F. Supp. 2d 133, 175 (D.D.C. 2012). The Court noted, “At best, the evidence may show that the districts are beginning to favor minority-preferred candidates, but section 5's effect prong protects only existing, not emerging, ability districts.” Id. According to the 2005-2009 ACS data available to redistricters, benchmark HD26 was 53.5% White Alone CVAP (though it was only 39.4% Anglo in terms of total population and 41.9% Anglo VAP), and 10.5% HCVAP, 12.6% Black Alone CVAP, and 22.2% Asian Alone CVAP. Thus, it remained majority-Anglo CVAP using 2005-2009 ACS data. However, using 2008-2012 ACS data, which the Court has concluded better reflects the true population in 2010, it had become majority-minority CVAP: the White Alone CVAP was 48.2%, Asian Alone CVAP was 25.9%, HCVAP was 12.7%, and Black Alone CVAP was 12.1%. In Plan H283, HD26 is again Anglo-majority CVAP: it is 57.3% White Alone CVAP, and 11.6% HCVAP, 10.6% Black Alone CVAP, and 19.6% Asian Alone CVAP using 2005-2009 ACS data. 58 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 59 of 171 (Korbel)). Proposed HD26 in Plan H202 would be 12.9% HCVAP, 14.5% Black Alone CVAP, 23.8% Asian CVAP (for a combined minority CVAP of 51.2%), and 47.7% Anglo CVAP using 2005-2009 ACS data. Joint Ex. J-25. The NAACP tenders Fairfax’s opinion that the district had become 57% minority CVAP by 2014 and that it complies with traditional redistricting criteria including compactness, contiguity, and respecting political subdivisions. Tr841-42, TrJ902. Rep. Senfronia Thompson testified about the coalition here and concluded HD26 would elect an Asian American and the candidate of choice of minority voters. TrJ1245-46. The NAACP Plaintiffs also assert that the evidence applicable to Harris County is “equally applicable to the creation of a new minority State House District in this area” because voting in the “larger Houston metro area is racially polarized, and the area has historically demonstrated high levels of political cohesion amongst minority voters.” Docket no. 406 at 33. Visually, HD26 is relatively compact looking, and where its shape is somewhat odd it appears to follow a river. Existing HD27 also appears visually compact and is virtually the same configuration as in Plan H283. However, the districts do split the cities of Sugar Land, Missouri City, and Arcola. Further, even when a district is visually or geographically compact, Plaintiffs must still proffer evidence that the minority communities within the proposed district are compact, taking into account traditional redistricting principles. See Gonzales v. Harris Cty., Tex., 601 F. App’x 255, *257 (5th Cir. Feb. 9, 2015) (affirming district court’s conclusion that although plaintiffs presented a “geographically compact hypothetical district” that satisfied the numerosity requirement, plaintiffs nevertheless failed to satisfy the compactness precondition because their plans did not respect traditional districting principles). Other than compactness scores, Plaintiffs have not proffered any 59 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 60 of 171 specific evidence concerning the compactness of the minority community in light of traditional districting principles such as respecting cities. Further, even if the NAACP Plaintiffs satisfied the first Gingles precondition, they did not proffer sufficient evidence supporting the second and third Gingles factors. The NAACP Plaintiffs offered only lay testimony concerning minority cohesion in Fort Bend County. While lay testimony is relevant, statistical data is essential when asserting cohesion among three different minority groups. The Court will not infer cohesion in Fort Bend County based on cohesion in nearby HD149 in Harris County.48 MALC offers Plan H329 that would create coalition districts HD26 and HD27. MALC-128. MALC also offers demonstration Plan H366 with “plug-in districts” for Fort Bend County. MALC introduced the expert opinion of Dr. Brischetto concerning racially polarized voting, though he only examined 2012 elections. The Court finds these plans should be addressed in the 2013 plan case, but notes that MALC must offer evidence demonstrating that the minority communities contained within these proposed districts are compact, taking into account traditional districting principles. Several Plaintiffs point to Fort Bend County as evidence of intentional vote dilution. The Perez and NAACP Plaintiffs contend that mapdrawers racially gerrymandered by splitting the minority community among HDs 26, 28, and 85. Docket no. 1303 at 5 (citing Martin report at 12). MALC notes that there are “several” precinct splits “between HD26 and other districts in Fort Bend County” and “[a]ll of the evidence points to the fact that the State used race-based mapping splitting 48 The Court has found that the Perez Plaintiffs did not bring § 2 results claims, though they did submit demonstration Plan H290 for Fort Bend County. To the extent this was offered as a Gingles plan, the Court finds that they did not proffer sufficient evidence that the minority communities within the proposed minority districts are compact, especially in light of the bizarre shapes and the fact that the cities of Sugar Land and Rosenberg are split. 60 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 61 of 171 precincts along racial lines to lessen minority voting strength.” Docket no. 1185 at 8. Plaintiffs note that Fort Bend County is only 38% Anglo, but Anglos control 71% of the 3.5 seats anchored there in Plan H283. Joint Expert Ex. E-5 (Martin report) at 12. Defendants argue that Fort Bend County was represented by two Republicans and one Democrat, and the incumbents signed off on the districts. Docket no. 1272 at 77; docket no. 1249-1 at 27-29 (citing April 27, 2011 House Journal Supp (D-190) at S766 wherein Rep. Howard says all three members worked together and signed off the configuration and do not want to change it). Defendants further assert that Plaintiffs fail to prove dilutive effect because it was not possible to create an additional district in which any single minority group was a majority of the citizen voting age population, and Plaintiffs failed to prove the necessary cohesion among Asian-American, African-American, and Hispanic voters for a coalition district. Docket no. 1272 at 79. NAACP Plaintiffs contend that Defendants’ response to claims that redistricters fragmented the minority community seems to rest solely on the fact that the map for the area was approved by all delegates, including one Democrat. Docket no. 1294 at 14. However, they argue that, rather than this being a member-driven map, Interiano testified that he drew it with input from the members, and he noted that the two Republican members wanted their districts to remain Republican through the decade even though this was a diverse and rapidly growing area of the state. Id. (citing TrJ1604-05, TrJ1607, TrJ1571). They argue that Defendants ignore Interiano’s role and stated purposes. The Court finds that Plaintiffs have failed to establish their claim of intentional vote dilution in Fort Bend County because they have failed to show that mapdrawers acted with a racial purpose as opposed to a partisan purpose. The district configuration came from the members, including minority Democrat member Reynolds. Though there are precinct splits, Plaintiffs failed to prove that 61 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 62 of 171 they were racial, and Defendants offered race-neutral explanations for the bizarre shape of HD26. Although Interiano testified that he wanted to keep the Republican districts Republican throughout the decade, there is no indication that he used race to do so. I. Dallas County In addition to the one person, one vote claims brought by certain Plaintiffs in Dallas County, Plaintiffs argue that mapdrawers used packing and cracking to dilute minority voting strength, despite the fact that Anglo population declined and all growth in the County was minority growth.49 Docket no. 1280 at 35-36 (“Areas in the county where the greatest minority population growth occurred were divided amongst several districts, with heavy minority populations being carved out and added to already existing minority districts.”) (citing TrJ1424 (Korbel)). Plaintiffs note that the 49 For their § 2 results claim, the NAACP offers demonstration Plan H202, with HD107 as a proposed new coalition African-American opportunity district with 23.9% HCVAP and 26.5% Black Alone CVAP using 2005-2009 ACS data. Docket no. 406 at 32; Joint Ex. J-25. This Court’s interim plan H309 did not alter the Dallas County configuration, though the current map makes some changes to districts 103 and 115. Thus, this claim remains ripe for determination in the 2013 plan case. The NAACP Plaintiffs offer Fairfax’s opinion that the district is reasonably compact, contiguous, respectful of political subdivisions, and capable of being legally enacted by the Legislature. Tr84041; docket no. 267-1 (Fairfax Supp. Report) at 7. However, on the current record Plaintiffs have failed to show that Plan H202 contains more compact minority opportunity districts than Plan H283 in Dallas County. Plaintiffs focus their evidence and briefing on establishing the Gingles preconditions for HD107, but fail to demonstrate that sufficient other districts make the preliminary showing and qualify as compact § 2 opportunity districts. To make the comparison required by De Grandy when Plaintiffs alter all fourteen districts in a county, as is done here, Plaintiffs must do more than simply point to a single new district that does not exist in Plan H283. Rather, at a minimum, they must show that all of their proposed minority opportunity districts within Dallas County in Plan H202 are reasonably compact districts under Gingles and that there are more than in Plan H283. Although they provide population and compactness statistics for their proposed districts, they fail to provide evidence that the minority populations contained therein are compact, taking into account traditional redistricting principles, as is required by the first Gingles precondition. Fairfax’s report was admitted without objection, but it contains only demographic information concerning the minority population percentages of the districts, various compactness measures of the districts (Reock, Schwartzberg, Polsby-Popper) (including for the districts in Plan H283), district population deviation, contiguity, and county and VTD splits. Fairfax notes that “[t]here are some compactness ratios that measure the dispersion of the district’s population” but “[f]or this analysis only the geographic dispersion measurements were calculated.” Docket no. 267-2 at 6 n.7; see also docket no. 267-1 at 3 n.3). Fairfax concluded that the districts in Plan H202 were all within 10% overall deviation and were contiguous and that Plan H202 “met standard districting criteria and could be adopted by the state legislature.” Docket no. 267-2 at 10. This is insufficient to satisfy Plaintiffs’ burden under Gingles. Rodriguez v. Harris Cty., Tex., 964 F. Supp. 2d 686, 741 (S.D. Tex. 2013) (“[T]he Court finds that the three quantitative metrics of compactness used by Plaintiffs are probative, but not themselves determinative, on the question of compactness.”). 62 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 63 of 171 Anglo population of Dallas County decreased by over 198,000, while the Hispanic population grew by 243,211, the African-American population grew by 73,016, and the Asian population grew by 30,302 (a total of approximately 350,000), yet no new minority districts were created and “there is some evidence that a minority opportunity seat in the county was lost.” Docket no. 1280 at 35 (citing TrJ1423 (Korbel)); docket no. 401 at 9. The Perez Plaintiffs argue that the fact that Anglos are only 33% of the Dallas County population but will control 58% of seats required sophisticated line drawing. Docket no. 401 at 9. Specifically, Plaintiffs primarily argue that: (1) in western Dallas County, existing Latino districts (HD103 and HD104) were packed (and overpopulated) to waste Latino votes and prevent additional minority opportunity or influence; (2) benchmark HD105 and HD106 were becoming minority-majority districts and on track to provide minority opportunity, but Anglo population in HD105 was increased so that it could not perform for minorities and HD106 was eliminated; and (3) in northeast Dallas County, significant minority concentration is cracked among five Anglodominated districts (107, 112, 102, 113, and 114) to eliminate emerging minority districts and dilute the minority vote. Defendants acknowledge that although it was a “drop-in county,” there was no delegation map, and Defendants state that “Downton therefore worked with several Dallas delegation members to ensure that districts with high SSVR maintained high SSVR” and “worked to protect incumbents to the extent Dallas’s population decrease allowed.” Docket no. 457 at 70; docket no. 1249-1 at 1822. Downton testified that because the Dallas County delegation was losing two Republican seats and could not agree on a County proposal, he drew the Dallas County districts, starting with HD103 and HD104, to ensure their SSVR levels remained at benchmark. Defendants contend that “race was 63 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 64 of 171 a consideration in plan H283 only [to] the extent necessary to comply with the State’s legal obligations.” Docket no. 457 at 70. Specifically, they note that changes were made to HD103 and HD104 to respond to Hanna’s concerns, and that the configurations of HD104 and HD105 were attributable to partisan and neutral motives (pairing Anderson and Harper-Brown, limiting the number of districts within Grand Prairie) and VRA compliance (maintaining HD104’s SSVR over 50% and maintaining HD103’s SSVR near benchmark). Docket no. 1249-1 at 20. Defendants contend that no additional Latino and African-American opportunity districts could be drawn while maintaining existing opportunity districts. The United States asserts that mapdrawers “intentionally prevented the emergence of a Hispanic opportunity district in Dallas County.” Docket no. 1279 at 51. The United States notes that due to population changes, HD106 elected the Hispanic-preferred candidate in 2006 and 2008, and HD105 came within 20 votes of electing the minority-preferred candidate in 2008. Id.50 It asserts that the districts showed a strong trend toward electing the Hispanic-preferred candidate, especially in Presidential election years. Id. at 51-52. The United States argues that, “[w]ith the additional minority population growth in the area, state officials realized that the minority population had a very real possibility of electing its candidate of choice in 2012, and those officials impermissibly used race to make HD105 in the 2011 Plan more Anglo.” Id. at 51.51 The United States contends that mapdrawers reversed the demographic trend in HD105 by increasing Anglo 50 In the § 5 preclearance litigation, the D.C. Court rejected the claim that HD106 was a lost ability district. Texas v. United States, 887 F. Supp. 2d 133 at 175. 51 The Perez Plaintiffs contend that HD105 can only be explained by racial line drawing. Docket no. 1263 at 5. They assert that, as Korbel explained, HD105 had experienced a drop in Anglo VAP to less than 40%, so to “remedy this,” “precinct splits took a long finger of heavily Hispanic blocks (now precincts 4653, 4654, and 4659) to be packed into HD103, and a lightning bolt went into former HD106 to add concentrations of Anglo voters.” Id. (citing Perez-133 at 12). 64 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 65 of 171 VAP by 5%, which was accomplished by “stretching the district almost the entire length of Dallas.” Id. at 52. It notes that this demographic change had a corresponding effect in the election analysis, with HD105’s election score dropping from 7/9 contests in 2008 to 0/9. Id.; docket no. 1278 FF532.52 The Perez Plaintiffs assert that the “manipulation” of HD103, HD104, and HD105 was racial and intentionally frustrated the creation of an additional minority district. Docket no. 401 at 9-11. The Perez Plaintiffs acknowledge that the explanation for the shape of HD105 was the need to pair Anderson and Harper-Brown in a Republican district, but they contend that this intentionally diluted minority voting strength in HD105 and forced Hispanic population into HD104 and that fingers were sent into northwest Irving from HD103 to withdraw Latino population. Id. at 9-10 (citing Martin Tr323). They note that HD103 became the most overpopulated district in the County and that overpopulation of these west Dallas districts frustrates the creation of an additional minority district. Id. at 10-11. Similarly, the Perez and the NAACP Plaintiffs contend that making HD105 Anglo required “major manipulation.” Docket no. 1303 at 3. They argue that precinct splits removed heavy Hispanic blocks into HD103 and HD104 and packed them to preserve HD105 as Anglo. They thus contend that HD103, 104, and 105 “manifest racial gerrymandering and minority vote dilution.” Id. To the extent Plaintiffs rely on the fact that all the growth in Dallas County was minority and that Anglo population decreased, the Court notes that the CVAP growth was not nearly as high as the total minority population growth. Dallas went from 14.11% HCVAP in 2000 to 19.17% under 52 The United States analogizes these facts to what occurred in LULAC v. Perry, 548 U.S. at 400. Docket no. 1279 at 52. In LULAC, Hispanic voters had not yet elected their preferred candidate within the district, but were poised to do so. 65 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 66 of 171 the 2005-2009 ACS data available to the Legislature (the 2008-2012 ACS data indicates 20.39% HCVAP). D-230, D-218, D-231. Under the 2005-2009 ACS data, though Dallas had a total population of 2,383,125, it was estimated that approximately 423,000 were not citizens. D-218. Further, while the total Hispanic population was estimated at approximately 900,000, only about 550,000 of those were citizens, and only 256,195 were Hispanic citizens of voting age. Nevertheless, two out of fourteen districts (about 14%) remains less than proportional to the HCVAP. Though fairness would seem to indicate that Hispanics be given more districts to achieve proportionality because their population growth prevented the loss of even more districts, mapdrawers were not required to draw Latino districts simply to achieve proportionality. In that regard, the Court finds that Plaintiffs have failed to prove intentional vote dilution with regard to their cracking claims in northeast Dallas County. Specifically, they have failed to show that mapdrawers acted with an intent other than maintaining their Republican districts or that they used race for partisan advantage. Because of the high correlation between race and party, splitting Democrats, eliminating Democratic districts, and shoring up districts for Republican incumbents necessarily affects minorities, but that alone is not intentional vote dilution based on race. The Court does find, however, that Plaintiffs have proven an improper use of race in western Dallas County to dilute Latino voting strength. The configuration of HD103, HD104, and HD105 is undisputably based in large part on race. Downton admitted to using racial shading at the block level to remove Hispanics from HD105 and place them into HD103 and HD104. Downton denies that any use of race was racially discriminatory, however, arguing that the shape of HD105 is governed by the need to pair the residences of Republican incumbents Harper-Brown and Anderson 66 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 67 of 171 within HD105 and make the district more Republican to ensure that one of them could get re-elected. He further contends that HD103 was drawn in part with Rep. Anchia and that race was used to ensure that its SSVR remained at benchmark levels to avoid retrogression. Although he did not confer with Rep. Alonzo about HD104, he asserts that it was drawn to ensure that its SSVR was above 50%. While these explanations have some superficial truth, the Court finds that the configuration uses race in part in a racially discriminatory manner to intentionally dilute minority voting strength to protect a Republican incumbent in HD105 by intentionally making the district more Anglo. Mapdrawers and redistricting leadership were hostile to the creation of a new Latino district in Dallas County because they felt it would be a Democrat district and result in the further loss of a Republican seat there beyond the two already required by the lack of growth in Dallas County overall. With regard to Dallas County, David Hanna expressly stated that they would already be losing two Republican seats but it may get “worse” because they might have to draw a third Latino district, which would mean the loss of a third Republican seat. D-192. Rather than exploring whether any additional minority districts could be drawn or maintained to recognize the population growth, they eliminated districts that were on track to perform for minority voters. Faced with existing districts 103 and 104 that mapdrawers felt had to be maintained and two Republican districts that had become majority-minority in terms of VAP and were becoming less reliable for Republicans, mapdrawers decided to eliminate HD106, the district that had actually become majority-minority CVAP (using 2008-2012 ACS data, though still Anglo-CVAP majority using 2005-2009 ACS data available to the Legislature) and had elected the minority-preferred candidate in two of five elections. Consistent with their incumbency-protection plan, they paired the 67 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 68 of 171 Republican incumbents of HD105 and HD106 in a shored-up district. While this could be acceptable partisan gerrymandering (at least under current Supreme Court law), their use of race in doing so renders it impermissible. Downton shored up HD105 by making it more Anglo (and thus more Republican)—he admitted to splitting precincts to put the Hispanic population into HD104 and HD103 and the Anglo population in HD105. Though he claimed this use of race was only to preserve the SSVR of those districts, the Court finds that this was yet another example of mapdrawers using superficial compliance with the VRA to dilute minority voting strength rather than enhancing it. Although Hanna had suggested giving thought to raising the SSVR of HD104 above 50%, he also clearly stated that it could be argued that HD104 “will likely perform at 45.6 SSVR since this is similar to the performing level it was drawn at in 2001.” D-122. Downton relied on the 50% SSVR threshold in bad faith to waste Latino votes in HD104 because there is no indication that HD104 needed to be at 50% SSVR to continue performing. Further, although HD103 was maintained at benchmark SSVR, there is no indication that mapdrawers thought this was actually necessary for VRA compliance. Downton admitted that he did not think HD103 was an opportunity district (based on his 50% SSVR requirement), and testified that they maintained the SSVR only to “try to stave off any possible legal challenge.” Downton 8-31-11 depo. (Joint Ex. J-62) at 97. Jeff Archer had stated that HD103 was not a Latino opportunity district, see D-134, and there is no indication that any election analysis was done indicating that its SSVR number needed to be maintained to comply with the VRA or to keep the district performing. Thus, Downton’s testimony is not credible. The Court finds that Downton did not believe that raising the SSVR of HD103 was necessary 68 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 69 of 171 but he saw it as an opportunity to make HD105 more Anglo and safer for the Republican by placing that Latino population into HD103. Thus, although the use of race was claimed to comply with the VRA, the Court finds that the true motive was to dilute Latino voting strength in west Dallas County by unnecessarily placing Latinos in HD103 and HD104 while simultaneously making HD105 more Anglo to protect the Anglo Republican incumbent who would emerge from the Republican primary. The Court finds that the § 2 results claims are best addressed in the 2013 plan case. However, the Court finds that mapdrawers improperly used race with an intent to dilute Latino voting strength by wasting Latino votes in HD103 and HD104 and creating a more Anglo HD105 to protect the Anglo Republican incumbent in the general election. This intentional vote dilution in Dallas County violates § 2 and the Fourteenth Amendment. J. Tarrant County The Perez and NAACP Plaintiffs cite Tarrant County as an example of intentional vote dilution through intentional minority fragmentation. Docket no. 1263 at 5; docket no. 1303 at 3. The Perez Plaintiffs argue that Plan H283 does not recognize the substantial minority growth in Tarrant County, freezing the number of minority districts at three. Docket no. 401 at 14. They assert “that Districts 90, 93, 95 and 96 reflect racial gerrymander and consequent dilution.” Docket no. 1303 at 3. Citing Martin’s report and Korbel’s report, they contend that Plan H283 minimized minority seats by packing minorities into HD90 and HD95 through bizarre configurations to dilute HD96 (preserving the Republican incumbent) and by creating HD93 as an elongated spike that splintered the minority community. Docket no. 1303 at 3; docket no. 401 at 14. The configurations of HD93, HD90, and HD95 are indeed bizarre, and Plaintiffs contend their bizarre shape can only be explained by racial gerrymandering. The evidence indicates that the 69 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 70 of 171 initial County configuration in Plan H113 was a delegation-proposed map, although the two minority-district representatives Veasey and Burnam asserted that they approved of only their own districts. Although Martin opined that HD90 and HD95 are packed, there is no indication that they were packed at this point. Their Black+Hispanic total population and B+HVAP numbers remained relatively close to and slightly below benchmark, despite the fact that substantial population had to be added to the districts. Because Hanna raised concerns about the SSVR in HD90 dropping from the benchmark (from 45% to 40% total SSVR) and MALDEF had asserted that its SSVR could be raised over 50%, and because they believed they could “offset” the loss of HD33, Downton and redistricting leadership decided to raise the SSVR of HD90 above benchmark levels to 50.1% (non-suspense), without consulting the representatives of those districts. There is no indication that HD90’s SSVR needed to be increased to that level to remain a performing ability-to-elect district. There is undisputed evidence that mapdrawers manipulated the population based on race, removing some areas and swapping out population to increase the SSVR of HD90 from 40/41.9% in Plan H113 to 47.9/50.1%. There is little indication that they cared about traditional redistricting factors or maintaining communities of interest; they were only concerned with raising the SSVR to 50.1%. The district became the most underpopulated district in the plan. The African-American community of Como was removed from HD90 and placed into a Republican district represented by Anglo Rep. Geren. Other population was swapped between HD90 and HD95, and between HD90 and HD93. Overall, substantial Anglo population was removed (approximately 12,000 total population and 11,000 VAP), while Black+Hispanic total population and B+HVAP was increased. While at first blush, this appears to be an extreme instance of packing in HD90, the changes 70 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 71 of 171 cannot be evaluated in isolation. As noted, the African-American community of Como was removed and placed into an Anglo district; this is not packing. Much population was swapped between HD90 and HD95, another minority district. In HD95, substantial Anglo population was added, while substantial minority population was removed, lowering the Black+Hispanic total population percentage as well as B+HVAP. Considering the population changes in HD90 and HD95 together, total population decreased by 4,100 Anglo and 4,200 Black+Hispanic, and voting age population decreased by about 3,600 Anglo and 3,300 B+HVAP. Thus, claims that HD90 and HD95 were packed are not supported by the evidence. However, increasing the SSVR of HD90 was merely superficial compliance with the VRA, invoked in bad faith to actually undermine Latino voting strength. As discussed, Defendants ignored DOJ guidance that ability to elect was not measured simply by a demographic criterion and claimed that they could “offset” the loss of existing Latino opportunity and ability district HD33 by increasing the SSVR of an existing ability district above 50%, even though they knew this did not create a new ability district. Further, although they did not use the 50% SSVR pretext to pack HD90 and HD95, they did use it to shore up HD93 as an Anglo district. Between Plan H113 and Plan H283, approximately 1,460 Black+Hispanic total population and 830 B+HVAP were removed from HD93 and approximately 435 Anglo total population and 413 Anglo VAP were moved into HD93, increasing the Anglo CVAP from 65.7% to 66.5%. The Court finds that mapdrawers acted with racially discriminatory intent to dilute Latino voting strength in Tarrant County. K. McLennan County MALC, the NAACP Plaintiffs, and the Perez Plaintiffs assert that there was intentional vote 71 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 72 of 171 dilution in McLennan County.53 Docket no. 1263 at 4; docket no. 1280 at 34-35. They note that benchmark HD57 was created in the Graves v. Barnes litigation, 378 F. Supp. 640 (W.D. Tex. 1974), and the minority community remained intact in that district for almost 40 years, electing the minority candidate of choice, until 2010. Docket no. 1263 at 4 (citing TrJ1442-43 (Korbel)); docket no. 1280 at 34. Benchmark HD57 had become majority-minority in terms of total population, though it remained majority-Anglo CVAP. Perez Plaintiffs and the NAACP Plaintiffs argue that majority-minority HD57 was radically altered and converted to an Anglo-majority district. Docket no. 1303 at 5. County Commissioner Lester Leon Gibson testified that the minority community had elected their candidate of choice in HD57 from 1998 until the 2010 election, when Anglo Republican Marva Beck won the district. TrJ1828-29 (Gibson). Plaintiffs contend that, to protect that incumbent, mapdrawers removed largely minority areas of Waco from HD57 and replaced them with more Anglo areas (creating a district that was 55.2% Anglo in terms of total population), splitting minority communities of interest and diluting minority voting strength.54 Korbel testified that, instead of preserving the majority-minority character of HD57, Plan H283 removed 23,000 people, 70% of whom were minority, and imported 20,000 persons, who were more than 80% Anglo, cracking minority voting strength. TrJ1444 (Korbel); see alsodocket no. 1263 at 4. The Legislature changed 53 Defendants point to the ambiguity in the Perez Plaintiff’s live pleading concerning whether they are alleging claims against Plan H283. Docket no. 1272 at 92 n.35. As discussed previously, it was clear during trial and from the Perez Plaintiffs’ arguments and post-trial briefing that they are challenging Plan H283. 54 Plaintiffs showed in Plan H363/Plan H367 that the same three districts could be configured in a way that would have created a majority-B+HCVAP district using 2008-2012 ACS data. This district maintains much of the benchmark configuration, including portions of McLennan County and all of Falls and Robertson Counties, but adds a portion of Brazos County including part of the City of Bryan. However, Plaintiffs failed to proffer sufficient evidence showing that the minority community in HD57 is compact, taking into account traditional districting principles. Plaintiffs also offered insufficient evidence of racially polarized voting for McLennan County. Plaintiffs have thus failed to prove a § 2 results claim with regard to Plan H283 but may raise this claim in the 2013 plan case. 72 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 73 of 171 the district number to HD12 and took out minority precincts in McLennan and Brazos Counties and added Limestone County. TrJ1443-44 (Korbel); TrJ1841 (Gibson). The Perez Plaintiffs contend that the reconfiguration intentionally cracked and fragmented minority voters to dilute their voting strength. Docket no. 1263 at 4. Korbel testified that the majority-minority character of the district could have been maintained, thus preserving minority opportunity to elect. TrJ1445 (Korbel); see also docket no. 1280 (NAACP Brief) at 35. Defendants contend that Plaintiffs have failed to prove that the districts were drawn with the purpose of diluting any group’s voting strength, and that Commissioner Gibson was not familiar with the 2011 House plan and had not looked at any figures showing whether there was vote dilution. Docket no. 1272 at 93. Although McLennan County fits the pattern of diluting minority voting strength in districts with significant minority population that had elected a Republican who was not the minority candidate of choice in 2010, Plaintiffs have failed to provide sufficient evidence of racial means or motive to establish an intentional vote dilution claim. It is insufficient to show effects, without evidence demonstrating intent. Coupled with the lack of statistical evidence of racially polarized voting in this County, the lack of evidence of intent leads the Court to conclude that Plaintiffs have failed to prove intentional vote dilution. L. Bell County MALC and the NAACP Plaintiffs contend that § 2 required the creation of a minority coalition district in Bell County. MALC, NAACP, and Perez Plaintiffs further contend that Bell County is evidence of intentional vote dilution. MALC’s Third Amended Complaint alleges that in Bell County, African-American and Latino population growth exceeded 51.64% and the minority 73 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 74 of 171 community is geographically compact and politically cohesive in the City of Killeen. Docket no. 897 ¶ 54. MALC contends that Plan H283 “unnecessarily fragments the minority community of Killeen to minimize its political impact on Texas House elections.” Id. The Perez Plaintiffs also allege intentional discrimination with regard to HD54. Docket no. 960 (Sixth Am. Compl.) ¶ 28.55 The NAACP Plaintiffs say that the City of Killeen is exceptionally diverse in part because of its unique relationship with Ft. Hood and that, because of its unique interests, it benefits from being kept whole. Docket no. 1280 at 31-32 (citing TrJ1706-7 (Jones)). It was kept nearly entirely whole in the benchmark HD54, with only about 200 individuals from Killeen excluded. TrJ1403 (Korbel). The NAACP Plaintiffs note that the City of Killeen experienced tremendous population growth over the decade, and HD54 was overpopulated. Docket no. 1280 at 32 (citing TrJ1706 (Jones); Tr TrJ1401-2 (Korbel)). MALC, Perez, and the NAACP Plaintiffs further note that, once Burnet County was removed, HD54 was short 13,000 voters, but instead of adding voters to the existing core of HD54, which already contained almost the entire City of Killeen, mapdrawers removed approximately 33,000 voters from Killeen, 2/3 of whom were minority, and added in Anglo voters. Docket no. 1280 at 32 (citing Tr1402-5 (Korbel)); docket no. 1263 at 5. In Plan H283, both HD54 and HD55 are majority-Anglo CVAP and the City of Killeen is split between the two districts. NAACP Plaintiffs contend that Plan H202 “created a new African American [coalition] opportunity district in Bell County.” Docket no. 406 at 33. In 2011, HD54 was proposed as a multi- 55 Although Defendants again say it is unclear if the Perez Plaintiffs are challenging the 2011 plan in addition to the 2013 plan, the Perez Plaintiffs presented claims against the 2011 plan at trial, briefed this claim in their 2014 PostTrial Brief (docket no. 1263 at 4), and included proposed conclusions of law after the 2014 trial that include HD54 in Plan H283 as having the purpose and effect of minimizing and diluting minority voting strength in violation of § 2 and the 14th Am (docket no. 1265 CL2). The Perez Plaintiffs’ joint briefing with the NAACP asserts that HD54 was racially gerrymandered to protect a Republican incumbent by splitting Killeen and fragmenting the minority population using race as a proxy for partisanship. Docket no. 1303 at 4. 74 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 75 of 171 ethnic minority coalition of 28.7% Black Alone CVAP, 17.7% HCVAP, 3.2% Asian CVAP, and .8% Indian American (46.4% Anglo CVAP) using 2005-2009 ACS data. Joint Map Ex. J-25. Fairfax testified that it would be a majority B+HCVAP district (30.99% BCVAP and 22.3% HCVAP for combined B+HCVAP of 53.29%) in 2014. TrJ912. NAACP Plaintiffs note that Fairfax found it to be a majority-minority district and that it was compact, contiguous, respected political subdivisions, complied with traditional redistricting criteria, and could be adopted by the Legislature. Tr842-43. Although the NAACP Plaintiffs offered no expert evidence of racially polarized voting specific to Bell County in 2011, they contended that “the statewide analyses performed by Dr. Kousser, Dr. Ansolabahere, and Dr. Lichtman finding racially polarized voting in Texas are all applicable to Bell County,” and they also relied on the lay testimony of Phyllis Jones, a long-time resident of Bell County. Docket no. 406 at 34; docket no. 408 at 8. They further noted that districts between 35-45% BVAP and total minority population around 70% are generally effective minority coalition districts, citing congressional districts 9, 18, and 30 as examples. Docket no. 406 at 35. In addition, Defendants stipulated during trial that racially polarized voting exists throughout Texas, other than in Nueces and Kleberg Counties. Plaintiffs fail to prove that HD54 as drawn in Plan H202 was required by § 2 in 2011. According to the ACS data available to the Legislature, it was not majority-minority-CVAP without combining three or four minority groups, and there is no indication that Plaintiffs had provided the necessary evidence concerning multi-ethnic cohesion. Thus, the Legislature’s intentional failure to create the district was not, standing alone, intentional vote dilution.56 56 To the extent Plaintiffs offer post-2011 plans and expert evidence of racially polarized voting in 2012 through Dr. Brischetto, those § 2 claims are better addressed in the 2013 plan trial. 75 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 76 of 171 However, the evidence does indicate that mapdrawers (specifically Aycock) intentionally racially gerrymandered the districts to dilute the minority vote by moving minority population out of HD54 and moving Anglo population in, thus cracking and diluting the minority vote to ensure Anglo control over both districts. Aycock was not the minority candidate of choice and often voted against NAACP positions. TrJ1703-5 (Jones); TrJ1751 (Aycock). The Perez Plaintiffs note that Killeen is heavily minority, “there has emerged an effective minority political coalition around municipal policies,” and that a “minority candidate [Brown] recently ran a reasonably strong campaign against the incumbent,” which Plan H283 addressed “by cynically splitting the Killeen minority community.” Docket no. 1263 at 4-5; TrJ1705 (Jones); see also TrJ1696-97 (Jones) (noting that minority voters had elected their candidate of choice to city council). Defendants assert that HD54 “was drawn for race-neutral reasons and without a racially discriminatory purpose.” Docket no. 1276 CL54. They contend that the Bell County configuration was created by the two members of the delegation (Jimmie Don Aycock (HD54) and Ralph Sheffield (HD55)), that it reflects the “give and take” between them, and that the demographic figures in H100 and H283 are largely the same. Docket no. 1249-1 at 32-34; docket no. 1272 at 49-51; docket no. 1276 FF145. Defendants note that most of Killeen is in HD54 and that Killeen had been divided in prior plans (and is divided in NAACP Plaintiffs’ proposed Plan H202), that minority CVAP increased in HD54 and Anglo CVAP decreased from 59.4% to 56.1%, and that Aycock testified that he felt that Lampasas County had more in common with Killeen than Temple and Belton (TrJ17334). Defendants argue that, as drawn in Plan H283, “the Bell County House districts reflect the goals identified by Representative Aycock and they do not diminish minority voting strength.” Docket no. 1272 at 51. 76 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 77 of 171 After the loss of Burnet County due to the population increase, Reps. Aycock and Sheffield were limited in their district configurations. Rep. Aycock drew the lines and “knew where the voters were and where [he] wanted to draw those lines.” TrJ1755 (Aycock). Aycock wanted more “Republican” areas and he knew that a minority coalition district would not likely have enough Republicans to re-elect him. TrJ1741-44, TrJ1769 (Aycock). Aycock drew the district to split the City of Killeen and the minority community. Although plans were introduced during the session (such as Plan H202 and Plan H232) that would have kept the City of Killeen more whole in one district, these plans were rejected, and Aycock testified it was because Lampasas County was more aligned with western Bell County than with Temple/Belton, so it was beneficial for them to remain in a district with Killeen. As noted in the fact findings, the Court does not find this testimony credible. Aycock also objected to plan H201 because it had a “land bridge” joining Killeen with Temple and Belton, though he admitted that Plan H283 includes land bridges in other areas and he voted for it. This again indicates that Aycock’s objections to minority-proposed plans were pretextual. Further, the fact that Anglo CVAP decreased in HD54 was due to the fact that minority population growth accounted for more than 70% of the growth in Bell and Lampasas Counties and the exclusion of more Anglo Burnet County. The Court finds that the decision to split Killeen and the minority community within it was to ensure that HD54 and HD55 remained Anglo-majority, and would reelect Republican incumbents. Minority communities were removed and Anglo areas were moved in, which made it more difficult for minority voters in HD54 to elect their candidate of choice. Plaintiffs presented some lay testimony of minority cohesion through Phyllis Jones of 77 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 78 of 171 Killeen.57 Dr. Brischetto found racially polarized voting in the 2012 elections in Bell County. Thus, splitting the minority community was an effective way to dilute the minority vote and ensure that the bloc-voting Anglo majority would defeat minority-preferred candidates. The Court thus finds evidence of intentional vote dilution in Bell County in violation of § 2 of the VRA and the Fourteenth Amendment. M. Midland/Ector Counties MALC contends that a new Latino opportunity district was required in Midland and Ector Counties. Defendants assert that, under the 2010 census, both Midland and Ector Counties were entitled to the same number of districts (0.8) as in 2000, so the configuration in this area is largely unchanged. Docket no. 1249-1 at 37. Following the County Line Rule, the two districts in Plan H283 are composed of whole counties—HD81 contains Andrews, Winkler, Ector, and Ward Counties, while HD82 contains Dawson, Martin, Midland, Crane, and Upton Counties. No party has asserted that a Latino opportunity district could be drawn in this area utilizing whole counties and respecting the County Line Rule. MALC proposes a new Latino opportunity district (HD81) in this area in Plan H205, Plan H295, Plan H329, and Plan H360, all of which involve County Line Rule violations. In Plan H205, proposed HD81 splits Midland and Ector Counties (along with the cities of Midland and Odessa) and joins them with six whole counties to the west plus an odd-shaped extension into Winkler County that splits the cities of Kermit and Monahans. Excising the Hispanic populations from four cities and towns spread across three counties does not respect political 57 Jones also testified that minority voters are sometimes faced with lack of translators or more rigorous questioning about ID. TrJ1699-1703. 78 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 79 of 171 subdivisions. These Hispanic populations are not compact, taking into account traditional districting principles, and appear to be joined based solely on their race. In Plan H295’s HD81,58 Midland and Ector Counties are again split, along with the cities of Midland and Odessa, and these partial counties are joined with six whole counties and a bizarrely shaped extension into Tom Green County excising a cross-section of the City of San Angelo and dividing it into three pieces. This plan suffers the same deficiency as Plan H205, but in this case the Hispanic populations of Midland and Odessa are excised and joined with the Hispanic population in San Angelo several counties over. MALC has failed to show that the Latino community contained in HD81 is compact, taking into account traditional districting principles. MALC has failed to satisfy the first Gingles precondition for its plans presented during the 2011 legislature and 2011 trial phase of this case.59 The Court also finds that the Legislature’s failure to create these proposed districts in 2011 was not intentional vote dilution. N. Lubbock County MALC contends that an additional Latino opportunity district was required in Lubbock County and the surrounding areas. In the benchmark plan, HD84 was centered in the City of Lubbock, though it did not encompass all of Lubbock and included areas to the north and east beyond the city limits. The remainder of Lubbock in the southwest and the rest of the County was placed in HD83 and joined with other counties. In Plan H283, all of HD84 is located within the City of Lubbock. The rest of the City and County is joined with almost all new counties (except for 58 Plan H296 has the same proposed HD81. 59 MALC also offered Plan H329 and Plan H360. The Court finds that the § 2 results claims based on these plans are best addressed in the 2013 plan trial. 79 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 80 of 171 Gaines County) in HD83. The idea of creating a Latino opportunity district based in part in Lubbock was presented to the Legislature via MALDEF’s Plan H115. That plan creates a long, bizarre district that connects two whole counties with parts of nine of other counties. No party offers this plan as a demonstration plan, but Figueroa testified before the Legislature that this plan was drawn in part to “see if Latinos were being split because of the county line rule.” D-595 at 74. The proposed district not only snakes through various counties but splits numerous cities (including Amarillo, Hereford, Dimmitt, Plainview, Littlefield, and Slaton) in addition to dividing the City of Lubbock into three districts to reach 51.6% HCVAP. When asked how this map showed respect for political subdivisions, Figueroa responded that it did not split VTDs and that it drew “as many counties as we could whole” though he acknowledged that only two of ten counties are whole in HD87. D-595 at 16-17. The shape of Plaintiffs’ proposed district suggests that it is not the County Line Rule, but the dispersion of the Latino population that is responsible for the lack of a Latino opportunity district in this area. MALC offers Plan H205, which was also introduced during the 2011 legislative session, with proposed HD84 (53.3% HCVAP) that excises the middle of the City of Lubbock and joins it with all of Crosby County and parts of twelve more counties stretching north and south in a bizarrely shaped, convoluted district. Defendants argue that HD84 violates the County Line Rule eleven times, is clearly drawn based on race, and has compactness scores that are two times higher (i.e., worse) than any of the district scores in Plan H283. Docket no. 457 at 29 (citing Joint Ex. J-26, Giberson Report (Ex. E-18) at 7-8); docket no. 645 at 16. Like in Plan H115, HD84 splits the City of Lubbock among three districts, and splits numerous other cities such as Littlefield, Tulia, Dimmitt, Hereford, Amarillo, Brownfield, and 80 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 81 of 171 Lamesa. It has low compactness scores and in fact appears “irrational on its face.” Again, the shape of the proposed district suggests that the Latino population in West Texas is too dispersed to be considered compact, and is not compact taking into account traditional redistricting principles such as respect for political subdivisions. See Shaw v. Reno, 509 U.S. 630, 651-52 (1993) (noting that it is “permissible for a State, employing sound districting principles such as compactness and population equality, to attempt to prevent racial minorities from being repeatedly outvoted by creating districts that will afford fair representation to the members of those racial groups who are sufficiently numerous and whose residential patterns afford the opportunity of creating districts in which they will be in the majority.”) (quoting United Jewish Organizations of Williamsburgh, Inc. v. Carey, 430 U.S. 144, 168 (1977) (emphasis in Shaw)). A district that includes “individuals who belong to the same race, but who are otherwise widely separated by geographical and political boundaries, and who may have little in common with one another but the color of their skin, bears an uncomfortable resemblance to political apartheid.” Shaw, 509 U.S. at 647. “It reinforces the perception that members of the same racial group—regardless of their age, education, economic status, or the community in which they live—think alike, share the same political interests, and will prefer the same candidates at the polls” and such “impermissible racial stereotypes” have been rejected. Id. MALC has not proffered sufficient evidence to demonstrate that the populations included in this proposed district were included for reasons other than their race. Thus, MALC has failed to satisfy the first Gingles precondition for its plans presented during the 2011 legislature and 2011 trial phase of this case.60 60 During the 83rd legislative session in 2013, MALC offered Plan H329 with HD88 in Lubbock County. Docket no. 1273 at 12; MALC-100. HD88 has 47.2% HCVAP using 2005-2009 ACS data and 50.9% HCVAP using 2008-2012 ACS data. D-322; MALC-102. MALC asserts that HD88 is “reasonably compact, split only the same county 81 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 82 of 171 The Court further finds that the Legislature’s failure to draw this proposed district in 2011 was not intentional vote dilution. O. Statewide and totality of circumstances The Task Force asserts that Plan H283 systematically thwarts Latino opportunity wherever it could have been added and has an overall discriminatory effect on Latinos which, in the face of overwhelming population growth, is circumstantial evidence of discriminatory intent. Docket no. 1282 at 26-27. The Task Force points to the strong racial tension and heated debate about Latinos, Spanish-speakers, undocumented immigrants, and voter ID legislation. Id. at 29. Flores observed the session and testified that it was the most racially acrimonious he had ever seen. Tr436. The Task Force Plaintiffs further note the rushed process with limited opportunity for public input and the failure to employ proper VRA standards (such as focusing solely on 50% SSVR) or examine § 2 compliance. Docket no. 1282 at 47-50. They argue that redistricters took away opportunities where Latinos were about to oust Republican incumbents in HD78 and HD117 and that Plan H283 disadvantages Latinos to bolster the re-election chances of those incumbents, who were not the Latino candidates of choice. Docket no. 1282 at 66. already split under H283 in Lubbock, and brought together a cohesive Hispanic population to provide an opportunity for electoral success.” Docket no. 1273 at 12. The Court finds that the results claim based on Plan H329 is best addressed in the 2013 plan case. However, the Court notes that MALC’s focus on comparing compactness scores in its proposed plans with those in the enacted plan is misplaced. MALC argues throughout its briefing that where its proposed districts are comparably compact or more compact than enacted districts, that shows compactness and/or that MALC did not ignore traditional districting principles. E.g., docket no. 459 at 14; docket no. 1273. Whether a district is compact for § 2 purposes is not determined by comparing the district’s compactness score to districts in the enacted plan, which is unquestionably politically gerrymandered. Rather, “§ 2 embodies its own requirement of compactness” regardless of whether the state has drawn compact districts. Diaz v. Silver, 978 F. Supp. 96, 130 (E.D.N.Y. 1997). Thus, MALC must offer evidence demonstrating that the minority population within its proposed district is compact, taking into account traditional districting principles. In that regard, although demonstration districts are drawn with a focus on achieving a 50% HCVAP, are not required to have the “least possible amount of irregularity,” and are not necessarily the district that must be implemented by the State, it is not enough to draw a district solely by gathering in Hispanic population and then asserting that the included Hispanics share some interests in common. 82 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 83 of 171 The Perez Plaintiffs and their expert Ed Martin contend that mapdrawers and redistricting leadership utilized three main techniques to intentionally dilute the minority vote in Texas to protect Anglo Republican incumbents. These techniques include: (1) packing existing minority districts to prevent the creation of new minority districts by using an arbitrary 50% SSVR standard (except in HD78 where it would have negative consequences for the Republican incumbent); (2) cracking the burgeoning minority population to eliminate effective minority coalition districts; and (3) using population deviations.61 Joint Expert Ex. E-5 (Martin report) at 3. By using these techniques, Martin opined that Plan H283 actually reduced the number of effective minority districts62 by at least five and up to eight districts, despite the massive minority population growth. Id. at 4. In contrast, the Perez Plaintiffs argue, Plan H232 contained the same number of County Line Rule violations (one) as Plan H283 yet created at least nine, and possibly twelve, more minority districts than Plan H283. Id. The Court agrees that the overall configuration of Plan H283 is the product of intentional vote dilution. The evidence of the mapdrawing process supports the conclusion that mapdrawers were motivated in part by an intent to dilute minority voting strength. Discussions among mapdrawers demonstrated a hostility to creating any new minority districts, as those were seen to be a loss of Republican seats, despite the massive minority population growth statewide. Although delegations worked on their respective areas, the overall map was drawn in secret, with no one seeing the statewide map until Plan H113 was released on Wednesday, April 13. 61 Plaintiffs bring Larios-type one person, one vote claims based on the population deviations, and these claims are discussed below. However, as mentioned above, certain Plaintiffs also rely on the population deviations as evidence of intentional vote dilution under § 2 of the VRA. 62 Martin defined an effective minority district as one with less than 50% Anglo VAP that had performed for minority voters in the past. Joint Expert Ex. E-5 at 4. 83 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 84 of 171 Mapdrawers intentionally paid little attention to § 2 compliance, believing that minority members in delegations would ensure VRA compliance, but the member-driven process and the limited role of minority members in Harris County were not conducive to compliance. After Plan H113 was released on April 13, the process was extremely rushed, with public hearings on Friday, April 15 and Sunday, April 17, and the plan voted out after the HRC meeting on Tuesday, April 19 (with new Plan H134 released the day before). Few changes were made and no minority opportunity was added. When faced with the fact that the number of opportunity districts had decreased, redistricters chose to artificially inflate the SSVR in existing ability districts HD90 and HD148 to claim nonretrogression and to thwart claims that additional opportunity districts were required, even though they had no reason to believe that doing so was required by § 2 or § 5. Despite the massive minority population growth, Plan H283 not only failed to create any new minority opportunity districts, it reduced the number of minority opportunity districts by eliminating HD33 in Nueces County. Redistricters also intentionally disadvantaged Hispanic voters in Hispanic districts that had elected Republicans in 2010 who were not Latino candidates of choice; they used race impermissibly and intentionally diluted Latino voting strength in HD78, HD117, and HD41 to protect Republican incumbents who were not Latino candidates of choice. The impact of the plan was certainly to reduce minority voting opportunity statewide, resulting in even less proportional representation for minority voters. Given the existence of racially polarized voting throughout Texas and the evidence of the various Arlington Heights factors (incorporated from the fact findings and the Court’s Opinion on Plan C185), as well as the totality of circumstances, the Court finds invidious discriminatory purpose underlies Plan H283. As discussed in the Court’s Opinion on Plan C185, it is not necessary for the 84 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 85 of 171 Court to find that mapdrawers or redistricters acted with actual racial animus or hostility. Rather it is sufficient that redistricters acted to undermine minority voting strength for partisan advantage. Further, although various mapdrawers drew different portions of the map, they did so with authority from the Legislature, and thus their motives, knowledge, and intent must be imputed to the Legislature when it enacted the plan. Thus, Plaintiffs have shown that, with regard to the various specific regions in which this Court has found intentional vote dilution, as well as statewide, redistricters acted at least in part with a racially discriminatory motive and intended to dilute minority voting strength in violation of § 2 of the VRA and the Fourteenth Amendment. III. Shaw-type racial gerrymandering claims In response to this Court’s request for briefing on claims after the Alabama LBC opinion was issued, MALC argued that it asserted Shaw-type racial gerrymandering claims because it alleged “racial gerrymandering” in its Third Amended Complaint. Docket no. 1306. MALC asserts that it has alleged racial gerrymandering claims to districts 26, 34, 41, 54, 77, 78, 103, 104, 105, 117, and 148. However, the Court finds that MALC did not sufficiently plead a Shaw-type racial gerrymandering claim, and that its racial gerrymandering claims are only asserted in the context of intentional vote dilution claims. The Perez and NAACP Plaintiffs asserted a number of “racial gerrymandering” claims related to districts in Dallas County, Tarrant County, Harris County, Bell County, McLennan County, Fort Bend County, and joined with MALC in its challenges to Bexar and Nueces Counties and HD41. Docket no. 1303. However, again the Court finds that these claims were phrased in terms of intentional vote dilution, not Shaw-type racial gerrymandering claims. The Task Force Plaintiffs assert that they have alleged Shaw-type racial gerrymandering 85 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 86 of 171 claims against HD117 in Bexar County and “House districts in El Paso County” in Plan H283. Docket no. 1308 at 2. Their current pleading, the Fourth Amended Complaint (docket no. 891 ¶37), alleges that mapdrawers used race as a predominant factor to allocate Latino voters into and out of HD117 and across districts in El Paso County. These claims are brought in addition to their intentional vote dilution racial gerrymandering claims. Defendants concede that the Task Force Plaintiffs have alleged Shaw-type claims with regard to HD117 and the “El Paso districts” but assert that they lack standing to bring the claim against HD117 because they failed to prove that any member of the association resides in the district. The evidence63 indicates that (1) one named Plaintiff (Sergio Coronado) resides in HD78 in Plan H283; (2) one named Plaintiff (Socorro Ramos) resides in HD75 in Plan H283; (3) one named Plaintiff (Armando Cortez) resides in HD119 in Plan H283; (4) Latino registered voter members of HOPE and MABA reside in HD117 in Plan H283; and (5) Latino registered voter members of HOPE and MABA reside in HD77 in Plan H283.64 Docket no. 1314. This is sufficient to establish standing for the Shaw-type claims asserted against HD117 in Bexar County and against HD75, HD77, and HD78 in El Paso County. The Task Force Plaintiffs rely on both district-specific and statewide evidence to support their Shaw-type claims focused on HD117 and El Paso. Docket no. 1308 at 4. In terms of statewide policies focusing on race, they point to evidence that “Texas redistricters mechanically applied a strict threshold of 50% SSVR statewide to measure compliance with section 5, despite the State’s 63 As discussed in the congressional plan opinion, the Task Force Plaintiffs’ motion to supply evidence concerning the residences of its members is granted and the information is filed under seal. 64 Although the Task Force asserts that these individuals reside in “the antler,” a search of the addresses indicates that they actually live in the “head” of HD77. 86 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 87 of 171 awareness that the U.S. Department of Justice did not mechanically apply a demographic threshold to evaluate retrogression.” Docket no. 1308 at 5. They argue that “Texas redistricters insisted that districts be drawn with greater than 50% SSVR in order to count as ability to elect districts in the State’s retrogression analysis.” Id.65 The Task Force Plaintiffs contend that the Alabama LBC decision made clear that mechanical quotas are not the measure of ability to elect or retrogression. A. HD117 The Task Force Plaintiffs argue that “Texas’s mechanical 50% SSVR test to claim nonretrogression . . . drove redistricters to use race as the predominant and controlling factor in crafting HD117.” Docket no. 1308 at 7. They further note that Texas claimed in its post-trial brief that it drew HD117 “to comply with the Voting Rights Act by maintaining the district’s SSVR over 50%.” Id. at 5 (citing Post-Trial Brief (docket no. 1272) at 55). It is undisputed redistricters determined that HD117 had to maintain an SSVR above 50%. The evidence further indicates that incumbent John Garza and redistricters also determined that it would not exceed 50.1% SSVR. The Task Force Plaintiffs note that Interiano explained that he had to perform a careful balancing act in RedAppl, shading for SSVR and using election results to craft the final HD117. Docket no. 1308 at 7. The Task Force Plaintiffs contend that “[t]he State’s mechanical use of the 50% SSVR standard to claim non-retrogression contorted the redistricting process by making race a predominant factor and led redistricters far afield of traditional redistricting 65 The Task Force Plaintiffs note that redistricters increased the SSVR of HD90 and HD148 over 50% in Plan H283 to count them as additional Latino districts in the retrogression analysis, even though they were already Latino ability-to-elect districts. Docket no. 1308 at 5-6 (citing Post-Trial Brief at 58-59). The Court notes that although several parties complain about the State’s increase of SSVR in HD90 and HD148 (and the Task Force uses this to support its assertion that redistricters used 50% SSVR as a metric), no party asserts a Shaw-type claim directly against these districts. In this context, they point to those districts as further evidence of mapdrawers’ view of the importance of having SSVR above 50% in certain districts. 87 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 88 of 171 considerations in crafting . . . HD117.” Id. at 8. Plaintiffs argue that redistricters also used race, Latino registration and electoral data to swap territory into and out of HD117 to create the district as Latino majority but not likely to elect the Latino candidate of choice; they achieved their 50.1% SSVR goal, yet Latino performance dropped when compared to benchmark. The Task Force argues that Garza’s racial statements and Interiano’s method of swapping territory in and out of the district while striving for exactly 50.1% SSVR constitute direct evidence of the use of race as a predominant factor when crafting the district. Id. at 11. The first step in a Shaw-type racial gerrymandering claim analysis is determining whether traditional redistricting principles were subordinated to race. The Court finds that Plaintiffs have made this showing with regard to HD117. Defendants admit, and the evidence is clear, that Defendants believed that maintaining the district above 50% SSVR was essential to avoid retrogression. However, they refused to draw a district that included any more than 50.1% SSVR. Maintaining this particular racial datapoint for the district was the dominant factor used by redistricters in deciding which population to include within or without the district, indicating that racial criteria predominated. Further evidence that race predominated is that Interiano’s other explanations for why certain territory was included or excluded were not credible, leaving racial criteria as the primary explanation. The Court thus considers whether this use of race served a compelling state interest. The Court finds that it did not because, although race was used to reach the 50.1% SSVR target set by redistricters, race was not used to further compliance with the VRA. Rather, redistricters insisted that the district not only have exactly 50.1% SSVR but that it also perform less favorably for Latino 88 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 89 of 171 voters. Redistricters would not accept Farias’s proposed amendment that created an HD117 with 50.1% SSVR but would also maintain Farias’s representative-constituent relationship with Whispering Winds and Somerset. Solomons’ objections–that it created jagged and awkward lines and increased the number of split precincts–were pretextual, given that these criteria had not been used to draw the map or reject other proposals, and other districts drawn by Solomons’ mapdrawers (such as the HD77/HD78 border and HD41) had more split precincts. Solomons rejected the proposal because maintaining the racial quota while at the same time ensuring the least possible Hispanic opportunity was paramount. Accordingly, redistricters lacked a compelling state interest for their use of race. The Court thus finds that the Task Force Plaintiffs have established a Shaw-type racial gerrymandering claim with regard to HD117. B. HD77 and HD78 in El Paso The Task Force Plaintiffs point to the fact that in El Paso, Downton used block-level racial data while mapping, and they argue that mapdrawers subordinated traditional redistricting criteria by splitting precincts (Downton split 14 along the HD77/78 boundary). The Task Force Plaintiffs further point to the fact that in places where it employs race-based redistricting, Plan H283 splits precincts, divides communities of interest, and does not protect partisan interests, flaunting traditional redistricting criteria. Docket no. 1308 at 14. Defendants assert that race did not predominate in the configuration of the El Paso County districts and “the basic configuration of the districts was determined by the El Paso delegation, and the evidence shows that Representative Marquez drew HD77 (including the so-called ‘antlers’) to maintain its partisan character and to increase her chances of re-election.” Docket no. 1310 at 17 89 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 90 of 171 (citations omitted). Defendants further contend that Downton modified the El Paso delegation’s proposed districts to increase SSVR in HD78 based on a specific concern raised by Hanna. The Court finds it to be a close question whether race predominated in the allocation of persons within and without HD78. While the initial configuration of the district was not race-based, there is clear evidence that Marquez modified HD77 and HD78 to assign high-SSVR precincts to her district HD77, and that Downton made further race-based changes to then increase the SSVR in HD78. While the Court finds both uses of race to be unjustified, the Court does not find that race predominated in the drawing of the districts as a whole. Accordingly, the Court finds that the Task Force has failed to establish this Shaw-type claim. IV. One Person, One Vote Claims Based on Population Deviation (“Larios Claims”)66 A. Background Several plaintiffs allege that specific areas and districts in Plan H283 violate the one person, one vote principle because of systematic underpopulation of majority-Anglo districts combined with systematic overpopulation of majority-Latino districts in urban areas, indicating a discriminatory intent that weakens the strength of votes in the overpopulated districts and systematically disadvantages minority districts. These challenges are brought primarily in reliance on Larios v. Cox, 300 F. Supp. 2d 1320 (N.D. Ga.), summarily aff’d 542 U.S. 947 (2004), in which a three-judge panel for the Northern District of Georgia invalidated two Georgia plans with top-to-bottom deviation of less than 10% because they violated the one person, one vote principle. LULAC and 66 As noted above, several Plaintiffs bring these claims under the Equal Protection Clause as violations of the one person, one vote principle. However, they also (as does the United States) incorporate the use of population deviations into their intentional vote dilution claims under § 2 and the Fourteenth Amendment. To limit redundancy, the Court has largely limited the population deviation analysis to the one person, one vote claims except where necessary to the analysis of the intentional vote dilution claims. 90 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 91 of 171 the Perez plaintiffs assert similar challenges of this sort, as does MALC. The nature of these claims, and precisely what they challenge, is discussed later. The court will first describe the general, top-to-bottom deviation in Plan H283 for context and background. Then, the court will summarize the governing legal standards for Larios-type one person, one vote claims, which will include a discussion of the legal nature of what these claims challenge from a geographic point of view and whether Plaintiffs have standing to bring these claims. Finally, and most importantly, the court will analyze whether the population deviations in Plan H283 are explained by legitimate legislative policies in ways that justify the challenged deviations. B. Plan-wide deviation in Plan H283 Ideal population for single-member district plans is the total state population divided by the total number of districts. For the Texas House, ideal population of a district based on the 2010 Census is 167,637. HD90 in Tarrant County is the smallest district in Plan H283, with a deviation of -4.90% and population 8,209 below the ideal.67 D-109 at 27, 34. HD61 in Parker and Wise Counties is the largest district in the Plan, with a deviation of 5.02% and population 8,417 above the ideal. Id. at 27, 32. Therefore, Plan H283 has an overall population deviation of 9.92%.68 Id. at 27. Martin gave Plan H283 an “F” on good-faith effort to minimize deviation, pointing out that there are fourteen counties with two or more whole districts contained within them, and they account 67 A district’s deviation percent from the ideal is its relative deviation, and is measured by dividing the absolute deviation of the district from the ideal by the ideal population. 68 “Overall population deviation is the difference in population between the two districts with the greatest disparity. Average population deviation is the average of all districts’ deviation from perfect one-person, one-vote allocation.” Abrams v. Johnson, 521 U.S. 74, 98 (1997). The difference between -4.90% and 5.02% is the total deviation—9.92%. 91 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 92 of 171 for 93 of the 150 districts. Tr379-81. These 93 districts could have been drawn with zero or minimal population deviations. Tr380. Yet eight of the fourteen counties (accounting for 75 districts) have a total deviation over 3.5% from the county ideal. Tr380, Tr383. Kousser testified that Plan H283 was “not very compliant” with the one person, one vote principle and that “the population disparities are clearly correlated with partisanship and ethnicity.” Tr248-49. He found that population deviations, which were greater in Plan H283 than in the benchmark, were not justified by any legitimate reason and heavily burdened Latinos. Tr249, Tr313. In support of his conclusions, Kousser provided a statistical analysis of the Plan. He stated that if the Legislature had been trying to minimize deviations, a histogram of population deviations would have resembled a normal curve, with the largest number of districts clustered around zero deviations. Tr233. Instead, he noted, it is more U-shaped, with the largest number of districts between four and five percent underpopulated (23 districts) and four and five percent overpopulated (20 districts). Joint Expert Ex. E-2 at 61-66. Plan H100 only had 37 districts with deviations over 4%. Id. at 6263. Turning to the racial aspects of his conclusions, Kousser stated that Anglo-majority districts are more underpopulated than overpopulated and Latino-majority districts are primarily overpopulated: 34/80 Anglo districts are overpopulated and 46/80 are underpopulated, while 22/37 Latino districts are overpopulated and 15/37 are underpopulated. Tr234; Joint Expert Ex. E-2 at 6465. Five of these 15 underpopulated Latino districts, however, are contained in El Paso County, a drop-in county that as a whole is underpopulated, which forces all of its districts to be underpopulated. Tr235; Joint Expert Ex. E-2 at 64. Not counting El Paso County, Latino districts are underpopulated at a rate of 31.25% (10/32) while Anglo districts are underpopulated at a rate of 92 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 93 of 171 57.5% (46/80); there are more than twice as many overpopulated Latino districts (22) as underpopulated ones (10). Tr235; Joint Expert Ex. E-2 at 64. In addition, Kousser noted irregularities in urban districts, both Republican and Democrat, but opined that these irregularities were more severe in urban Democrat districts and were worst in urban Democrat Latino districts. Tr235-37. In terms of urban Republican districts, 19 were overpopulated and 15 were underpopulated, with an average deviation of .47%. Joint Expert Ex. E-2 at 66. Urban Democrat districts were overpopulated at approximately the same rate (22 overpopulated, 17 underpopulated), but the deviations were more severe at .94%. Id. And, as noted, Urban Democrat Latino districts fared the worst—13 were overpopulated compared to only 6 underpopulated, with an average deviation of 1.17%. Id. Almost all districts represented by Latino Democrats were overpopulated. Tr236. Kousser found no pattern to explain the deviations in urban Republican districts, but found a bias towards overpopulation in urban Democrat districts. Id. He noted that this pattern demonstrates the Legislature’s partisan and racially discriminatory intent. Joint Expert Ex. E-2 at 67. Dr. Arrington opined that usually the population deviations in a districting plan would be unimodal, with most of the population deviations clustering near the ideal population, but that was not true in Plan H283. US-355 at 7. He added that when use of the County Line Rule is taken into account, there is no statewide, systematic overpopulation of minority districts. TrJ182. In spite of this statewide conclusion, Arrington pointed out specific instances of population deviation in Dallas, Harris, and Hidalgo Counties that he felt exemplified racially discriminatory effects that must have been intentional. US-355 at 10-13; TrJ222. Defendants’ expert, Dr. Alford, provided contrary testimony. He did not think that a 93 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 94 of 171 unimodal distribution around zero would occur if redistricters started with existing districts and modified them based on new populations, rather than starting from scratch. TrJ1887-88. This, he said, was because mapdrawers were redistributing population along county lines and preserving constituency representational relations. TrJ1888-89. Alford pointed out that both Anglo and minority districts in Plan H283 have a bimodal pattern, indicating that population equality was balanced against other redistricting interests. TrJ1890. Alford also analyzed the average deviation of Anglo districts compared with that of minority districts based on CVAP. TrJ1868-71. He found that the average minority district is overpopulated by 232 persons and that the average Anglo district is underpopulated by 124 persons. TrJ1870. Using SSVR, he found that districts with SSVR above 50% were undersized by 242 and districts with SSVR below 50% were oversized by 63. Id. In the context of an ideal district size of approximately 168,000, Alford found all of these deviations to be quite minor. TrJ1871. C. Applicable law of a Larios violation of the one person, one vote principle 1. Background “The conception of political equality from the Declaration of Independence, to Lincoln’s Gettysburg Address, to the Fifteenth, Seventeenth, and Nineteenth Amendments can mean only one thing—one person, one vote.” Gray v. Sanders, 372 U.S. 368, 381 (1963). Over time, the Supreme Court and lower courts have spoken extensively on this principle, violations of which are justiciable through the Fourteenth Amendment’s Equal Protection Clause. Baker v. Carr, 369 U.S. 186, 237 (1962). In short, the theory behind the principle is that “the vote of any citizen [must be] approximately equal in weight to that of any other citizen in the State.” Reynolds v. Sims, 377 U.S. 533, 579 (1964). As alleged here, this principle is violated because some districts are 94 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 95 of 171 underpopulated compared to ideal (giving voters in those districts greater say in electing a representative) and others are overpopulated compared to ideal (diminishing the value of a vote in these districts). “Simply stated, an individual’s right to vote . . . is unconstitutionally impaired when its weight is in a substantial fashion diluted when compared with votes of citizens living in other parts of the State.” Id. at 568. Reynolds v. Sims set forth the backdrop for one person, one vote claims, holding that, “as a basic constitutional standard, the Equal Protection Clause requires that the seats in both houses of a bicameral state legislature must be apportioned on a population basis.” Id. Recognizing that “[m]athematical exactness or precision is hardly a workable constitutional requirement,” Reynolds did not require “identical number[s],” but rather “an honest and good faith effort to construct districts, in both houses of [a state’s] legislature, as nearly of equal population as is practicable.” Id. at 577. Otherwise, Reynolds provided scant guidance, leaving room for future courts to determine how much deviation is permissible, how to allocate the burden of proof on these questions, and which reasons justify a deviation of a certain size. Reynolds did, however, provide two points of guidance. First, it described the reasons that might justify some deviation from the ideal as “legitimate considerations incident to the effectuation of a rational state policy.” Id. at 579. Second, it noted that a higher deviation is permissible in state legislative districts than congressional districts because state legislative districts outnumber congressional districts. Id. at 577-78. Because the challenges here are made to Texas House districts, this pronouncement from Reynolds guides the Court’s analysis. 2. Burden of proof i. early cases — the 10% threshold 95 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 96 of 171 Moving past Reynolds, the Court later stated that “minor deviations from mathematical equality among state legislative districts are insufficient to make out a prima facie case of invidious discrimination under the Fourteenth Amendment so as to require justification by the State.” Gaffney v. Cummings, 412 U.S. 735, 745 (1973) (emphasis added). Relying on Reynolds, the Court recognized the importance of factors aside from strict adherence to a 0% deviation, noting that “[a]n unrealistic overemphasis on raw population figures, a mere nose count in the districts, may submerge these other considerations and itself furnish a ready tool for ignoring factors that in day-to-day operation are important to an acceptable representation and apportionment arrangement.” Id. at 749. Having established that strict population equality would yield to other legitimate reapportionment considerations, the Supreme Court began to provide more specificity concerning what constituted a “minor” deviation. In White v. Regester, 412 U.S. 755, 764 (1973), the Court held that plaintiffs challenging a Texas House plan with a total deviation of 9.9% did not “satisf[y] the threshold requirement of proving a prima facie case of invidious discrimination under the Equal Protection Clause.” The Court recognized, however, that it was “[v]ery likely [that] larger differences between districts would not be tolerable without justification ‘based on legitimate considerations incident to the effectuation of a rational state policy.’” Id. (quoting Reynolds, 377 U.S. at 579). Later, the Supreme Court found that deviations of 16.5% and 19.3% “can hardly be characterized as de minimis” and that they therefore “substantially exceed the ‘under 10%’ deviations the Court has previously considered to be of prima facie constitutional validity only in the context of legislatively enacted apportionments.” Connor v. Finch, 431 U.S. 407, 418 (1977) (citing Gaffney, 412 U.S. 735 (1973) and Regester, 412 U.S. 755 (1973)) (emphasis added). Accordingly, the Court in Finch stated that “a legislatively crafted apportionment with deviations 96 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 97 of 171 of this magnitude could be justified only if it were ‘based on legitimate considerations incident to the effectuation of a rational state policy.’” Id. (quoting Reynolds, 377 U.S. at 579). This line of cases culminated in Brown v. Thomson, 462 U.S. 835 (1983). There, the Court explicitly stated the numerical point at which population deviations are treated differently by the one person, one vote principle: “Our decisions have established, as a general matter, that an apportionment plan with a maximum population deviation under 10% falls within this category of minor deviations. A plan with larger disparities in population, however, creates a prima facie case of discrimination and therefore must be justified by the State.” Id. at 842-43 (emphasis added). ii. Larios v. Cox (1) Three-judge panel opinion In Larios, a three-judge panel of the Northern District of Georgia struck down two Georgia plans with less than 10% deviation because they violated the one person, one vote principle. The court held that Georgia’s state legislative reapportionment plans plainly violate the one person, one vote principle embodied in the Equal Protection Clause because each deviates from population equality by a total of 9.98% of the ideal district population and there are no legitimate, consistently applied state policies which justify these population deviations. Instead, the plans arbitrarily and discriminatorily dilute and debase the weight of certain citizens’ votes by intentionally and systematically underpopulating districts in rural south Georgia and inner-city Atlanta, correspondingly overpopulating the districts in suburban areas surrounding Atlanta, and by underpopulating the districts held by incumbent Democrats. Larios, 300 F. Supp. 2d at 1322. Before analyzing the differences for these population deviations, the Court set forth the applicable burdens of proof, with heavy reference to Brown v. Thomson and the cases before it. The court stated that “state legislative plans with population deviations of less than 10% may be 97 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 98 of 171 challenged based on alleged violations of the one person, one vote principle,” but the 10% threshold “serves as the determining point for allocating the burden of proof” in such a challenge. Id. at 1340 (quoting Daly v. Hunt, 93 F.3d 1212, 1220 (4th Cir. 1996)). The court focused on the Supreme Court’s repeated use of the phrase “prima facie” in the Brown line of cases to indicate that 10% “is not a safe harbor”: Had the [Supreme Court] intended to foreclose all one person, one vote challenges to plans with population deviations not rising to the 10% level, the Court would undoubtedly have said as much, rather than expressing that such plans are merely “prima facie”—in other words, rebuttably—constitutional. And in this case, because the population deviations in both the House Plan and the 2002 Senate Plan are 9.98%, the ten percent rule applies (albeit barely). In short, the legislative plans are presumptively constitutional, and the burden lies on the plaintiffs to rebut that presumption. Id. at 1341. The court cited numerous district and circuit court opinions for the proposition that “plans with population deviations of less than 10% . . . are not automatically immune from constitutional attack.” Id. at 1340 (collecting cases). Applying this framework, the court found two reasons for the systematic population deviations in Georgia’s plans. First, the court noted that these deviations favored rural and inner-city interests at the expense of suburban areas north, east, and west of Atlanta. Id. at 1327–29. Second, based on circumstantial evidence, the court found that there was “an intentional effort to allow incumbent Democrats to maintain or increase their delegation, primarily by systematically underpopulating the districts held by incumbent Democrats, by overpopulating those of Republicans, and by deliberately pairing numerous Republican incumbents against one another.” Id. at 1329–31. The court recognized that traditional redistricting principles, when applied in a consistent and nondiscriminatory manner, may justify some level of population deviation. Id. at 1331; see also 98 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 99 of 171 Karcher v. Daggett, 462 U.S. 725, 741 (1983) (“The showing required to justify population deviations is flexible, depending on the size of the deviations, the importance of the State’s interests, the consistency with which the plan as a whole reflects those interests, and the availability of alternatives that might substantially vindicate those interests yet approximate population equality more closely. By necessity, whether deviations are justified requires case-by-case attention to these factors.”). Yet, in light of the regional favoritism and discriminatory incumbent protection, the court found no such justifications in the record. Id. at 1334. With regard to incumbency protection, the court noted that it can be a “permissible cause of population deviations,” but “only when it is limited to the avoidance of contests between incumbents and is applied in a consistent and nondiscriminatory manner.” Id. (emphasis added). The court recognized Supreme Court precedent indicating that “an interest in avoiding contests between incumbents may justify deviations from exact population equality, not that general protection of incumbents may also justify deviations.” Id. at 1348 (citing Bush v. Vera, 517 U.S. 952, 964 (1996); Karcher, 462 U.S. at 740; Burns v. Richardson, 384 U.S. 73, 89 n.16 (1966)). A permissible use of incumbency protection of this sort, however, was not supported by the record: In this case, it is clear that many of the incumbent-protecting population deviations were caused not by the legitimate state interest in avoiding contests between incumbents, but, rather by the more aggressive goal of allowing incumbents to avoid taking on more new constituents than was absolutely necessary to stay within 5% of the ideal district size. The defendant has not attempted to present any evidence that any of the population deviations were truly necessary to avoid pairing two incumbents. In short, in this case, the interest of incumbent protection was not applied in a reasonably consistent and nondiscriminatory way and cannot be used to justify the population deviations. Id. at 1349. With regard to the mapdrawers’ awareness of population deviations and the potential 99 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 100 of 171 ramifications of a deviation above 10%, the court noted that the mapdrawers “made no effort to make the districts as nearly of equal population as was practicable. In fact, it is quite apparent on this record that legislators and plan drafters made a concerted effort to contain population deviations to ± 5%, and no further, as they operated on the belief that there was a safe harbor of ± 5%.” Id. at 1341. Addressing the perception of a 10% deviation safe harbor, the court stated that [s]uch use of a 10% population window as a safe harbor may well violate the fundamental one person, one vote command of Reynolds, requiring that states “make an honest and good faith effort to construct districts . . . as nearly of equal population as practicable” and deviate from this principle only where “divergences . . . are based on legitimate considerations incident to the effectuation of a rational state policy.” The use of a 10% safe harbor may also conflict with the Roman Court’s observation that “the constitutionally permissible bounds of discretion in deviating from apportionment according to population” cannot be stated in a uniform mathematical formula, as it assumes that 10% is such a mathematical formula. Id. (internal citations omitted). Ultimately, however, the court did not answer this question “because the policies the [10% population window] was used to promote in this case were not ‘free from any taint of arbitrariness or discrimination’” based on the lack of traditional redistricting principles, disfavoring of suburban areas in favor of rural and inner-city populations, and discriminatory and inconsistent protection of incumbents. Id. at 1341–42 (quoting Roman v. Sincock, 377 U.S. 695, 710 (1964)). The court did not address whether “partisan advantage alone would have been enough to justify minor population deviations,” though it did point out that “the Supreme Court has never sanctioned partisan advantage as a legitimate justification for population deviations.” Id. at 1351. (2) Supreme Court affirmance A per curium Supreme Court summarily affirmed the district court’s opinion. Cox v. Larios, 542 U.S. 947 (2004). Justice Stevens wrote a concurrence, joined by Justice Breyer. Id. (Stevens, 100 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 101 of 171 J., concurring). Stevens characterized the appeal as an “invit[ation] . . . to weaken the one-person, one-vote standard by creating a safe harbor for population deviations of less than 10 percent, within which districting decisions could be made for any reason whatsoever.” Id. at 949. He was satisfied that the Court “properly reject[ed] that invitation” because one person, one vote remained the only clear limitation on partisan gerrymandering, and districting based solely on partisan interests was impermissible. Id. at 949–51. Justice Scalia dissented, and would have set the case for argument because it was not clear “whether a districting plan that satisfies [the] 10% criterion may nevertheless be invalidated on the basis of circumstantial evidence of partisan political motivation.” Id. at 951 (emphasis original). Scalia characterized the use of “politics as usual” as itself a traditional redistricting criterion, “so long as it does not go too far.” Id. at 952. He stated that “[i]t is not obvious to me that a legislature goes too far when it stays within the 10% disparity in population our cases allow. To say that it does is to invite allegations of political motivation whenever there is a population disparity, and thus to destroy the 10% safe harbor our cases provide.” Id. iii. Harris v. Arizona Independent Redistricting Commission Harris has been called “the Supreme Court’s most recent, and arguably most lucid, pronouncement as to plaintiffs’ burdens in one person, one vote cases below the 10% deviation threshold.” Raleigh Wake Citizens Ass’n v. Wake Cty. Bd. of Elections, 827 F.3d 333, 341 (4th Cir. 2016). There, a group of Arizona voters brought a one person, one vote challenge to a redistricting plan for the state legislature with a maximum population deviation of 8.8%. Harris, 136 S. Ct. at 1305–06, 1309. A three-judge panel entered judgment against the plaintiffs after a bench trial, finding that “the population deviations were primarily a result of good-faith efforts to comply with 101 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 102 of 171 the Voting Rights Act . . . even though partisanship played some role.” Id. at 1306. On appeal, the Supreme Court unanimously affirmed. Id. Much of the Court’s relatively brief opinion assessed the evidence presented at trial as it related to the plan’s deviations and the reasons for them. Id. at 1307–09. The Court ultimately affirmed because the district court’s conclusion that the deviations were justified by the legitimate factor of attempting to comply with the VRA was “well supported in the record.” Id. at 1309. In particular, the Court pointed out that the population deviations resulted from a complex and thorough series of attempts at compliance with different requirements of the VRA, “even though partisanship played some role.” Id. at 1308–09 (emphasis added). Beyond this factual conclusion, the Court clearly pronounced the applicable legal standard: In sum, in a case like this one, those attacking a state-approved plan must show that it is more probable than not that a deviation of less than 10% reflects the predominance of illegitimate reapportionment factors rather than the “legitimate considerations” to which we have referred in Reynolds and later cases. Given the inherent difficulty of measuring and comparing factors that may legitimately account for small deviations from strict mathematical equality, we believe that attacks on deviations under 10% will succeed only rarely, in unusual cases. Id. at 1307. The voters specifically attempted to invoke Larios, arguing that it warranted a different result, but the Supreme Court disagreed: In [Larios], however, unlike the present case, the district court found that those attacking the plan had shown that it was more probable than not that the use of illegitimate factors significantly explained deviations from numerical equality among districts. The district court produced many examples showing that population deviation as well as the shape of many districts “did not result from any attempt to create districts that were compact or contiguous, or to keep counties whole, or to preserve the cores of prior districts.” No legitimate purposes could explain them. It is appellants’ inability to show that the present plan’s deviations and boundary shapes result from the predominance of similarly illegitimate factors that makes [Larios] 102 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 103 of 171 inapposite here. Even assuming, without deciding, that partisanship is an illegitimate redistricting factor, appellants have not carried their burden. Id. at 1310 (internal citations omitted). While the Harris Court’s treatment of Larios gives a good indication as to the applicable legal standard, it leaves open the question of whether and to what extent partisanship can justify a deviation of less than 10%. Id. iv. Raleigh Wake Citizens Association v. Wake County Board of Elections In Raleigh Wake, a group of voters in North Carolina brought one person, one vote challenges to two redistricting plans for local governmental units that had a total population deviation below 10%. Raleigh Wake, 827 F.3d at 338–39. After a bench trial, the district court found that the plans did not violate the one person, one vote principle. Id. at 340. The Fourth Circuit reversed, holding that the voters proved their one person, one vote claims. Id. at 338. Raleigh Wake is not binding on this court, but it provides a thorough and insightful discussion of the applicable legal standard and the relationship between Larios and Harris. In Raleigh Wake, the Fourth Circuit applied the standard of review for state legislative plans as stated by the Court in Harris—“to succeed on the merits, plaintiffs in one person, one vote cases with population deviations below 10% must show by a preponderance of the evidence that improper considerations predominate in explaining the deviations.” Id. at 342 (referencing Larios and Harris). The court contrasted the result of Harris with that of Larios, but explained these different results as the result of applying different facts to the same legal standard. Id. at 341. In Harris, the plaintiffs “foundered not because their one person, one vote challenge failed as a matter of law, but because they did not muster the evidence needed to show it to be ‘more probable than not that [the] deviation of less than 10% reflect[ed] the predominance of illegitimate reapportionment factors.’” Id. (quoting 103 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 104 of 171 Harris) (alteration in original). In Larios, on the other hand, “the plaintiffs succeeded in proving their one person, one vote claims.” Id. (emphasis added). In other words, Larios and Harris are two sides of the same coin, and a joint reading of these cases clarifies the legal standard to be applied. v. summary of the burden of proof The persuasive analysis of Raleigh Wake, along with Harris and its treatment of Larios provide a clear statement of the applicable burden—because Plan H283 contains a top-to-bottom deviation of less than 10%, Plaintiffs must show that it is more probable than not that the deviations reflect the predominance of illegitimate reapportionment factors rather than the “legitimate considerations” identified by Reynolds and later cases. Harris, 136 S. Ct. at 1307. Population deviations of less than 10% are not within a safe harbor and may still be challenged. See Moore v. Itawamba Cty., Miss., 431 F.3d 257, 259 (5th Cir. 2005) (“[A] deviation less than ten percent is not a safe harbor, barring any claim of discrimination.”). 3. Geographical nature of a Larios one person, one vote claim i. applicable law Although Plaintiffs originally asserted primarily statewide claims in their pleadings, as the case developed, they challenged population deviations in discrete geographic areas. They briefed these challenges to population deviations as both statewide claims and claims made in specific geographic areas. In their Post-Trial Brief, Defendants raise, but do not develop, a discussion-worthy point about the nature of a Larios one person, one vote claim by implying that such claims do not exist at a localized level, and can be made only against a plan as a whole. See Docket no. 1272 at 147 (“Even if Larios could support a county-specific one-person-one-vote claim . . .”). This point raises a valid distinction between what was challenged in Larios and what is challenged here. In 104 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 105 of 171 Larios, the plaintiffs challenged the plan as a whole by pointing out broad, statewide deficiencies and systematic deviations that disadvantaged suburban counties surrounding Atlanta. Here, the claims are geographically more dispersed and localized, with Plaintiffs pointing out deviations in specific areas across the state. Thus, the Court must address the conceptual question of whether a Larios claim can be made against specific districts in a plan that has a top-to-bottom deviation below 10%. On this point, the court finds guidance in Alabama Legislative Black Caucus v. Alabama, 135 S. Ct. 1257, 1265-68 (2015). There, two groups brought racial gerrymandering claims against an Alabama redistricting plan. Id. at 1263. Following a bench trial, a three-judge panel rejected these claims for several reasons. Id. at 1264. The court placed the challenges into two categories: (1) claims that the plans as a whole constituted racial gerrymanders, and (2) claims that the state racially gerrymandered four specific districts. Id. It then proceeded to dispose of these claims by “referr[ing] to the racial gerrymandering claims as claims that race improperly motivated the drawing of boundary lines of the State considered as a whole.” Id. at 1265 (emphasis original). The Supreme Court disagreed with this characterization of a racial gerrymandering claim as a statewide attack on a plan as a whole. Instead, the Court held, such a claim “applies to the boundaries of individual districts. It applies district-by-district. It does not apply to a state considered as an undifferentiated ‘whole.’” Id. The Supreme Court’s reasoning for this conclusion was two-fold. First, it noted that it has “consistently described a claim of racial gerrymandering as a claim that race was improperly used in the drawing of the boundaries of one or more specific electoral districts.” Id. (emphasis original). Second, and consistent with its first point, the Court recognized that the nature of the harm arising from a racial gerrymandering claim is individual and 105 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 106 of 171 personal: [These harms] include being “personally . . . subjected to [a] racial classification,” as well as being represented by a legislator who believes his “primary obligation is to represent only the members” of a particular racial group. They directly threaten a voter who lives in the district attacked. But they do not so keenly threaten a voter who lives elsewhere in the State. Id. (internal citations omitted). In light of its view of these claims as district-specific challenges, the Supreme Court criticized the panel’s repeated characterization of the claims as attacks on the plan “as a whole.” It noted that, as an evidentiary matter, statewide evidence may be used to “prove racial gerrymandering in a particular district,” but this was different from a statewide challenge to the entire plan. Id. (emphasis added). Building off this, the Court next clarified that “voters might make the claim that every individual district in a State suffers from racial gerrymandering,” but that this was not the same as “a claim of racial gerrymandering with respect to the State as an undifferentiated whole.” Id at 1265-66 (emphasis original). Ultimately, the Supreme Court remanded for further findings with respect to the individually challenged districts because “[a] showing that race-based criteria did not significantly affect the drawing of some Alabama districts . . . would have done little to defeat a claim that race-based criteria predominantly affected the drawing of other Alabama districts, such as Alabama’s majority-minority districts primarily at issue here.” Id. at 1266 (emphasis in original). The obvious distinction between Alabama LBC and the present case is that Alabama LBC dealt with a racial gerrymandering claim. In light of the reasoning in Alabama LBC, this is a meaningful distinction—racial gerrymandering claims are consistently described and applied on a 106 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 107 of 171 district-by-district level, as the Supreme Court’s language in prior cases indicates.69 Recognizing that the Court in Alabama LBC relied heavily on its prior linguistic formulations of racial gerrymandering claims in determining their geographic scope, this Court will conduct a similar analysis with respect to the language that describes one person, one vote challenges based on population deviations under 10%. Early (pre-Larios) cases involving one person, one vote claims speak of challenges to “schemes” and to “plans,” not to individual districts. See, e.g., Reynolds, 377 U.S. at 560, 586 (“[A]ttacks on state legislative apportionment schemes, such as that involved in the instant cases, are principally based on the Equal Protection Clause . . . Since, under neither the existing apportionment provisions nor either of the proposed plans was either of the houses of the Alabama Legislature apportioned on a population basis, the District Court correctly held that all three of these schemes were constitutionally invalid.” (emphasis added)); Gaffney, 412 U.S. at 751-52 (“We must, therefore, respond to appellees’ claims in this case that even if acceptable populationwise, the Apportionment Board’s plan was invidiously discriminatory because a ‘political fairness principle’ was followed in making up the districts in both the House and Senate.” (emphasis added)); Regester, 412 U.S. at 763-64 (reversing a three-judge district court judgment that “the apportionment plan for the State of Texas is unconstitutional as unjustifiably remote from the ideal of ‘one man, one vote.’” 69 The Court quoted language from several seminal racial gerrymandering cases for this proposition. E.g., Shaw v. Reno, 509 U.S. 630, 649 (1993) (racial gerrymandering violation consists of “separat[ing] voters into different districts on the basis of race” (emphasis added)); Bush v. Vera, 517 U.S. 952, 965 (1996) (principal opinion) (“[Courts] must scrutinize each challenged district . . . ” (emphasis added)); Miller v. Johnson, 515 U.S. 900, 916 (1995) (“The plaintiff’s burden is to show . . . that race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without a particular district.” (emphasis added)). 107 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 108 of 171 (emphasis added)).70 Larios itself squarely fits this mold. Larios, 300 F. Supp. 2d at 1357 (“Judgment is entered in favor of the plaintiffs and against the defendant on Count I (Violation of Right to Equal Protection of the Laws), insofar as the plaintiffs have alleged that Georgia’s state legislative reapportionment plans violate the one person, one vote principle.” (emphasis added)). Even cases since Larios have allowed similar one person, one vote claims made on the basis of top-to-bottom deviation to proceed against a plan as a whole. See Harris, 136 S. Ct. at 1305 (“Appellants, a group of Arizona voters, challenge a redistricting plan for the State’s legislature on the ground that the plan’s districts are insufficiently equal in population . . . We consequently affirm a 3–judge Federal District Court decision upholding the plan.” (emphasis added)). On the other hand, one person, one vote claims against specific geographical areas have also been brought. In Connor v. Finch, one group of appellants challenged Mississippi’s court-ordered Senate Plan “as a whole” but did not make a “blanket challenge to the entire House Plan,” instead pointing only to several specific districts. Finch, 431 U.S. at 421 n.19. Nevertheless, the Court gave these appellants the benefit of the doubt, scrutinizing the entire House Plan even though only several specific districts were challenged: In the context of a court-ordered plan that results in the sort of systemic violation revealed by the figures in this record, it is hardly appropriate to confine our scrutiny to particularly egregious, but localized examples of violations specifically relied on by the parties. And even if the constitutional validity of the entire court-ordered House plan could not appropriately be viewed as an issue implicitly raised by the parties, this Court has the authority and the duty in exceptional circumstances to notice federal-court errors to which no exception has been taken, when they seriously 70 Regester also upheld the district court’s conclusion that multi-member districts in two specific counties were unconstitutional. Regester, 412 U.S. at 765-70. These determinations, however, were not made in the context of a one person, one vote challenge to population deviation, but rather through a different legal theory entirely. See id. 108 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 109 of 171 affect the fairness, integrity or public reputation of judicial proceedings. Id. (emphasis added) (internal citations omitted). Though the Finch Court was not entirely clear on its actions, it appears to have converted a claim against specific geographic regions within a plan into a claim against the plan as a whole due to the widespread impropriety of the deviations and public impact, even though the plaintiffs did not bring their claims as such. This question of local-versus-statewide was even more acute in Brown v. Thomson, in which the Court ultimately refused to invalidate a Wyoming Plan with an 89% top-to-bottom deviation because only one specific geographic feature of the Plan was challenged. Brown, 462 U.S. at 84648. The Court focused on one narrow issue, as framed by the district court in reliance on expert testimony—whether “the ‘dilution’ of the plaintiffs’ votes is de minimis when Niobrara County has its own representative” despite being underpopulated by over 60%. Id. at 841. The Court answered this narrow question in the negative because the deviation in Niobrara County was entirely the result of consistently applied, nondiscriminatory state policies. Id. at 844-46. Despite this holding, the 5-4 Court had to address the elephant in the room—the plan’s overall deviation of 89%. The Court explained that the question was not before it because the plaintiffs did not challenge the entirety of the Wyoming Plan: [W]e are not required to decide whether Wyoming’s nondiscriminatory adherence to county boundaries justifies the population deviations that exist throughout Wyoming’s representative districts. Appellants deliberately have limited their challenge to the alleged dilution of their voting power resulting from the one representative given to Niobrara County. The issue therefore is not whether a 16% average deviation and an 89% maximum deviation, considering the state apportionment plan as a whole, are constitutionally permissible. Rather, the issue is whether Wyoming’s policy of preserving county boundaries justifies the additional deviations from population equality resulting from the provision of representation to Niobrara County. 109 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 110 of 171 Id. at 846 (footnotes omitted). Justice O’Connor joined the majority, but wrote a separate concurrence, expressing “the gravest doubts” about the constitutionality of the statewide legislative plan, but agreeing that the relevant geographic focus was Niobrara County rather than the plan as a whole, based on the plaintiffs’ characterization of their claims. Id. at 849-50 (O’Connor, J., concurring) (joined by Stevens, J.). Justice Brennan, joined by three others, dissented. Id. at 850-61 (Brennan, J., dissenting). The dissenting justices would have reached the question of the entire plan’s constitutionality under the one person, one vote principle in light of its 89% deviation. Id. at 853-56 (Brennan, J., dissenting). Though the dissent does not squarely explain why it would do so even though the plaintiffs did not bring this claim, it points out the unlikelihood that “any future plaintiffs challenging a state reapportionment scheme as unconstitutional will be so unwise as to limit their challenge to the scheme’s single most objectionable feature . . . [A]t least plaintiffs henceforth will know better than to exercise moderation or restraint in mounting constitutional attacks on state apportionment statutes, lest they forfeit their small claim by omitting to assert a big one.” Id. at 851 (Brennan, J., dissenting). From the dissent’s point of view, the practical import of the Brown holding is limited because plaintiffs in future cases will bring appropriately broad claims. The split between the majority and dissent in Brown explains the geographic scope of a one person, one vote claim—it is whatever the plaintiff makes it. These challenges can be plan-wide, location-specific, or both. Were it otherwise, there would be no cause for disagreement among the justices in Brown because this disagreement is over whether the plaintiff had in fact brought a statewide challenge or a limited, localized challenge. The majority found that the plaintiff only 110 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 111 of 171 challenged one specific county and its district, while the dissent nevertheless felt that the plan as a whole was subject to scrutiny in that case. In either event, implicit in this disagreement is that both a state-wide challenge and a localized challenge are legally cognizable theories under the one person, one vote principle, and that it is up to the plaintiff to choose his challenge. Though it appears that the method of choice for plaintiffs in these cases is to challenge a plan as a whole, e.g., Larios, 300 F. Supp. 2d 1320, this trend of strategic choices is simply consistent with Justice Brennan’s prediction that plaintiffs in such cases would bring both narrow and broad claims. The second portion of the Court’s reasoning in Alabama LBC was that racial gerrymandering claims were properly characterized as district-by-district claims because the “harms” that arise from these violations are “personal” in that a voter in a racially gerrymandered district is personally subjected to a racial classification, is personally represented by a legislator who believes he is only there to represent members of a racial group, and is personally affected so long as he lives in the challenged district. This aspect of the Court’s reasoning is both similar to and distinct from a one person, one vote challenge. By their very nature, a plan with population deviations will overvalue voters in underpopulated districts and undervalue voters in overpopulated districts. The voter in the overpopulated district will be personally affected by overpopulation because his vote will matter less, whereas another voter in another part of the state will not necessarily suffer this same harm unless his district is also overpopulated. Thus, like a voter in a racially gerrymandered district, a voter in an overpopulated district suffers a harm that is personal to him and is not necessarily experienced by a voter in another part of the state. This explains why a one person, one vote challenge, like a racial gerrymandering claim, can be brought on a district-by-district basis. 111 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 112 of 171 On the other hand, the harms that result from population deviation can be broader. In cases like Larios and Harris, the fact that the top-to-bottom deviation in a plan as a whole is less than 10% determines the burden of proof for analyzing compliance with the one person, one vote principle. A different burden (“a prima facie case of discrimination”) applies in cases with a statewide deviation of greater than 10% in a plan as a whole. Brown, 462 U.S. at 842-43 (citing Swann v. Adams, 385 U.S. 440, 444 (1967) (discussing plan-wide deviations of 30% and 40%)). Because the law treats plans differently based on the amount of overall plan-wide deviation, these challenges can be and frequently are made to a plan as a whole. In sum, if a plaintiff characterizes his one person, one vote claim as a challenge to an entire plan with an overall deviation of less than 10%, the geographic nature of that claim is plan-wide. See, e.g., Harris, 136 S. Ct. at 1305; Larios, 300 F. Supp. 2d at 1357. On the other hand, if a plaintiff characterizes his one person, one vote claim as a challenge to a specific piece within an entire plan with an overall deviation of less than 10%, the geographic nature of that claim is limited to that specific piece of the plan. See, e.g., Brown, 462 U.S. at 846. Both are valid legal theories for challenging a plan, and both claims can be made together. ii. Plaintiffs’ claims Because the geographical nature of Plaintiffs’ claims depends on how they have presented these claims, the Court must assess the pleadings and briefing to determine precisely what is being challenged. This is no easy task. LULAC and the Perez Plaintiffs assert challenges of this sort, even though Defendants urge that the Perez Plaintiffs have not brought such a claim. See Docket no. 1272 at 3 n.5, 49 n.23. The 112 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 113 of 171 Perez Plaintiffs’ complaint is not a model of clarity,71 but it, along with their other briefing and the issues presented at trial, sufficiently states that they are making these challenges. Precisely what these Plaintiffs are alleging, however, is a moving target. Overlooking the apparent error in the Perez Plaintiffs’ and LULAC’s Joint Sixth Amended Complaint, the Court will look to this live pleading and assume that these parties meant to join in the same claims set forth therein. Under this reading, the Perez Plaintiffs and LULAC potentially challenge the following districts in Plan H283 as violating one person, one vote principles: 102, 105, 107, 113 (Dallas County); 132, 135 (Harris County); 26 (Fort Bend County); 93, 96 (Tarrant County); 54 (Bell County); and 56 (McLennan County). Docket no. 960 ¶¶ 12, 28. The Complaint seeks a declaration that “the existing plans for election of the Texas House of Representatives . . . [are] in violation of . . .the 14th Amendment,” which indicates that the Perez plaintiffs and LULAC also challenge Plan H283 as a whole. Id. at 6. However, because they also bring intentional vote dilution claims under the Fourteenth Amendment, it is difficult to discern whether all of these areas involve alleged one person, one vote violations. The complaint is only a starting point. The Perez Plaintiffs have identified other geographic areas throughout their briefing, though they do not always point to specific districts. In addition to the above-listed districts, the Perez Plaintiffs point out additional districts when discussing potential issues with population deviation. For example, in their Post-Trial Brief, they add districts 137, 144, 71 LULAC joined the Perez Plaintiffs in their live complaint, which states “Plaintiff LULAC joins this complaint for the limited purpose of challenging the redistricting of the Texas House of Representatives as set forth in paragraph 32 below.” Docket no. 960 at 2 (Perez Plaintiffs’ and LULAC’s Sixth Amended Complaint). This complaint, however, ends at paragraph 28, in which Plaintiffs assert what the court views as a Larios-type challenge to the Texas House plan. See id. Notably, however, the Perez Plaintiffs’ previous complaint, which is not joined by LULAC, lists its Larios-type claims in paragraph 32. Docket no. 898 at 8 (Perez plaintiffs’ Fifth Amended Complaint). It appears that the Perez Plaintiffs made an error in explaining which claims LULAC joins, and intended to state in their Sixth Amended Complaint that LULAC joins the Larios claims in paragraph 28. 113 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 114 of 171 and 149 in Harris County. Docket no. 1303 at 1. In addition, the Perez Plaintiffs bring up certain counties in their briefs but it is unclear whether they rely upon the Fourteenth Amendment for the intentional vote dilution or one person, one vote claims, or both. E.g., Docket no. 1303 at 5 (discussing McLennan County). In their briefing, it is also unclear whether they are challenging specific geographic areas (whether certain districts, certain counties, or certain unnamed districts in certain named counties), whether they are pointing to these problem areas as evidence of a plan-wide claim, or both. For guidance, the Court looks to the Perez Plaintiffs’ Proposed Conclusions of Law. Docket no. 1265. There, the second conclusion of law lists the following districts as violating the Fourteenth Amendment, though again it is unclear whether this is intended to be limited to their intentional vote dilution claims: HDs 101, 102, 103, 104, 105 (Dallas County); HDs 132, 137, 145, 148, 149 (Harris County); HD93 (Tarrant County); HD117 (Bexar County); HD41 (Hidalgo County); “McLennan County District 57 in the Benchmark Plan” (which is associated with HD56); HD 54 (Bell County); and HD26 (Fort Bend County). Id. In addition, the third proposed conclusion of law states, “HB 283 has unconstitutionally excessive population deviation among the House districts in Dallas, Harris and Tarrant Counties. These deviations are the result of gerrymandering along racial ethnic and partisan lines, in violation of the principles established in Cox v. Larios, 542 US 947 (2004). Accordingly, the redistricting plans for these counties are void.” Id. Between the districts identified in the Sixth Amended Complaint and these proposed findings of fact and conclusions of law, the Court construes this to be the scope of the Perez Plaintiffs’ claims. At their broadest, the Perez Plaintiffs challenge Plan H283, as a whole, along with the entire counties of Dallas, Harris, and Tarrant. Construing the claims as such reconciles the main 114 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 115 of 171 differences between the identified districts in the Complaint and conclusions of law, which are mostly in Dallas and Harris Counties. Finally, they challenge the identified districts in Hidalgo, Bell, and Fort Bend Counties, as well as all districts in Dallas, Harris, and Tarrant Counties individually. Importantly, and more generally, the Perez Plaintiffs and LULAC allege Larios violations “[p]rimarily in the multi-member urban counties,” but make clear that they are not alleging violations “mainly in the ex-urban and rural areas of the state” where “there may be a plausible explanation for the population deviations” in the County Line Rule. Docket no. 738 at 21-22. They state that “the entire House plan does not have to be redrawn.” Id. The Court construes these more generalized claims as an adoption by LULAC of the claims that the Perez Plaintiffs have brought to specific districts and counties in the “multi-member urban counties” throughout the state. This would encompass the aforementioned counties and districts, along with Nueces County. MALC asserts similar claims. See Docket no. 734 at 8–13. Broadly, MALC’s claim is that Latino-majority districts within whole counties were overpopulated. Id. Aside from the generalized claim, MALC’s briefing identifies several particularly problematic areas in Harris County (HDs 145, 147, 149) and Hidalgo County (HDs 30, 31, 36, 39, 41). Id. It appears that MALC also asserts a claim against Plan H283 as a whole. Docket no. 1275 at 45–47 (MALC’s Proposed Conclusions of Law). Id. D. Standing Before the Court turns to the merits of the Plaintiffs’ claims, it must consider whether they have standing on the Larios-type one person, one vote claims that they assert. The elements of standing are well-established: the irreducible constitutional minimum of standing contains three elements. First, the 115 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 116 of 171 plaintiff must have suffered an “injury in fact”—an invasion of a legally protected interest which is (a) concrete and particularized, and (b) “actual or imminent, not ‘conjectural’ or ‘hypothetical[.]’” Second, there must be a causal connection between the injury and the conduct complained of—the injury has to be “fairly . . . trace[able] to the challenged action of the defendant, and not . . . th[e] result [of] the independent action of some third party not before the court.” Third, it must be “likely,” as opposed to merely “speculative,” that the injury will be “redressed by a favorable decision.” Lujan v. Defs. of Wildlife, 504 U.S. 555, 560–61 (1992) (internal citations omitted) (most alterations original). “An association has standing to bring suit on behalf of its members when its members would have standing to sue in their own right, the interests at stake are germane to the organization’s purpose, and neither the claim asserted nor the relief requested requires the participation of individual members in the lawsuit.” Friends of the Earth, Inc. v. Laidlaw Envtl. Services (TOC), Inc., 528 U.S. 167, 181 (2000). The geographic nature of the parties’ claims makes a complex, party-by-party standing inquiry unnecessary because as discussed above, LULAC asserts the broadest claims with respect to individual districts, and LULAC has members in each of the challenged districts identified above.72 In order to establish its standing, LULAC recently filed a motion for leave to supplement the record with a declaration from its Texas State Director, Elia Mendoza. Docket no. 1342. In the declaration, Mendoza states that “there are multiple members of LULAC who are eligible voters in 72 Again, we find guidance from Alabama LBC. Aside from addressing the geographic nature of the claims brought, the Supreme Court reviewed the district court panel’s opinion on standing. Alabama LBC, 135 S. Ct at 1268. The district court sua sponte dismissed claims brought by the Alabama Democratic Conference—one of the two plaintiffs—because it found that the Conference did not prove that it had members in the challenged districts. Id. at 1268–70. The Supreme Court remanded, stating that “elementary principles of procedural fairness required that the District Court, rather than acting sua sponte [and dismissing], give the conference an opportunity to provide evidence of member residence” by providing a membership list. Id. What necessitated the standing analysis in the first place, however, was that the Conference brought a broader list of claims than the other plaintiff, the Alabama Legislative Black Caucus, which had adequately proven its standing but brought a narrower list of claims. Id. at 1264. Here, LULAC is both the party with the broadest standing and the party bringing the broadest array of claims. Therefore, unlike in Alabama LBC, the standing inquiry is less important, and the Court rests its analysis of the merits on the standing of LULAC, which covers all of the Larios claims in the lawsuit. 116 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 117 of 171 each of the . . . State Legislative Districts” in Dallas County, Tarrant County, Harris County, Bell County (including specifically the city of Killeen), the “Lower Rio Grande Valley Counties” (which would encompass Hidalgo County), and Fort Bend County, among others. Docket no. 1342-1 at 3. In addition, LULAC identifies specific members by name in HD149 (Harris County), HD54 (a challenged Bell County district), HD34 (a challenged Nueces County district), and HD36 (a challenged Hidalgo County district). Id. Defendants oppose LULAC’s request for leave to supplement the record with Mendoza’s declaration. Docket no. 1343. They argue that LULAC does not justify its failure to introduce this evidence sooner, and that supplementing the record now is prejudicial because they never had an opportunity to challenge the declaration. Id. at 3-4, 6-7. In addition, Defendants point to a stipulation that they made with LULAC to end a previous discovery dispute between the parties. Id. at 4 (citing docket nos. 955, 956, 975, 1000). This dispute arose from LULAC’s alleged failure to supplement its initial disclosures and respond to certain interrogatories and requests for production. See docket no. 955 at 2. The Court granted Defendants’ initial motion to compel. Docket nos. 955, 956. After LULAC allegedly failed to comply with that order, Defendants filed two motions for sanctions. Docket nos. 975, 1000.73 By their motion for sanctions, Defendants sought to prevent LULAC from introducing evidence and witnesses at trial relating to the merits of its claims. See docket no. 1000 at 3-5. Ultimately, the parties resolved this dispute through a stipulation, which stated: In connection with Defendants [sic] Second Motion for Sanctions Against the LULAC Plaintiffs, [Defendants] and [the LULAC Plaintiffs], hereby agree as follows: 1. 73 Defendants withdrew the initial motion for sanctions, docket nos. 975, 977, and later filed the second motion, which is similar to the first and seeks the same relief as the first, docket no. 1000. 117 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 118 of 171 LULAC’s only witness at the 2014 trial will be their disclosed expert George Korbel. LULAC agrees that it shall not call any other witnesses at the 2014 trial. 2. At trial, LULAC will only offer exhibits that were: a. previously offered by LULAC in the 2011 Section 2 trial of this matter; or b. provided to Defendants on May 24, 2014, and at the deposition of George Korbel scheduled for May 27, 2014. Docket no. 1002 (emphasis added). LULAC’s motion for leave to supplement the record is granted. As an initial matter, the Court points out that the prior discovery dispute centered around information unrelated to LULAC’s standing or the residency of its members, and the resulting stipulation between Defendants and LULAC was made “[i]n connection with” that dispute. As Defendants’ motion for sanctions made clear, Defendants were only seeking LULAC’s contentions and proposed testimony related to the merits of its case. See docket no. 1000 at 3-5.74 The language of the discovery requests at issue and a letter from Defendants’ attorneys to LULAC’s attorney likewise show that this discovery dispute was never about evidence, facts, or information relating to LULAC’s standing. See generally docket nos. 975-5, 975-6, 975-7, 975-8, 975-9. Thus, because LULAC’s stipulation was made “in connection with” a discovery dispute surrounding the merits of LULAC’s contentions, that stipulation does not extend to the present request for leave, which is limited to only evidence on LULAC’s standing. Beyond this, “[i]n deciding whether to reopen evidence, a court should weigh the importance 74 Without rehashing the entirety of the relief that Defendants sought by their motion for sanctions, the Court recounts two illustrative examples: The State Defendants request an order prohibiting the LULAC plaintiffs from presenting evidence, testimony or otherwise supporting their claims on the following issues, is appropriate at this time: [sic] 1. LULAC’s contention that the Texas Legislature engaged in intentional discrimination against minorities in passing the 2011 Congressional Plan (C185) and the 2011 Texas House of Representatives Plan (Plan H283). (RFP 2). . . . 17. Testimony by unknown individuals involved in preparing plans of apportionment, and testimony about such plans. (2011 Disclosures at ¶ 5). Docket no. 1000 at 3, 5. 118 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 119 of 171 and probative value of the evidence, the reason for the moving party’s failure to introduce the evidence earlier, and the possibility of prejudice to the non-moving party.” Aransas Project v. Shaw, 775 F.3d 641, 655 (5th Cir. 2014) (internal quotations omitted). Courts are permitted “wide discretion” in exercising their powers over the presentation of evidence, which includes the reopening of evidence. In re Perry, 345 F.3d 303, 309–10 (5th Cir. 2003). LULAC considers its failure to present the declaration sooner as a simple technicality, docket no. 1342 at 2, while Defendants complain that they will be severely prejudiced by being unable to challenge the declaration, docket no. 1343 at 6-7. Neither party addresses the importance and probative value of the declaration, but both are extremely high and weigh heavily in favor of granting leave. Without the declaration, LULAC’s standing will be entirely unclear, and such a deficiency will have come to light only long after the end of trial and after the Court has expended vast resources in analyzing the parties’ contentions on the merits. Similarly, the probative value of LULAC’s declaration is high, as the Supreme Court has recognized in similar circumstances that there is “no reason to believe that [an association plaintiff in redistricting litigation] would have been unable to provide a list of members . . . had it been asked.” Alabama LBC, 135 S. Ct at 1269–70.75 Further supporting Plaintiffs’ request for leave is Alabama LBC, which guides courts to recognize “elementary principles of procedural fairness” and consider allowing presentation of a membership list before dismissing redistricting claims for a lack of standing. Id. Consistency also dictates this result. In its previous order addressing the Plaintiffs’ challenges to the congressional maps, the Court addressed a similar request from the Task Force for 75 To be clear, this is not to say, as LULAC requests, that the Court takes judicial notice of LULAC’s membership list. Instead, the Court simply recognizes that the organization’s declaration of the residences of its own members is highly probative of where those members reside. 119 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 120 of 171 leave to file under seal a list of its members. Docket no. 1339 at 100-01 (granting docket no. 1314). The Task Force and Defendants presented similar arguments concerning the excuse for the failure to present standing evidence sooner and the prejudice to Defendants in granting the motion. Ultimately, the Court granted the Task Force’s motion: The Court initially found that the Task Force had satisfied the pleading standard for associational standing, and neither the Court nor Defendants again raised the issue. Given the undisputed existence of standing with regard to CD6 and the strong likelihood that an individual member of the Task Force’s member organizations would have standing to challenge CD26, the Task Force was not dilatory in seeking to admit evidence, especially given that the Shaw claims have not been a primary focus of the litigation, the issue had not been pressed by Defendants or the Court, and the Task Force had First Amendment and privacy concerns about releasing member information. Defendants are not surprised by this claim and were not so concerned about the issue to have raised it again via motion for summary judgment or otherwise. See Mullaney v. Anderson, 342 U.S. 415, 417 (1952) (allowing joinder of parties on appeal to address lack of standing and noting that “[t]o . . . require the new plaintiffs to start over in the District Court would entail needless waste and runs counter to effective judicial administration”). Docket no. 1339. This logic holds true here. The Larios claims at issue in this standing analysis have not been the primary focus of the litigation up to this point. Prejudice to Defendants is mitigated by the fact that they were on notice of the declaration (which LULAC originally filed in April 2015 as an exhibit to its briefing on Alabama LBC, docket no. 1302-1), and did not seek summary judgment or otherwise press the standing issue. Accordingly, LULAC’s motion for leave is granted. Having granted LULAC’s motion for leave to supplement the record with Mendoza’s declaration that sets forth the residences of LULAC’s members, the Court finds that LULAC has standing to bring Larios one person, one vote claims because it has individual members in each of the challenged areas. Because no other plaintiff asserts a Larios claim in an area where LULAC does 120 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 121 of 171 not, whether other plaintiffs have standing is irrelevant, and the Court must analyze the merits of these claims regardless. See supra footnote 71 and accompanying text. E. Analysis of the merits of the population deviation claims “[T]he proper judicial approach is to ascertain whether, under the particular circumstances existing in the individual State whose legislative apportionment is at issue, there has been a faithful adherence to a plan of population-based representation, with such minor deviations only as may occur in recognizing certain factors that are free from any taint of arbitrariness or discrimination.” Roman v. Sincock, 377 U.S. 695, 710 (1964). The Roman Court’s pronouncement that minor deviations (below 10%) can only be justified by “certain factors that are free from any taint of arbitrariness or discrimination” is often quoted, and provides a high-level guide for what types of factors plaintiffs may use as evidence to invalidate a plan with a deviation below 10%. See, e.g., Larios, 300 F. Supp. 2d at 1340; Gaffney, 412 U.S. at 777 n.4; Finch, 431 U.S. at 415; Brown, 462 U.S. at 843; Harris v. Ariz. Indep. Redistricting Comm’n, 993 F. Supp. 2d 1042, 1090 (D. Ariz. 2014), aff'd, 136 S. Ct. 1301 (2016). The Supreme Court has identified several specific traditional redistricting criteria that, when applied in a nondiscriminatory manner, will justify deviations. Karcher v. Daggett, 462 U.S. 725, 740 (1983). This list includes “making districts compact, respecting municipal boundaries, preserving the cores of prior districts, and avoiding contests between incumbent Representatives.” Id. A Larios-type one person, one vote challenge necessarily focuses on the reasons behind the population deviations because of the applicable burden of proof—the plaintiffs must show that it is more probable than not that a deviation of less than 10% reflects the predominance of illegitimate reapportionment factors rather than the “legitimate considerations” identified by Reynolds and later 121 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 122 of 171 cases. See Harris, 136 S. Ct. at 1307. In general, the parties contest the legitimacy of the reasons behind the deviations, to the extent that there are reasons given. Looking to these reasons, Defendants have not consistently argued that making districts compact and preserving the cores of prior districts created the deviations, even though these are permissible reasons for population deviations under Karcher. In addition, Defendants point to further reasons that are not identified in Karcher but that nonetheless warrant full discussion. The court now turns to these reasons. 1. Mistaken belief in a 10% safe harbor The mapdrawers’ most common explanation for the deviations is that they believed that under Larios, a map with a deviation under 10% could not be challenged. Tr994 (Downton); Tr1473-74 (Interiano). Mapdrawers believed that they would not have to proffer any further explanation. Id. Because of this belief, population deviations were intentionally not minimized beyond 10%, and were not explained or justified on the legislative record during the session. Tr1494 (Interiano): Downton 8-12-11 depo. (Joint Ex. J-62) at 14-15. Mapdrawers were aware of the one person, one vote principle and of Larios v. Cox. Interiano 8-26-11 depo. (Joint Ex. J-61) at 11-12. Hanna stated at a March 2011 HRC hearing that a 10% deviation is allowed “without justification.” D-590 at 40. He cautioned, however, that Larios might mean that “the old ten percent rule, which was a safe harbor, may not be any more.” Id. Hanna went further: “[i]f all [of a] certain kind of district are drawing small, then we start to run into the sort of issues that they [had] in [Larios v.] Cox. Or if all the districts are drawing big, we start to run into that question. That’s not fully developed jurisprudence yet but that starts to get me a little bit concerned.” Id. at 45. Hanna advised that, to avoid Larios problems, the Legislature should 122 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 123 of 171 “[e]nsure that all deviations are justified by a legitimate, consistently applied policy such as the preservation of county lines” and should “[m]ake sure that deviations in a plan with a total deviation of less than 10 percent do not consistently advantage or disadvantage one or more regions, racial or ethnic groups, or political parties.” Perez-132 at 1. At the April 15 HRC meeting, Rep. Madden and Luis Figueroa of MALDEF also discussed Larios, and Rep. Madden noted that the case “indicated that you had to pay attention to deviations from the standard so that you didn’t have one group as under and other groups as over.” D-595 at 57. The TLC’s publication, State and Federal Law Governing Redistricting in Texas included similar admonitions: [I]t is reasonably clear that, so long as the total deviation range of a state legislative plan remains under 10 percent, the state is not required to strive for a more exacting level of population equality. Within this range, the state is relatively free to use population deviation for any rational purpose. However, as discussed . . . , a discriminatory scheme of population deviation may be invalid for other reasons, even if the range of deviation is less than 10 percent. . . . Even if a legislative plan has an overall range of population deviation less than 10 percent, a pattern of population deviation within that range to further invidious intentional discrimination or that inadvertently results in the systematic underrepresentation of a racial or ethnic group may be held invalid on other grounds. . . . To minimize the chance of a successful challenge under the somewhat amorphous Larios standard, mapmakers may want to consider, in a legislative redistricting plan with an overall deviation of less than 10 percent, avoiding deviations that consistently advantage or disadvantage a particular political, racial, or ethnic group or region of the state. US-357 at 36-37, 39. In spite of these warnings, mapdrawers assumed they could have a 10% deviation without justification, and drew the map on that assumption. Yet this assumption is not only wrong in light of the governing law, but also in light of the record—mapdrawers were repeatedly warned that under Larios, plans with deviations below 10% were not immune from challenge. 123 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 124 of 171 More importantly, this mistake, even if it explains the existence of the deviations, is not a justification for the deviations. It may be true in part that mapdrawers did not pay attention to deviations (so long as they were under 10%) because they mistakenly believed that they had a 10% safe harbor.76 But the free-wheeling use of deviations up to 10% without furthering any other constitutionally acceptable redistricting criteria is not itself a constitutionally acceptable redistricting criterion. In Reynolds, the Supreme Court stated that “[s]o long as the divergences from a strict population standard are based on legitimate considerations incident to the effectuation of a rational state policy, some deviations from the equal-population principle are constitutionally permissible.” Reynolds, 377 U.S. at 579. This pronouncement set up a simple framework—deviations may exist, but they must be justified by the right reasons. Cases beyond Reynolds have added the 10% threshold to this framework as the leeway for mapdrawers to further other legitimate criteria; the threshold itself is not a legitimate criterion. To hold otherwise would overlook the Supreme Court’s language in Roman describing which criteria may justify minor deviations below 10%: “[F]actors that are free from any taint of arbitrariness or discrimination.” Roman, 377 U.S. at 710 (emphasis added). Still, though the mistaken belief in a 10% safe harbor does not justify a deviation, the burden of proof is on Plaintiffs to show that it is more probable than not that a deviation of less than 10% reflects the predominance of illegitimate reapportionment factors rather than the “legitimate considerations” identified by Reynolds and later cases. Harris, 136 S. Ct. at 1307. Defendants argue that Plaintiffs have not carried this burden because the mapdrawers’ mistaken belief was 76 As discussed, this factual assertion is not credible—mapdrawers were repeatedly warned that the Plan might be scrutinized even if its overall deviation was below 10%. 124 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 125 of 171 accompanied by the occasional use of other traditional redistricting criteria and was not accompanied by the use of any illegitimate factors. On this point, Defendants argue that Larios should yield to Rodriguez v. Pataki, 308 F. Supp. 2d 346 (S.D.N.Y. 2004). There, the court found “that an express objective of staying within a ten-percent deviation while pursuing other legitimate goals provides no support to the plaintiffs’ claim of invidious or arbitrary discrimination or of bad faith.” Id. at 367 (emphasis added). The court ultimately granted summary judgment against the plaintiffs’ one person, one vote challenge to a New York state plan with an overall deviation of 9.78%. Id. at 365, 370-71. Pataki and Larios are not inconsistent, though they are factually distinct. On the impermissible and inconsistent use of incumbency protection, the Pataki court noted that the New York plan, unlike the Georgia plan in Larios, “reflect[ed] a legitimate effort to avoid incumbent pairs.” Id. at 367-68 n.23. As for favoritism of some geographic areas of a state at the expense of others, the New York plan “appl[ied] the accepted principle of maintaining the core of districts” in both the allegedly favored and disfavored districts. Id. at 370 n.27. Importantly, the Pataki court identified “many permissible redistricting considerations” that influenced the mapdrawers: “[The State Legislative Task Force on Demographic Research and Reapportionment] was interested in contiguity, compactness, preserving the cores of existing districts, desiring not to pit incumbents against one another, respecting then-current political subdivisions and county lines, and staying within the ten-percent-deviation parameter of Brown.” Id. at 367-68. In Larios, on the other hand, no similar considerations were at work. Larios, 300 F. Supp. 2d at 1331. Though Georgia argued that its use of incumbency protection could justify the deviation, the court found that the use of incumbency protection was inconsistent and discriminatory. 125 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 126 of 171 Id. The state did not attempt to identify any other “state polic[y] that, when applied in a consistent and nondiscriminatory manner [could] justify some population deviation.” Id. (citing Karcher 462 U.S. 725 (1983)).77 Pataki is similar to the present case in one main way—in both cases, mapdrawers knew that the legal implications of their plans would change when overall population deviation exceeded 10%. This similarity alone, however, is not in any way outcome determinative, just as this fact alone was not dispositive in Pataki. Though the Pataki court recognized that “[n]o invidious purpose can be inferred from such conduct alone,” the court there found that New York’s mapdrawers were mindful of the 10% threshold “while pursuing other legitimate goals.” Pataki, 308 F. Supp. 2d at 367 (emphasis added). As explained in the court’s contrast to Larios, the New York mapdrawers’ awareness of the threshold was accompanied by several other traditional, neutrally applied redistricting criteria that were present in the record. Therefore, this Court must assess the other purported justifications for the population deviations to determine whether Plaintiffs have carried their burden. Specifically, the Court will search for “factors that are free from any taint of arbitrariness or discrimination” that may have caused the deviations. Roman, 377 U.S. at 710. 2. Adherence to the County Line Rule Mapdrawers adhered to the County Line Rule, as they interpreted it, and Defendants argue that this caused the deviations in Plan H283. Under this Rule, whenever a single county has sufficient population to be entitled to a Representative, 77 “The other policies [that could justify the deviations] were not causes of the population deviations in the House Plan and 2002 Senate Plan; nor indeed, were they priorities at all in drafting the plans. In fact, the defendant has never claimed that they were.” Larios, 300 F. Supp. 2d at 1331. 126 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 127 of 171 such county shall be formed into a separate Representative District, and when two or more counties are required to make up the ratio of representation, such counties shall be contiguous to each other; and when any one county has more than sufficient population to be entitled to one or more Representatives, such Representative or Representatives shall be apportioned to such county, and for any surplus of population it may be joined in a Representative District with any other contiguous county or counties. TEX. CONST. art. III, § 26. A rule of this sort was identified by the Karcher Court as a “legislative polic[y] [that] might justify some variance,” so long as it is evenly applied. See Karcher, 462 U.S. at 740 (“Any number of consistently applied legislative policies might justify some variance, including, for instance . . . respecting municipal boundaries.”). Therefore, the Court will analyze whether the County Line Rule was in fact a cause of the deviations in Plan H283 to determine whether it justifies them. It is clear that use of the County Line Rule would necessarily result in a top-to-bottom deviation of close to 10%, which explains why the County Line Rule is a legislative policy that can justify deviations from the statewide ideal. Using whole counties necessarily limits mapdrawers’ ability to control the population within districts, and adherence to the County Line Rule would cause some top-to-bottom deviation in Plan H283, even if used in consistent and neutral ways on a countyby-county basis. For example, HD61 in Parker and Wise Counties (which is not challenged) is the most overpopulated district in Plan H283 with a deviation of 5.02% (8,417 persons), but adherence to the County Line Rule explains this deviation—the Rule necessitates the joinder of these two whole counties and the addition of their population. However, the most underpopulated district, HD90, is one of several districts within Tarrant County, and thus the County Line Rule did not dictate its exact size. Therefore, the County Line Rule does not explain the entire top-to-bottom deviation of Plan H283. 127 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 128 of 171 Moreover, application of the County Line Rule would not justify deviations among districts within multi-district counties (also referred to as “drop-in counties”). Such districts could have equal population based on a county ideal size or could all have roughly equal deviations from statewide ideal. Interiano calculated deviation only statewide, without looking at individual counties except to fit them in to the statewide deviation. Interiano 8-26-11 depo. (Joint Ex. J-61) at 10-11. Similarly, Solomons was not aware of any legal justification for intra-county deviations, but assumed the staff would know why there were deviations. Tr1596-98 (Solomons). Accordingly, many deviations (discussed below), which occur within counties, are not the result of the County Line Rule. Dr. Alford made a broad-based attempt to justify the county-level deviations in light of the County Line Rule. TrJ1872. He reached this general conclusion in two ways. First, he analyzed adjusted average population deviation by taking into account drop-in counties and examining their deviation from their county ideals. TrJ1872-73. To do this, he calculated a new ideal for all districts outside of drop-in counties, which he said would take out the possibility that data for these districts are being obscured by drop-in county districts. Id. He found that minority-majority CVAP districts are underpopulated by 172 persons and majority-Anglo districts are overpopulated by 91 persons. TrJ1873-74; D-168 at 24. Using SSVR, he found that SSVR-majority districts are overpopulated by 80 persons while districts below 50% SSVR are underpopulated by 21. TrJ1874; D-168 at 24. Second, Alford analyzed deviations in all drop-in counties as a collective whole by using adjusted countywide ideals. TrJ1875. He found on average that minority-majority-CVAP districts are underpopulated by 149 persons and Anglo-CVAP-majority districts are overpopulated by 141 persons. TrJ1876; D-168 at 25. He further found that the average SSVR-majority district was 626 128 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 129 of 171 persons underpopulated and the average <50% SSVR district was 209 persons overpopulated. Id. Crucially, however, Alford did not analyze any individual drop-in county, nor did he look for discriminatory patterns in particular areas.78 TrJ1900. This deficiency in Alford’s analyses makes his results consistent with Arrington’s general view of Plan H283—when viewed in light of the County Line Rule, Plan H283 may not exemplify a statewide, systematic pattern of overpopulation of minority districts, but there may nonetheless be individual geographic areas within the Plan where the one person, one vote principle was violated. TrJ1880 (Alford); TrJ1821 (Arrington). Kousser went further, testifying that deviations and different treatment of urban Democrat districts—particularly Latino ones—simply cannot be explained by partisanship or the County Line Rule. Tr235-37. Martin’s testimony also indicates the County Line Rule cannot explain many of the county-level deviations in Plan H283. According to Martin, there are 14 counties with two or more whole districts contained within them. Tr379. Of the 150 districts in the Plan, these 14 counties account for 93 of the districts. Tr380. Martin testified that each of these 93 districts could have been drawn to zero deviation, but that instead, eight of the 14 counties (accounting for 75 districts) have a deviation over 3.5% from the county ideal. Tr380-81. Implicit in this analysis is that each of these 93 districts could have been drawn without any deviation and without violating the County Line Rule because these districts are wholly contained within a county. According to Martin, there was never any other justification given for the deviations in Plan H283. Tr381. Aside from the broad expert testimony that reveals flaws in Defendants’ reliance on the 78 The Court finds this particularly important because different mapdrawers were involved in drawing the different challenged areas. 129 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 130 of 171 County Line Rule to justify deviations, a close-up look at its applications in several individual areas (as discussed below) shows that it usually is not the cause of the challenged deviations and therefore cannot justify them. 3. Avoiding contests between incumbents and protecting incumbents In Karcher, the Supreme Court specifically identified “avoiding contests between incumbent Representatives” as a “legislative polic[y] [that] might justify some variance,” so long as it was consistently applied and nondiscriminatory. Karcher, 462 U.S. at 740; see also LULAC v. Perry, 548 U.S. 399, 412 (2006) (recognizing that avoiding the pairing of incumbents has been recognized as a neutral redistricting standard). This purported justification was at the heart of Larios; other than this factor, none of the other policies identified in Karcher were at issue. Larios, 300 F. Supp. 2d at 1331. Looking to the numbers, the Larios court pointed out that in the House plan, “forty-seven incumbents were paired, including thirty-seven Republicans, which was 50% of the Republican caucus, but only nine Democrats, comprising less than 9% of that caucus (as well as one Independent [who was considered a Republican]).” Id. at 1329. Six of the 21 districts involved in the pairing of incumbents were multi-member districts, meaning that 28 of the paired incumbents could be reelected while the other 19 would be unseated. Id. In the Senate plan, there were six incumbent pairings—four were Republican–Republican and two were Republican–Democrat. Id. The result of these pairings was the lost seats of 18 House Republican incumbents (compared to three Democrats) and four Senate Republican incumbents (compared to zero Democrats). Id. at 1329-30. In addition to the raw numbers, the Larios court looked to the shapes of the districts, noting that they were drawn to force Republicans to run against one another, along with the process for 130 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 131 of 171 obtaining such a design: “[I]ndividual Democratic incumbents negotiat[ed] with the plans’ drafters to draw them the safest possible districts.” Id. at 1330. Not only did Democratic incumbents take these measures to secure their own seats, but they also “targeted particular Republicans to prevent their re-election.” Id. In conclusion, the court tied these incumbency pairing measures together with the population deviations, noting that Republican-leaning districts are vastly more overpopulated as a whole than Democratic-leaning districts. Indeed, by one measure, the House Plan contains fifty overpopulated and thirteen underpopulated Republican-leaning districts, compared to only twenty-two overpopulated and fifty-nine underpopulated Democratic-leaning districts, and the 2002 Senate Plan contains nineteen overpopulated and seven underpopulated Republican-leaning districts, compared to only eight overpopulated and twenty-two underpopulated Democratic-leaning districts. Moreover, as we have noted, the large positive deviations often occurred in districts that paired Republican incumbents. Id. at 1331 (internal citations omitted). In part, this feature of Larios has been left open; as the court there stated, “[w]e need not resolve the issue of whether or when partisan advantage alone may justify deviations in population, because here the redistricting plans are plainly unlawful . . . [P]artisan interests are bound up inextricably with the interests of regionalism and incumbent protection. It is simply not possible to draw out and isolate the political goals in these plans from the plainly unlawful objective of regional protection or from the inconsistently applied objective of incumbent protection.” Id. at 1352. In Harris, the Court refused to answer a similar question of the permissible use of politics. 136 S. Ct. at 1310. With this guidance in mind, an extensive analysis of incumbency pairing and protection is not necessary here, nor is an ultimate answer on the question of what role “politics as usual” may have played. The role of incumbency pairing in this case is much smaller than in Larios, where it 131 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 132 of 171 was hotly contested and played a large role in the statewide outcome. Here, Plan H283 paired fourteen incumbents, only two of whom were Democrats. This is a far cry from the systematic level of pairing mostly Republican incumbents in Larios. Solomons set the standard early in the process that only members of the same party would be paired, and this rule was consistently followed. Tr1427 (Interiano). The fact that only two Democrats were paired is due to the operation of the County Line Rule and redistricting leadership’s views concerning whether districts were protected by the VRA. There is no indication that population deviations were used to target Democrat incumbents. However, Defendants also have not offered any evidence showing that any particular population deviations challenged by Plaintiffs were the result of avoiding incumbent pairings. In short, because neither side has presented evidence that mapdrawers created deviations to avoid the pairing of incumbents, this factor has little evidentiary value. 4. Compliance with the VRA Aside from broadly stating the applicable burden of proof in challenges to plans where population deviation is below 10%, Harris provides an analysis of whether and how attempts to comply with the VRA can justify deviations. There, a three-judge district court panel found that the Arizona plan’s 8.8% deviation was “primarily a result of good-faith efforts to comply with the Voting Rights Act . . . even though partisanship played some role.” Harris, 136 S. Ct. at 1306 (quoting the district court panel opinion, 993 F. Supp. 2d 1042, 1046 (D. Ariz. 2014)). The Supreme Court affirmed, with the conclusion that the record showed that Arizona’s mapdrawers made a goodfaith attempt at compliance with the VRA, which is a “legitimate state consideration that can justify 132 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 133 of 171 some deviation from perfect equality of population.”79 Id. When creating the Arizona plan, an independent five-member redistricting commission created by the Arizona Constitution started with a grid-patterned map with a total deviation of 4.07%. Id. at 1308. After reviewing the grid and consulting with several legal and mapdrawing experts, all five Commissioners agreed that the plan would need to by precleared by the Justice Department under the VRA, and that the former benchmark plan contained 10 ability-to-elect districts. Id. After conducting an “iterative process” in which the Commissioners strengthened minority voting strength by adjusting the initial grid’s boundaries, they published a new plan with 10 ability-to-elect districts for public comment. Id. Around the same time, however, the Department of Justice indicated that it might not view the new plan as containing 10 ability-to-elect districts, but per normal DOJ protocol, it did not specify precisely how it would calculate ability-to-elect districts. Id. The calculations behind this determination were complex and in part unpredictable. Id. “The upshot [of the DOJ’s process and calculations] was not random decision-making but the process did create an inevitable degree of uncertainty.” Id. 79 In finding that attempted VRA compliance was a “legitimate consideration” that might justify population deviations, the Court stated: In cases decided before Shelby County v. Holder, 570 U.S. –, 133 S. Ct. 2612 (2013), Members of the Court expressed the view that compliance with § 5 of the Voting Rights Act is also a legitimate state consideration that can justify some deviation from perfect equality of population. See League of United Latin American Citizens v. Perry, 548 U.S. 399, 518 (2006) (Scalia, J., concurring in judgment in part and dissenting in part, joined in relevant part by Roberts, C.J., Thomas & Alito, JJ.); id., at 475, n. 12, (Stevens, J., concurring in part and dissenting in part, joined in relevant part by Breyer, J.); id., at 485 n. 2 (Souter, J., concurring in part and dissenting in part, joined by Ginsburg, J.); see also Vieth v. Jubelirer, 541 U.S. 267, 284 (2004) (plurality opinion) (listing examples of traditional redistricting criteria, including “compliance with requirements of the [Voting Rights Act]”). It was proper for the Commission to proceed on that basis here. Harris, 136 S. Ct. at 1306-07 (citations altered). 133 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 134 of 171 This uncertainty, the Court said, “could lead a redistricting commission, as it led Arizona’s, to make serious efforts to make certain that the districts it believed were ability-to-elect districts did in fact meet the criteria that the Department might reasonably apply.” Id. The Commission reconsidered the boundaries of its plan, and decided to strengthen its ability-to-elect districts by making several changes. Id. Several of these changes were uncontroversial and were unanimously approved by the Commission. Id. at 1308-09. Others were not—a change to one of the boundaries was characterized as “‘hyperpacking’ Republicans into other districts” by a Republican member of the Commission. Id. at 1309. A Democrat proponent of the change replied that it could create an eleventh ability to elect district, and the change was ultimately approved. Id. When Arizona submitted this plan to the DOJ for preclearance, it cited these changes in support of its argument, even though the DOJ ultimately concluded that the changed district was not an ability-to-elect one. Id. The Supreme Court found that the district court’s conclusion that deviations were caused by good-faith attempts at compliance with the VRA was well supported in the record. Id. In sum, “‘the record b[ore] out’ that the deviations ‘predominantly reflected . . . efforts to achieve compliance with the federal Voting Rights Act, not to secure political advantage for one party.’ In other words, the plaintiffs in Harris foundered not because their one person, one vote challenge failed as a matter of law, but because they did not muster the evidence needed to show it to be ‘more probable than not that [the] deviation of less than 10% reflect[ed] the predominance of illegitimate reapportionment factors.’” Raleigh Wake, 827 F.3d at 341 (citations omitted). 5. Use of illegitimate redistricting factors Because the burden of proof on these claims still rests on Plaintiffs to affirmatively show that 134 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 135 of 171 it is more probable than not that a deviation of less than 10% reflects the predominance of illegitimate reapportionment factors rather than legitimate considerations, the Court must assess the use of illegitimate factors as well as the legitimate ones discussed above. See Harris, 136 S. Ct. at 1307. In short, Plaintiffs argue that the State intentionally discriminated against minority voters by overpopulating minority districts and underpopulating Anglo districts. Such a practice is, of course, an illegitimate redistricting factor. See Gaffney, 412 U.S. at 751. 6. Analysis i. Nueces County Nueces County is a drop-in county with two districts, HD32 and HD34. Under 2010 census figures, it was apportioned 2.0295 districts based on the statewide ideal. Because the County had more than enough population for two districts, mapdrawers could have faithfully applied the County Line Rule and drawn two districts wholly contained in Nueces County, each with roughly equal population that was only slightly above the statewide ideal. In fact, mapdrawers did so in the initial versions of the House plan, including in the first public version Plan H113. In Plan H113, HD32 and HD34 were roughly equally overpopulated—HD32 (Hunter’s district) was 1.61% overpopulated and HD34 (the Latino opportunity district) was 1.34% overpopulated. Yet further changes were made that resulted in HD32 being underpopulated by 563 persons (a deviation of -0.34% from the statewide ideal) and HD34 being overpopulated by 5,512 (a deviation of 3.29% from the statewide ideal), nearly ten times the size of the underpopulation in HD32.80 As the numbers reflect, HD34 is significantly more Hispanic in population and significantly 80 Between Plan H113 and H153, HD34 went from 1.34% overpopulated to 2.89% overpopulated; from Plan H153 to Plan H283, overpopulation in HD34 rose further still to 3.29%. 135 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 136 of 171 more overpopulated than HD32, which is significantly more Anglo and is underpopulated, despite the fact that there is enough population in Nueces County for both districts to be overpopulated.81 Based on this evidence of the use of race in drawing the lines in Nueces County to overpopulate the minority district while underpopulating the Anglo district, the Court finds that Plaintiffs have proven Larios claims here. ii. Hidalgo County/HD41 The County Line Rule also does not explain the population deviations in the Hidalgo County districts. Hidalgo County was not a drop-in county because it had surplus population. In Plan H283, four districts are wholly contained within Hidalgo County—HDs 36, 39, 40, and 41. Three of these four districts are overpopulated—HD36 has a 2.61% deviation (4,368 persons), HD39 has a 4.62% deviation (7,746 persons), and HD40 has a 3.49% deviation (5,856 persons). The fourth, HD41, is underpopulated by 4.41% (-7,399 persons), even though Hidalgo County is overpopulated to the point where it must shed excess population into HD31, which includes Starr and Zapata Counties to the west. A comparison of the racial makeup of the underpopulated HD41 to the three overpopulated districts in Hidalgo County reveals that HD41 is significantly less Hispanic than the other three districts. The overpopulated districts all have an HCVAP above 80% (HD36 is 88.7%, HD39 is 82.4%, and HD40 is 89%), while the underpopulated HD41 is 72.1% HCVAP. The difference is 81 The only potential explanation given by Defendants concerning population disparities in Nueces County is that mapdrawers sought to draw HD34 as a district that would perform reliably for Latino voters. However, that justification is demonstrably not the reason for the population disparity between HD32 and HD34 in Plan H283. When mapdrawers initially drew the two Nueces County districts with roughly equal population, David Hanna assessed HD34 as “clearly performing” in his retrogression memos, and Solomons later touted HD34 in Plan H113 as being a “performing” Latino district in Nueces County. D-327; D-206. Thus, the further changes that led to the population disparities cannot be explained by needing to make HD34 a “performing” Latino opportunity district. Thus, neither the County Line Rule nor compliance with the VRA explain the population deviations in Nueces County. 136 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 137 of 171 starker in terms of total SSVR—HD36 is 85.1%, HD39 is 82.2%, and HD40 is 85.8%, while HD41 is 63%.82 As in Nueces County, the County Line Rule does not explain deviations within Hidalgo County, and Defendants have not even suggested that these deviations were necessary to comply with the Rule. Rather, these deviations disadvantage the more Hispanic districts to the benefit of the less Hispanic HD41. And the record shows that the true reason for this deviation was to ensure that Republican incumbent Rep. Peña had a favorable district (one that had more Anglo population and less Hispanic population). Incumbency protection (not necessarily avoiding the pairing of incumbents) was a large factor in Hidalgo County (particularly in HD41). As discussed above, Hidalgo County contains four whole districts, three of which are overpopulated and one of which is underpopulated (HD41). Interiano drew the initial configuration of what became HD41 in the final Plan H283 with the express goal of giving Peña the best chance to get re-elected. Although Downton made some changes to address the fact that HD41 was underpopulated, the findings of fact detail the use and purposes of these changes—particularly, the many precinct splits, some of which use race as a proxy for partisanship—and based on this evidence we reach the unmistakable conclusion that the underpopulation of HD41 was intentional, politically motivated, and explained by race. Mapdrawers’ express and admitted political goal of protecting the incumbent Peña in HD41 went too far in its use of race and partisanship. Thus, though political motivation of protecting the Republican incumbent Peña was the proffered justification for the deviations in Hidalgo County, the means of accomplishing this goal (the illegitimate use of race to split Latinos off into the 82 HD31, the district with which Hidalgo County shares its surplus population, is overpopulated by .6% (999 persons) and is 90.8% SSVR. 137 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 138 of 171 overpopulated districts of Hidalgo County while hoarding Anglos into the underpopulated HD41 and intentionally leaving it underpopulated compared to its neighboring districts to avoid adding in more Hispanics) was not a legitimate, neutrally applied redistricting policy. The underpopulation of HD41 and overpopulation of the surrounding districts was intentional, with the stated goal of maximizing Republican performance and the unstated goal of limiting Hispanic population in HD41 to get Peña reelected. Tr1478 (Interiano). Implicit in this discussion so far is that the underpopulation of HD41 is the cause of the relative overpopulation of Hidalgo County’s other districts. Of course, adherence to the County Line Rule in a county that as a whole is overpopulated (such as Hidalgo County) will necessitate some overpopulation of all districts within the county, though it will not explain underpopulation of any of them (like HD41). But the uneven application of political protection, especially when furthered by race-based precinct splits that fit the larger allegation of burying Latino voters into overpopulated districts, is not a legitimate reapportionment policy that can justify the county-level deviations between HD41 and its surrounding districts. Accordingly, the purported use of incumbency protection in HD41 cannot justify the overpopulation of the surrounding districts in Hidalgo County. Defendants argue that population deviations in Hidalgo County are justified when viewed in terms of different population metrics. They point out that the VAP of the districts in Hidalgo County is as follows: 104,939 (HD31), 111,643 (HD36), 114,761 (HD39), 112,156 (HD40), and 111,689 (HD41, Peña’s district). Docket no. 457 at 41–42. They further point out that the CVAP of these districts is as follows: 56,380 (HD31), 67,940 (HD36), 74,275 (HD39), 69,900 (HD40), and 79,770 (HD41, Peña’s district). Id. at 42. Based on these figures, HD41 has the highest CVAP of all districts in Hidalgo County, and ranks in the middle in terms of VAP. Thus, Defendants argue, 138 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 139 of 171 whatever disparity exists among the total population in the five Hidalgo County House districts, it does not diminish the voting strength of voters outside HD41. Id. at 42. This argument is nothing more than an after-the-fact attempt to justify population deviations that were illegitimately drawn into Hidalgo County in the first place. Nothing in the record indicates that mapdrawers relied on VAP or CVAP to achieve population equality in Plan H283. Rather, it is undisputed that they measured their compliance with one person, one vote requirements by total population, and they cannot now use VAP or CVAP numbers to make post-hoc arguments that no violation occurred. The argument further misses the point because it does not explain how or why the population deviations exist. An emphasis on high VAP and CVAP in HD41 is not an explanation for the deviation but rather stems from the same common cause of the deviations—the admitted goal of intentionally underpopulating HD41 to create a less-Hispanic, Republican district that would ensure Peña’s reelection. A person with knowledge of the area would know how to do this by splitting precincts, and the precinct splits in HD41 reflect precisely this goal with the added side effect of increasing VAP and CVAP. Therefore, despite Defendants’ attempt to minimize the harm from deviations in HD41, the population deviations in Hidalgo County are not the result of any legitimate redistricting considerations, and instead were driven only by the partisan objective of reelecting Peña, which was achieved in part through the use of race. Absent the use of any other legitimate redistricting principles, these population deviations, like those in Larios, result from the “inconsistently applied objective of incumbent protection.” iii. Dallas County The 8.88% top-to-bottom deviation within Dallas County cannot be explained by the County Line Rule. Dallas County’s population decreased compared to the rest of the state, which resulted 139 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 140 of 171 in Dallas County having enough population for 14.1266 districts, such that all districts in Dallas County theoretically could have been overpopulated. That was not the case in Plan H283. Both Latino districts in Dallas County are overpopulated—HD103 deviates from the statewide ideal by 5% (8,379 persons) and HD104 deviates by 3.07% (5,147 persons). One of the four African-American districts is overpopulated—HD109 deviates from the statewide ideal by 3.9% (6,539 persons)—while the other three are underpopulated—HD100 is -3.87% (-6,494 persons), HD 110 is -0.05% (-90 persons), and HD111 is -0.39% (-658 persons). Therefore, three of the six minority districts in Dallas County are overpopulated, and three are underpopulated. This same ratio of overpopulated-to-underpopulated districts was true for Anglo districts in Dallas County, as four are overpopulated and four are underpopulated. But the underpopulation of Anglo districts in Dallas County was more severe,83 while the overpopulation was not as severe.84 Based on these numbers alone, minority voters in Dallas County fared worse than Anglo voters. This is especially true when looking specifically at Latino districts—HD103 was overpopulated by 5% and HD104 was overpopulated by 3.07%. The potentially problematic deviations in Dallas County are confined to the interplay between the two Latino districts (HDs 103 and 104) and the Anglo district to their west, HD105. The two Latino districts are overpopulated by an average of 4.04% from statewide ideal (HD103 by 5.00% and HD104 by 3.07%). HD105 is also severely overpopulated at 4.83%. As the following will make clear, the composition and overpopulation in these western Dallas County districts is mostly caused 83 The average deviation in an underpopulated minority district in Dallas County is -1.44%, while the average deviation of an underpopulated Anglo district in Dallas County is -1.85%. 84 The average deviation in an overpopulated minority district in Dallas County is 3.99%, while the average deviation of an overpopulated Anglo district in Dallas County is 3.10%. 140 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 141 of 171 by shifts of population among the three districts. Thus, it appears that the rest of Dallas County is not problematic, and simply does not match the Plaintiffs’ broad theory that minorities were systematically disadvantaged by being crowded into overpopulated districts.85 Kousser opined that HD103 and HD104 (the two overpopulated Latino districts) were overpopulated to “pack” as much Latino population as possible into those districts, making it much more difficult to draw additional opportunity districts for the minority population growth, which accounts for 100% of Dallas County’s 350,000 population growth (in fact, Anglo population decreased by 198,000 in Dallas County). Anchia, HD103’s representative, was himself displeased with the overpopulation of his district, and felt that there was an effort to put a lot of unnecessary population into the district. Arrington made a similar observation, noting that Hispanic districts are overpopulated, which makes it more difficult to create another minority district. TrJ222. Martin’s analysis was more technical, pointing out that the 8,091-person excess in HD105 (4.83% overpopulation) is Anglo population drawn from the south, which makes the district less likely to elect Latino candidates of choice. TrJ330-32. The only explanations given for this strange configuration of west Dallas County are that mapdrawers needed to pair Rep. Anderson with Rep. Harper-Brown and needed to maintain SSVR in HD103 and HD104. TrJ2018 (Downton). The map of these western districts in Dallas County supports the experts’ conclusions, as described in the discussion of Plaintiffs’ vote dilution claims above. The boundary of HD105 divides the cities of Grand Prairie and Irving, and splits 22 precincts along with numerous 85 The numbers bear out this conclusion. Not counting the two Latino districts that are subject to closer scrutiny, the average minority district in Dallas County was actually underpopulated at -.10% deviation from statewide ideal while the average Anglo district (not counting HD105) was overpopulated by .03%. Thus, because overpopulation in the Anglo districts is more severe than in the minority districts, Plaintiffs have not shown that Dallas County as a whole uses population deviations to systematically disadvantage minority voters. 141 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 142 of 171 communities of interest. Tr321-23 (Martin); US-387 at 7; US-299. It represents a drastic change from the benchmark, extending much further out with jagged, bizarrely shaped lines. TrJ1107-08 (McPhail). The portion of HD105 that reaches into Grand Prairie picks up “the most Anglo part of the city,” follows the “least Hispanic channel that could have been drawn,” and leaves seven split Latino precincts in HD104. Tr321-22 (Martin). These precinct splits placed a disproportionately affluent Anglo population in HD105 and a disproportionately low-income Hispanic population in HD104. E.g., TrJ151 (Arrington); TrJ599-609 (Lopez). HD103 contains a long, narrow arm of mostly Latino population that extends through HD105 in Irving and splits 10 precincts, removing all of this Hispanic population from HD105. Tr323 (Martin). Downton deliberately split precincts by using race to put Hispanic voters in HDs 103 and 104 while keeping them out of HD105. TrJ2020-24 (Downton). Increasing the Anglo population of HD105 prevents it from providing minority opportunity. The Perez Plaintiffs presented demonstration Plan H288, with a total deviation of 1.72% and which maintained HD105 with a combined B+HCVAP of 45.8% using 2005-2009 ACS data and 50.1% using 2008-2012 ACS data, while otherwise containing the same number of BCVAP-majority (3) and HCVAP-majority (1) districts as Plan H283 using 2008-2012 ACS data. Joint Ex. J-33; D-110. Although Downton indicated in his deposition that it would be hard to maintain the SSVR of HD103 at closer to ideal population, he agreed at trial that it was not necessary to have a 5% deviation in HD103 to maintain the Hispanic population, and that there were others ways to have done it, “but that was a policy decision.” TrJ2149. Maps such as Plan H288 show that the SSVR of HD103 could be maintained or even increased without such overpopulation. Joint Map. Ex. J-33 (HD103 is 38.7/40.6% SSVR while only 1.31% overpopulated). 142 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 143 of 171 Further, between Plan H113 and Plan H283, the total population was increased, but the nonsuspense SSVR (available to mappers in RedAppl) was not. The overpopulation increased from 4.59% to 5%, while non-suspense SSVR remained at 38.1% (total SSVR went up .1%). D-112 at 10, 30. Downton had also testified that HD103 was overpopulated because he worked with Rep. Anchia to include that population, but Rep. Anchia was not happy that his district was overpopulated and felt that extra population had been included unnecessarily. Tr999; TrJ2017. Thus, none of Downton’s explanations for the overpopulation of HD103 are credible. These facts are compelling. As indicated in the discussion on Plaintiffs’ § 2 claims, the mapdrawers intentionally diluted minority voting strength in western Dallas County in part by manipulating total population. But Plaintiffs overlook that the net result of this manipulation does not manifest itself in racially inequitable deviations among the three districts that were trading population for one person, one vote purposes—the Anglo HD105 deviates from the statewide ideal by 4.83%, which is more than the average deviation in HD103 and HD104. While the lines were drawn to push Latino population into the Latino districts, they were also drawn to push Anglo population into HD105. This practice in itself, as discussed above, may be illegal in other ways. But in terms of overpopulation, Anglos in HD105 fared just as poorly as did the Latinos of HDs 103 and 104. In short, the overpopulation in western Dallas County may be an intentional vote dilution tactic, but when it is viewed in light of the relationship between the districts that caused this overpopulation, the deviations themselves are not the basis for a one person, one vote claim.86 86 The Court notes a tension between the standard for a § 2 intentional vote dilution claim and a Larios claim, which helps explain why Plaintiffs succeed on the former but fail on the latter in western Dallas County. To succeed on a § 2 intentional vote dilution claim, Plaintiffs have the burden of showing that discriminatory purpose is a “motivating factor”; they need not show that race was the sole decisionmaking factor or motivation. Vill. of Arlington Heights v. Metro. Hous. Dev. Corp., 429 U.S. 252, 265-66 (1977); Veasey v. Abbott, 830 F.3d 216, 230 (5th Cir. 2016). To succeed on a Larios claim, Plaintiffs have the burden of showing that illegitimate reapportionment factors 143 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 144 of 171 Accordingly, Plaintiffs have not shown that Dallas County’s population deviations reflect the predominance of illegitimate redistricting factors. iv. Harris County As in Dallas County and several others, the Court first dispels the assertion that the County Line Rule explains deviations in Harris County. Overall deviation among districts within Harris County is 9.74%, but a map could have been drawn with only 1.81% deviation (this would have been easy, according to Martin). Tr356-58. Because the Legislature chose to go with 24 districts, there was enough population to draw all districts at or above the statewide ideal. Despite the ability to overpopulate all districts, some districts were underpopulated and some were overpopulated.87 Of the 24 districts in Harris County, 13 are Anglo majority and 11 are majority minority.88 Of the 13 Anglo districts, 10 were overpopulated while three were underpopulated. Of the 11 minority districts, eight were overpopulated while three were underpopulated. Thus, the percentage of Anglo districts that were overpopulated (10/13, or approximately 77%) was actually greater than the percentage for minority districts (8/11, or approximately 73%). “predominated” in bringing about the challenged population deviations. Harris, 136 S. Ct. at 1307. Of course, one such illegitimate reapportionment factor would be the improper use of race. Thus, a plaintiff challenging population deviations through a Larios-type one person, one vote claim would have to show the predominance of illegitimate reapportionment factors (including race), while a plaintiff bringing a § 2 intentional vote dilution claim would only have to show that a racially discriminatory purpose is a “motivating factor” because the VRA is concerned with racial motives and not others. Section 2 of the VRA is not concerned with whether a district is over- or under-populated per se, but whether its configuration was produced at least in part by a racially discriminatory/dilutive motive. In contrast, the harm in a one person, one vote context is being in an overpopulated district. In short, a Larios plaintiff can rely on evidence of a broader array of redistricting practices to reach the higher predominance standard; a § 2 intentional vote dilution plaintiff has to present evidence limited to racially discriminatory purpose, but need only satisfy the lower “motivating factor” standard. 87 This analysis of the County Line Rule does, however, belie MALC’s assertion that “[t]here is no explanation as to why three out of 4 majority Latino districts in Harris County are overpopulated.” Docket no. 1185 at 9. Because Harris County was drawn down to 24 districts, there was sufficient population in the County to overpopulate all districts. 88 HD148 and HD137 were both non-majority HCVAP, but were performing for Latinos and minorities. For purposes of this analysis, they are counted among the 11 minority districts. 144 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 145 of 171 The average Anglo district is overpopulated by 1.51%, while the average minority district is overpopulated by 1.97% compared to the statewide ideal. Based on the county ideal,89 the deviations appear smaller, though the difference between average deviation of an Anglo district (.21%) and that of a minority district (.25%) is roughly the same. These numbers take into account the three underpopulated Anglo and three underpopulated minority districts. Looking only at these underpopulated districts, underpopulated minority districts (-2.86% from statewide ideal) are actually more underpopulated than their Anglo counterparts (-2.09% from statewide ideal). The raw data in Harris County points in conflicting directions. The likelihood that an Anglo district will be overpopulated is higher than the likelihood that a minority district will be overpopulated. An average minority district, however, will have more overpopulation than an average Anglo district. Both statistical measures of the overpopulation in Harris County are close—the difference between the occurrence of overpopulation in a district of each type is only about 4%, and the difference in overpopulation between an average district of each type is only about .46%. Considering these numbers, Plaintiffs have not shown that it is more probable than not that the deviations reflect the predominance of illegitimate reapportionment factors rather than legitimate considerations. The difference in overpopulation between the average minority district (1.97%) and the average Anglo district (1.51%) is slight. A difference so slight may be the result of many incidental or legitimate reapportionment criteria. The Plaintiffs have simply not provided evidence to affirmatively state which illegitimate factors were used in drawing the Harris County lines, and 89 The total population of Harris County’s districts (4,092,459) divided by the number of districts in the County (24) yields the county ideal—170,519. 145 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 146 of 171 how deviations so small among the Anglo and minority districts resulted from these factors predominating in the mapdrawing process. At most, Plaintiffs have provided circumstantial evidence in the form of a pattern of overpopulation in minority districts in Harris County. But that evidence is particularly weak given the small difference in overpopulation between Anglo and minority districts and the higher percentage of overpopulated Anglo districts. Accordingly, Plaintiffs have failed to carry their burden. v. Tarrant County Application of the County Line Rule cannot explain the deviations within Tarrant County. Tarrant County contains the smallest district in the entire plan—HD90, with a deviation from the statewide ideal of -4.90% (8,209 persons). Despite this fact, the Court does not find that the population deviations within Tarrant County support Plaintiffs’ one person, one vote claim. The Court has found that redistricters increased the SSVR in HD90 in bad faith, but doing so required them to underpopulate HD90. Further, the minority districts in Tarrant County are all underpopulated (HD90 is -4.90%, HD95 is -3.58%, and HD101 is -1.77%), and HD90 and HD95 are the most underpopulated in the County, while two Anglo districts (HD97 and HD99) are overpopulated. The one person, one vote challenge in Tarrant County fails. vi. Fort Bend County Unlike in many of the other counties, where the County Line Rule does not explain the existence of deviations within a county, Fort Bend County’s relatively small deviations can be explained by the mapdrawers’ need to work within the confines of the County Line Rule. Suburban Fort Bend County experienced significant growth, so mapdrawers turned the County from two wholly contained districts with some surplus into three wholly contained districts with some surplus. 146 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 147 of 171 This spillover was joined with Wharton and Jackson Counties in HD85. The districts in Fort Bend County deviate very little from one another—HD26 deviates from the statewide ideal by -4.50% (7,546 persons), HD27 deviates by -4.51% (7,553 persons), and HD28 deviates by -4.33% (7,264 persons). Within the confines of the County Line Rule, the underpopulation of these districts was apparently caused by a legitimate need to increase the size of the spillover to combine it with two other counties to form a new district. In addition, HD27, the minority district, is the most underpopulated. These districts do not manifest problematic population deviations. Plaintiffs have not shown that it is more probable than not that the population deviations in Fort Bend County (and specifically, the challenged HD26) reflect the predominance of illegitimate reapportionment factors rather than legitimate considerations. The one person, one vote challenge to Fort Bend County fails. vii. Bell and Lampasas Counties The Court next looks to Plaintiffs’ challenges to Bell and Lampasas Counties and their two districts, HDs 54 and 55. In the prior plan, HD54 combined all of Burnet County, all of Lampasas County, and a small extension in the western side of Bell County. The extension into Bell County contained almost all of the city of Killeen, which previously contained only a minor split (separating only 200 residents from the city). HD55 was composed of the remainder of Bell County. In Plan H283, Burnet County is removed from HD54 and no longer combines with any part of Lampasas or Bell Counties. The departure of Burnet County’s population left HDs 54 and 55 without sufficient population to fill two full districts, and both could have been underpopulated compared to the statewide ideal. Yet HD54 (Aycock’s district), which combines all of Lampasas County with the western section of Bell County, is overpopulated compared to the statewide ideal 147 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 148 of 171 by 0.06%, while HD55 (the remainder of Bell County) is underpopulated at -3.26%. A comparison of the prior and current maps reveals that it would have been relatively easy to draw HD54’s extension into Bell County in a way that left both HDs 54 and 55 equally underpopulated, but this was not done. The changes in Plan H283 that created the existing deviations are problematic in terms of the vote dilution that they cause (which has been previously discussed) and the population deviations that they create (which are discussed below). With Burnet County removed, HD54 lost 42,000 voters, and went from 29,000 overpopulated to 13,000 underpopulated. TrJ1402 (Korbel). Mapdrawers then removed more population from HD54. To do this, they enlarged what was previously a minuscule, 200-resident split in the city of Killeen. The new split removed a larger, mostly minority segment of Killeen’s population—32,000 persons, about two-thirds of whom were minorities—from HD54 and placed it into HD55. TrJ1404 (Korbel). At the same time, Plan H283 made a change in the other direction by removing 47,000 mostly Anglo persons from southwest Bell County (formerly in HD55) and placing them in HD54. TrJ1405 (Korbel). The net result was a 15,000 population shift from HD55 to HD54, with HD54 becoming more Anglo by gaining residents from southwest Bell County and less minority by splitting Killeen. In terms of population deviation, these changes left HD54 at 0.06% overpopulation and HD55 at -3.26% underpopulation, even though both could have easily been underpopulated. The Court has already recounted both the goal of these changes and the means for achieving this goal. In short, the goal was to draw two Republican districts, but the means for accomplishing this goal was splitting minority population in Killeen and making Aycock’s district more Anglo. And, as previously detailed, this constituted evidence of intentional vote dilution under § 2 of the 148 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 149 of 171 VRA and the Fourteenth Amendment. Aside from the intentional vote dilution caused by the population shifts among HDs 54 and 55, these same shifts also caused population deviations in contravention of the one person, one vote principle. Because both districts could have been underpopulated compared to the statewide ideal, the one person, one vote principle, when combined with the County Line Rule, counsels that any deviations between the districts should be justified by legitimate redistricting criteria. Here, however, the motives for changes that caused the deviations were exceedingly political and racial. In order to ensure that both districts remained Republican, mapdrawers traded 47,000 mostly Anglo voters in HD55 for 32,000 mostly minority voters in HD54 by splitting Killeen. In doing so, mapdrawers intentionally used race in a way that would overwhelm the remaining Latino voters of HD54 with the new influx of Anglo voters while also stranding a large portion of Killeen’s minority voters in the already heavily Anglo HD55. As a result, mapdrawers exacerbated the existing population deviations, which could have easily been remedied. In light of the overtly political goals, the use of race, and the absence of any other justification for the population deviations, the Court finds that deviations here resulted from the predominance of the illegitimate use of race to ensure that both districts remained Republican. viii. Statewide Claims In sum, the Plaintiffs have shown by a preponderance of the evidence that population deviations in the following areas reflect the predominance of illegitimate reapportionment factors: Nueces County, Hidalgo County, and Bell/Lampasas Counties. Plaintiffs have not shown the same in the following areas: Dallas County, Harris County, Tarrant County, Fort Bend County, and everywhere else in the state. 149 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 150 of 171 The relative failure of Plaintiffs’ localized Larios claims obviates the need for a particularly detailed analysis of their statewide challenge to Plan H283 as an undifferentiated whole. Areas in which Plaintiffs’ Larios-type, one person, one vote claims succeeded directly implicate only nine districts out of the 150 in Plan H283. As described in detail above, though Defendants have not affirmatively explained many deviations in the remaining areas of the state, Plaintiffs have likewise failed to do so in those areas. Broadly, the evidence shows that deviations exist, and that mapdrawers worked under the mistaken belief that these deviations could exist without justification. Plaintiffs point to a conglomerated view of population deviations across the state and encourage the Court to draw the inference that these deviations must have been caused by mapdrawers’ desire to harm minority voters. Otherwise, the Plaintiffs point to little specific evidence in the record in support of their claims. Though the Court recognizes that such circumstantial evidence may be used to prove a violation of the one person, one vote principle, as was done in Larios, the evidence here simply is not strong enough to show the statewide predominance of illegitimate factors in creating the population deviations in Plan H283, except for in the few isolated areas described above. Plaintiffs’ statewide Larios challenge to Plan H283 fails. V. The Court turns briefly to address a fundamental flaw in Judge Smith’s dissent. His analysis conflates the standards for Shaw-type racial gerrymandering claims and intentional vote dilution claims into one test—“race for its own sake”—derived from the Supreme Court’s recent decision in Bethune-Hill. However, in first recognizing the Shaw-type racial gerrymandering claim, the Supreme Court cast it as “analytically distinct” from a vote dilution claim. Shaw v. Reno, 509 U.S. 150 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 151 of 171 630, 652 (1993). This view was reaffirmed in Miller v. Johnson, 515 U.S. 900, 911 (1995), when the Court stated that “Shaw recognized a claim ‘analytically distinct’ from a vote dilution claim.” The parties have recognized this distinction throughout this case, as this Court noted in its congressional plan opinion. Docket no. 1339 at 122 (“All parties agree that an intentional vote dilution and a Shaw-type racial gerrymandering case are analytically distinct.”). This Court also recognized in its opinion on the congressional plan that the Supreme Court recently reaffirmed in Alabama LBC that a Shaw-type racial gerrymandering claim is “that race was improperly used in the drawing of the boundaries of one or more specific electoral districts.” Docket no. 1339 at 30 (quoting Alabama Legislative Black Caucus v. Alabama, 135 S. Ct. 1257, 1265 (2015) (emphasis omitted)). The plaintiff’s evidentiary burden in a Shaw-type claim is “to show, either through circumstantial evidence of a district’s shape and demographics or more direct evidence going to legislative purpose, that race was the predominant factor motivating the legislature’s decision to place a significant number of voters within or without a particular district.” Alabama LBC, 135 S. Ct. at 1267. This Court further acknowledged that Bethune-Hill v. Virginia State Board of Elections, 137 S. Ct. 788 (2017), reaffirmed the basic racial predominance analysis explained in Miller and Shaw II, and the basic narrow tailoring analysis explained in Alabama LBC. Docket no. 1339 at 30 n.26. Bethune-Hill considered only Shaw-type racial gerrymandering claims, not vote dilution claims. The Supreme Court developed the predominance test to determine whether voters were being placed into districts on the basis of race as opposed to traditional districting principles. Although the parties have argued over whether certain language from the Shaw line of cases is relevant to and can be applied to the vote dilution analysis, no party, witness, or expert has asserted that the 151 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 152 of 171 predominance test applies to Plaintiffs’ vote dilution claims. Defendants argue that racial discrimination must have been a cause in fact of the legislative action, but they agree with Plaintiffs that the test is whether racial discrimination was a substantial or motivating factor behind enactment of the law. Docket no. 996 (Defendants’ Motion for Summary Judgment) at 5-6, 12 (“The question is whether intentional racial discrimination was a substantial or motivating factor in the enactment of the . . . plans and, if it was, whether those plans would have been enacted in the absence of a discriminatory purpose.”). In the congressional plan opinion, this Court agreed that “statements made about determining motive in the Shaw-type cases can have application to intentional vote dilution cases, even though Shaw-type cases involve a different motive than intentional vote dilution cases” because “[w]hether partisanship provides an explanation for the specific lines that were chosen is relevant to both motive inquiries.” Docket no. 1339 at 122-23. However, while the issue of race predominance must be determined in a Shaw-type claim, intentional vote dilution may be found when race motivated the districting decision, along with other factors. Regardless of whether political motives also played a part or even dominated, if racial discrimination was a motivating factor for the legislative action and the other criteria for an intentional vote dilution claim are satisfied, the districting decision violates the Constitution. VI. Summary With regard to the § 2 results claims, the Court finds that remedied claims are moot (specifically in El Paso, Bexar, and Harris Counties). With regard to non-remedied claims, the Court finds that Plaintiffs have either failed to prove a § 2 results claim as to Plan H283 or that such claims are best resolved in the 2013 plan case. 152 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 153 of 171 With regard to the intentional vote dilution claims under § 2 and the Fourteenth Amendment, the Court finds that Plaintiffs proved their claims in El Paso County (HD78), Bexar County (HD117), Nueces County (the elimination of HD33 and the configuration of HD32 and HD34), HD41 in the Valley, Harris County, western Dallas County (HD103, HD104, and HD105), Tarrant County (HD90, HD93), Bell County (HD54), and with regard to Plan H283 as a whole. With regard to the Shaw-type racial gerrymandering claims, the Court finds that only the Task Force alleged such claims, and its claim against HD117 in Bexar County succeeds but its claim against HD77 and HD78 in El Paso County does not. With regard to the one person, one vote Larios-type claims, the Court finds that Plaintiffs’ statewide claims fail, but they succeed in proving violations in the Nueces County districts (HD32 and HD34), the Hidalgo County districts (HD31, HD36, HD39, HD40, and HD41), and the Bell/Lampasas County districts (HD54 and HD55). SIGNED on this 20th day of April, 2017. _______________/s/__________________ XAVIER RODRIGUEZ UNITED STATES DISTRICT JUDGE _______________/s/__________________ ORLANDO L. GARCIA CHIEF UNITED STATES DISTRICT JUDGE 153 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 154 of 171 JERRY E. SMITH, Circuit Judge, dissenting: For Texas Democrats, 2010 was a perfect storm. The nationwide Republican surge, whether attributable to the Tea Party movement or otherwise, played out on the big screen in Texas, resulting, as the majority recognizes, in a colossal increase in the number of Republicans elected to the Texas House of Representatives and unexpected wins by Republicans in the Texas Congressional delegation. That was timed ideally with the fact that the November election was followed immediately by the results of the 2010 decennial census and redistricting. The 2011 Legislature squeezed every conceivable ounce of advantage from this coincidence, making every possible effort to maximize the number of Republicans to be elected and re-elected in 2012 and beyond. This litigation followed. The core question is whether the redistricting was lawful as an exercise of political power won in fair biennial elections or, instead, was in violation of the Voting Rights Act (“VRA”) and the Constitution. At one point in their impressive opinion, the judges in the majority correctly find that Plaintiffs have failed to prove intentional vote dilution . . . . Specifically, they have failed to show that mapdrawers acted with an intent other than maintaining their Republican districts or that they used race for partisan advantage. Because of the high correlation between race and party, splitting Democrats, eliminating Democratic districts, and shoring up districts for Republican incumbents necessarily affects minorities, but that alone is not intentional vote dilution based on race. Majority op. at 66. But alas, that announcement is made only in regard to “cracking claims in northeast Dallas County.” Because it more properly should be applied to the entirety of the state House and Congressional redistricting, I respectfully dissent. Despite its exhaustive and careful examination of the record, the majority commits grave error in its recitation of redistricting law, so its findings of fact based on that misreading do not enjoy 154 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 155 of 171 the protection of the clearly-erroneous standard. And even under the clearly-erroneous test, the majority’s findings are fatally infected, from start to finish, with the misunderstanding that race, rather than partisan advantage, was the main reason for the Congressional and state house districts drawn in 2011.1 Despite its honorable intentions, the majority concocts the most extreme possible reading of the raw record to justify findings that, if converted to corresponding remedies, will hand these plaintiffs pretty much everything they have sought, causing a wholesale revision in the State House and Congressional maps. In 2011, I warned that [t]he judges in the majority, with the purest of intentions, have . . . produced a runaway plan that imposes an extreme redistricting scheme for the Texas House of Representatives, untethered to the applicable caselaw. The practical effect is to award . . . a slam-dunk victory for the plaintiffs . . . at the expense of the redistricting plan enacted by the Legislature . . . . Perez v. Perry, 835 F. Supp. 2d 209, 218 (W.D. Tex. 2011) (Smith, J., dissenting). The majority there erred. The Supreme Court promptly and unanimously vacated and remanded. Perry v. Perez, 132 S. Ct. 934 (2012) (per curiam). In the intervening five years, apparently nothing much has changed in terms of the majority’s view of who wins and loses.2 Granted, there have been extensive proceedings, resulting in a massive record, and the procedural posture is much different from what it was in 2011–12. Nonetheless, the majority’s jaundiced perspective of the Texas Legislature and its minions seems fixed. We will see 1 The exception is Part IV of the majority opinion, regarding the one-person-one-vote claims based on population deviation, with which I concur. As the majority convincingly demonstrates, in the redistricting process the State erred in assuming, without support in the law, that the ten-percent test offers an unassailable safe haven. I also agree with the ruling on standing. 2 The majority’s enthusiasm for the plaintiffs’ case is shown by the fact that, in at least one instance, it even invents a claim that the plaintiffs missed¯what it believes is “a clear Shaw-type racial gerrymandering claim.” Majority op. at 57 n.46. 155 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 156 of 171 how it fares on ultimate appeal. In this opinion, I will reiterate that because of mootness, this court is without jurisdiction in regard to the 2011 state House and Congressional plans. I will show how the majority is wrong on the law. And I will demonstrate that the majority’s factual findings are so extreme as to defy logic and reason under this record. I. Mootness deprives this panel of jurisdiction. In dissenting from the majority’s March 10, 2017, opinion on the 2011 Congressional plan,3 I showed that the proceedings for the 2011 Congressional and State House plans are moot under the binding authority of Davis v. Abbott, 781 F.3d 207 (5th Cir.), cert. denied, 136 S. Ct. 534 (2015). Nothing has changed to fix that fatal defect. Like all district courts, this court has an obligation to examine its own jurisdiction. This panel should promptly dismiss the 2011 suit for want of jurisdiction. Had it done so in 2015, when certiorari was denied in Davis, these proceedings would be much further advanced. II. The majority misreads the applicable law. Reduced to its barest essentials, the key question of law (assuming this is not moot) is whether plaintiffs have established that the legislature employed “race for its own sake.” As recently as March 2017, the Supreme Court has repeated that the answer is “no.” As I showed in my recent dissent on the 2011 Congressional maps,4 the elephant in the room is Bethune-Hill v. Virginia Board of Elections, 137 S. Ct. 788 (2017). There the Court twice repeats 3 See Perez v. Abbott, 2017 U.S. Dist. LEXIS 35012, at *243–263 (W.D. Tex. Mar. 10, 2017) (Smith, J., dissenting). 4 See id. at *266–270 (Smith, J., dissenting). 156 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 157 of 171 the test, quoting Miller v. Johnson, 515 U.S. 900, 913 (1995), that plaintiffs must prove “that race for its own sake, and not other districting principles, was the legislature’s dominant and controlling rationale.” Bethune-Hill, 137 S. Ct. at 798. “[R]ace for its own sake” must be “the overriding reason for choosing one map over others.” Id. at 799. And how do we know whether “race for its own sake” drives a redistricting decision? It occurs when “race was the criterion that, in the State’s view, could not be compromised.” Id. at 798 (citation omitted). Under Bethune-Hill, a legislative decision is suspect when it treats a person of color or of a certain ethnicity differently for no reason other than race, based on some reprehensible basis such as perceived inferiority or on a theory that persons are essentially different because of race. Thus, a retailer in the Deep South in 1960 might turn away black customers, regardless of how well they are dressed, how they behave, or how much money they wish to spend, on no basis other than race. Or, in the electoral context, a minority voter, to register, might be required to satisfy extra requirements for the sole reason of race and not because of his or her legal qualifications or how he or she is likely to vote. To be sure, there is raw “race for the sake of race” in this record, but not from state actors or their witnesses. As the majority’s findings on the 2011 Congressional redistricting state (pages 369–72), Congressman Al Green, an African-American Democrat, as well as plaintiffs’ expert Dr. Richard Murray, testified to complain that Green’s Houston-area district, CD9, was initially drawn with “a clear African-American plurality, but at the time of redistricting Hispanic total population was beginning to exceed the Black population.” In the 2011 plan, CD9 “gain[ed] heavily Latino and upscale areas” (which can only be a proxy for “white people”), “which might change the character of the district as an African-American district as the area develops.” (Emphasis added.) Those 157 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 158 of 171 witnesses said this “could create tension,” and “the addition of upscale communities . . . [in currently undeveloped plots] will likely produce an influx of Anglos . . . .” The obvious point is that the potential new Hispanic and Anglo residents were singled out because of “race for its own sake.” Congressman Green, with commendable candor, was sending the message that he did not want more Anglo or Hispanic folks in his district because, being (in his view) different on account of race, they would change the “character” of the district. One shudders to think what the reaction would be if an Anglo legislator had testified that he preferred not to have minority families moving into undeveloped parts of his district because of their “character” or essential “nature.” “Race for its own sake” was also shown in Dallas. As the majority recounts in its Congressional findings (page 375), CD30 African-American Democrat Congresswoman Eddie Bernice Johnson, also with candor, referred approvingly to Black voters who “want to be able to identify some of the people that look like them, that think like them, that live with them, to represent them.” “[W]hen they see an opportunity to represent themselves with someone who looks like them, that is what they want . . . .” That is a stark example of “race for its own sake,” but it cannot be attributed to the State defendants or their agents. Back to the law. The issue is whether the Legislature drew lines for reasons of “race for the sake of race” or, instead, for reasons of politics. The issue is complicated, for example, by whether there are limits to non-racial partisan gerrymandering and by the extent to which, if any, a given record reflects the use of partisanship as a proxy for race. To analyze this crucial issue, I rely, in part, on a new essay by Professor Richard L. Hasen of the University of California, Irvine School of Law, entitled “Race or Party, Party as Race, or Party 158 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 159 of 171 All the Time: Three Uneasy Approaches to Conjoined Polarization in Redistricting and Voting Cases.”5 Though Professor Hasen zealously advances his policy preferences as an advocate for more aggressive application of the VRA and Equal Protection Clause, his recitation of the law and his analysis of possible outcomes are resourceful and evenhanded. Professor Hasen discusses what has been called “conjoined polarization,” defined as “[t]he more consistent alignment of race, party, and ideology since 1965.”6 Importantly, he acknowledges that “legal doctrine has not yet found a comfortable way to deal with it.”7 Indeed, that is the challenge in the instant case. There are significant clues from various Supreme Court decisions, culminating, as I have said, in Bethune-Hill’s reiteration of the “race for its own sake” test. The appeal of this court’s ultimate judgment will give the Supreme Court the opportunity to be more precise in explaining¯with the 2020 decennial census on the horizon¯how much latitude legislatures have to fashion districts for purely political purposes, to take advantage of earned wins at the ballot box, where racial considerations play a part in the process and where (as always) compliance with the VRA and the Constitution is paramount. Professor Hasen presents three possible approaches that the Supreme Court could adopt: (1) “race or party”; (2) “party as race”; and (3) “party all the time.” Professor Hasen is least fond of the first, although, as I will discuss, that is the exegesis most consonant with Texas’s electoral landscape and, more importantly, is the methodology that the en banc Fifth Circuit has announced without 5 Legal Studies Research Paper Series No. 2017–18, preliminary draft, Feb. 6, 2017, forthcoming in William & Mary Law Review (2018) (symposium), abstract available at https://ssrn.com/abstract=2912403. 6 Hasen, text accompanying note 5 (quoting Bruce E. Cain & Emily R. Zhang, Blurred Lines: Conjoined Polarization and Voting Rights, 77 OHIO STATE L.J. 867, 869 (2016)). 7 Id., text accompanying notes 6–7. 159 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 160 of 171 using that nomenclature. For “race or party,” “a court’s task is to decide whether a case is ‘really about race rather than party.”8 As Professor Hasen candidly admits, if the task is to decide whether the legislature acted on the basis of race or party, redistricting plaintiffs “most importantly”9 must confront Easley v. Cromartie, 532 U.S. 234 (2001). Unaware of Professor Hasen’s writings, I nonetheless noted Cromartie’s centrality in my congressional dissent, observing that the Court found no error where “the legislature drew boundaries that, in general, placed more-reliably Democratic voters inside the district, while placing less-reliable Democratic voters outside the district.”10 That is the situation here, which Professor Hasen recognizes but this panel majority does not. He appropriately quotes (but without approval) the State’s August 5, 2013, brief in the instant case: DOJ’s accusations of racial discrimination are baseless. In 2011, both houses of the Texas Legislature were controlled by large Republican majorities, and their redistricting decisions were designed to increase the Republican party’s electoral prospects at the expense of the Democrats. It is perfectly constitutional for a Republican-controlled legislature to make partisan districting decisions, even if there are incidental effects on minority voters who support Democratic candidates.[11] Professor Hasen seems to acknowledge that, in light of Texas’s argument here, Cromartie means trouble for the instant plaintiffs. “It is not as though the Court in [Cromartie] was unaware of creeping conjoined polarization.”12 “Texas’s argument follows a statement [from] the precursor to 8 Id., text accompanying notes 7–8. 9 Id., text accompanying note 36. 10 Perez, 2017 U.S. Dist. LEXIS 35012, at *270 (Smith, J., dissenting) (quoting Cromartie, 532 U.S. at 252). Although Cromartie is a racial-gerrymandering case, it nonetheless squarely addresses the choice of “race or party” in redistricting. 11 Hasen, text accompanying note 88 (quoting State’s brief at 19). 12 Id., text accompanying note 37. 160 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 161 of 171 [Cromartie]: ‘[A] jurisdiction may engage in constitutional political gerrymandering, even if it happens that the most loyal Democrats happen to be black Democrats and even if the State were conscious of that fact.’”13 At least two other binding precedents are steep barriers to these plaintiffs, given the undeniable fact of conjoined polarization in Texas. The first is the Texas redistricting decision in Bush v. Vera, 517 U.S. 952 (1996). In what Professor Hasen calls the “principal opinion,”14 Justice O’Connor boldly stated that “[i]f district lines merely correlate with race because they are drawn on the basis of political affiliation, which correlates with race, there is no racial classification to justify.” Id. at 968 (opinion of O’Connor, J.). Second is LULAC v. Clements, 999 F.2d 831, 854 (5th Cir. 1993) (en banc), in which the Fifth Circuit (by which, as the majority reminds us, we are firmly bound) held that “[Section] 2 is implicated only where Democrats lose because they are black, not where blacks lose because they are Democrats.” As Professor Hasen explains, “The court pointed to this fact and the support of minority candidates by party members of each party[15] as a reason for treating race and party as separate categories, even while acknowledging that ‘[m]inority voters . . . have tended to support the Democratic Party.’”16 Thus, the Fifth Circuit has recognized facts that constitute conjoined polarization in Texas and has squarely adopted the “race or party” approach as the correct 13 Id., text accompanying note 89 (quoting Hunt v. Cromartie, 526 U.S. 541, 551 (1999)). 14 Id. note 37. 15 Importantly, several of the districts most hotly disputed in the Texas 2011 redistricting litigation are districts in which the Republicans, in recent elections, have nominated and supported minority candidates. Examples are AfricanAmerican Republican Congressman Will Hurd in CD23; Hispanic Republican Congressman Francisco Conseco in CD23; Hispanic Republican Representative Aaron Peña in HD41; Hispanic Republican Jose Aliseda in HD35, and Hispanic Republican Representative John Garza in HD117. 16 Hasen, text accompanying note 74 (quoting LULAC v. Clements, 999 F.2d at 854). 161 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 162 of 171 methodology: As Professor Hasen explains, “Since . . . Clements, conjoined polarization has only increased in the state of Texas, rendering the factual premise that race and party can be separated in Texas more dubious.”17 In summary, in light of conjoined polarization in Texas, the disjunctive “race or party” formulation is the proper understanding of the law. If that were not otherwise obvious, it is compelled by Bethune-Hill, Easley v. Cromartie, and LULAC v. Clements. But because the judges in the majority seem not to embrace it, I will discuss the other two approaches briefly. Moving from “race or party,” Professor Hasen describes the second approach, “party as race,” as “treat[ing] party as a proxy for race, equating proof of discriminatory partisan intent with proof of discriminatory racial intent.”18 We need not dwell for long on this formulation, because “[t]he Fifth Circuit in [LULAC v. Clements] . . . rejected the view that partisanship can serve as a proxy for race, viewing partisanship and racial discrimination as two discrete categories.”19 The en banc Fifth Circuit therefore “rejected the argument that ‘the Republican and Democratic Parties are proxies for racial and ethnic groups in Texas’” as well as the notion that “‘“partisan politics” is “racial politics.’””20 And, as already shown, “ ‘[p]arty as race’ ” [was] denied by the Supreme Court in Hunt v. Cromartie . . . .”21 That leaves us with the third explication, “party all the time.”22 It 17 Id., text accompanying note 75. 18 Id., text accompanying note 9. 19 Id., text accompanying note 90. 20 Id., text accompanying note 71 (quoting LULAC v. Clements, 999 F.2d at 860). 21 Id., text accompanying note 125. 22 It strikes me that “party all the time” might be a good recruiting slogan for a college fraternity. But I digress. 162 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 163 of 171 deal[s] with the problem of conjoined polarization [by litigating] these cases not as race cases but as party cases, having courts rule that certain partisan actions are themselves illegal. Thus, rather than making partisanship or incumbency protection a defense in cases of racial gerrymandering, vote dilution, or vote denial, such conduct would make out a prima facie case for liability.[23] For Professor Hasen, this is admittedly not the law and is only one proposal on an articulately advanced wish list: He acknowledges that under his proposal, “[t]he achievement of partisan ends would not be considered a good reason [for legislative action,] as it appears to be in the redistricting context.”24 As a matter of strategy, “it often will be easier and more fruitful to try to attack the partisanship directly, relying on the racial litigation route only as a backstop.”25 There would be no need to prolong this discussion of “party all the time,” which amounts to “direct policing of partisanship,”26 except that the majority advances it in one discrete respect. In their opinions in both the 2011 Congressional and 2011 House cases, the judges in the majority exude discomfort with the indisputable fact that the 2011 Republican legislative majority took every possible partisan advantage of the super-majority it gained from the 2010 Republican sweep. The question for this panel is whether that exercise of raw political leverage is lawful, under the VRA and Constitution, even if many might see it as unfair or abusive of power. I addressed the issue of partisan advantage in my dissent in the 2011 Congressional case. Regarding the Legislature’s overt but unsuccessful attempt to unseat longtime Anglo Democrat 23 Hasen, text accompanying note 135. 24 Id., text accompanying note 143 (citation and internal quotation marks omitted). In fairness, I need to add that because of issues regarding judicial manageability, Professor Hasen is “ambivalent” on the efficacy of “party all the time” in redistricting, as distinguished from, e.g., registration and voter ID cases. Id., text accompanying note 147. 25 Hasen, text accompanying notes 152–153. 26 Hasen, text accompanying notes 148–149. 163 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 164 of 171 Congressman Lloyd Doggett in the Austin area, I noted that “[a]n attempt to dislodge an incumbent political adversary should logically be viewed as a permissible redistricting principle,” because a party’s disadvantaging its political foes, for reasons not related to invidious purposes such as race, is not meaningfully different from the protection of that party’s incumbents. Perez v. Abbott, 2017 U.S. Dist. LEXIS 35012, at *274 (Smith, J., dissenting). The majority disagreed, opining that unseating Doggett “ignor[ed] many if not most traditional redistricting principles.” Id. at *177 (majority opinion).27 In summary, the judges in the majority engage in a well-meaning and unintended misreading of the law. To be fair, there is some residual uncertainty from the Supreme Court’s redistricting pronouncements over the past decades. One thing we can be sure of, however, is that as of March 2017, these plaintiffs must show that the legislature employed “race for its own sake.” Bethune-Hill, 137 S. Ct at 798. The law forbids “the racial purpose of state action, not its stark manifestation.” Id. (emphasis added, citation omitted). “Race for its own sake” must be “the overriding reason for choosing one map over others.” Id. at 799. 27 At least one blogger was alarmed at my suggestion that legislative majorities should be allowed to treat adversaries as adversaries: Litigants should challenge the notion that “incumbency protection” includes an interest in “incumbency advantage.” Holding otherwise quickly leads to a perverse and startling result. In his Perez dissent, for example, Judge Smith argued that a state’s interest in “incumbency protection” implies an interest [in] “incumbency advantage” and, therefore, a constitutional interest in “punishment of enemies [quoting the dissenting passage described above].” Voting rights advocates should forcefully challenge this line of reasoning. This “blood sport republic” logic is fundamentally incompatible with a constitutional framework that demands¯at a minimum¯a rational basis for state action. The “partisan advantage defense” should not shield racial sorting, and Cromartie (and its progeny) should not be extended to condone illegitimate uses of political data.” G. Michael Parsons, Is Bethune-Hill a Major Voting Rights Victory or the Next Northwest Austin?, available at https://moderndemocracyblog.com/2017/03/30/is-bethune-hill-a-major-voting-rights-victory-or-the-next-northwestaustin/. Mr. Parsons, a private attorney/blogger, at least agrees with me that “[i]f anything, Bethune[-Hill] took a step backwards” for redistricting plaintiffs. Id. n.15. 164 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 165 of 171 III. The correct law applied to these facts. It is on this explanation of “race for its own sake,” issued by the Supreme Court just a few weeks ago, that the majority fatally stumbles. The majority is right only if the State sought, as its ultimate end, not to exploit partisan advantage but, instead, to disadvantage voters (either in part or in whole) because they were members of a racial or ethnic class and were, for that reason, inherently unworthy of equal treatment under the law. By that approach, the State would be treating members of a minority as having a certain identifiable “character,” possibly creating “tension,” as shown above by the testimony for CD9. By the same notion, per the record quoted above for CD30, if the State had used “race for its own sake,” it would have assigned voters together with “people that look like them that think like them, that live with them,” instead of doing its best to ascertain voters’ political affiliation in order to maintain numerical political hegemony. The main flaw in the majority’s reasoning is that it defies logical explanation. First, the record is absolutely devoid of indication that the Legislature treated race qua race; instead, it is replete with examples (possibly hundreds of them) of decisions made, and actions taken, to advance the Republican majority, whether in a quest to retain power for its own sake or to advance policy goals favored by one party and opposed by the other. Every crime requires a motive. No reason is advanced as to why or how it would have made sense for the legislative majority to disadvantage minority voters because they were minority voters, as the law requires the plaintiffs to show. Nothing in this record reflects partisanship as a proxy for race. If it did, that would mean that the legislators cared less about Republican strength than about racial disparagement¯that, in fact, in the 2011 Texas Legislature, political considerations played second fiddle to racial division: The stated objective was partisan power, but the real and secret goal 165 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 166 of 171 was subjugation of minorities. It is indeed possible to imagine a world¯maybe the Deep South in the 1920s¯in which the main guidance for legislative action was racial hatred or prejudice. But to their credit, these private plaintiffs make no such claim, nor do my distinguished colleagues in the majority believe it. This is not “race for the sake of race.” The majority may predictably respond that I have it backwards: that the issue is not whether political party was a proxy for race, but the flipside: whether race was a proxy for partisanship.28 That is to say, the plaintiffs and the majority seem to take the position that even assuming that the ultimate purpose was the non-invidious goal of maintaining partisan strength, race was often used in drawing lines, in the knowledge that the result of a given boundary change would likely be an increase in Republican electability. Thus, the majority posits that, for example, “turning on racial shading” in the computerized redistricting program is seldom or never acceptable, because race is the hidden reason a particular change (such as breaking a precinct) is made, and though the real and stated goal is purely partisan, the voters left in or out of the precinct are being so treated because of some stereotypical characteristic that is calculated to equate with political party. The difficulty with this reverse-proxy construct is that it is still not “race for the sake of race.” The correct construct is dictated by Cromartie, in which such “precinct swapping,” 532 U.S. at 257, was center stage. The test is whether the Legislature’s decision “is unexplainable on grounds other than race.” Id. at 242 (citations omitted). “Caution is especially appropriate in this case, where the State has articulated a legitimate political explanation for its districting decision, and the voting population is one in which race and political affiliation are highly correlated.” Id. Importantly, in Cromartie the Court approved of a precinct split that “plac[ed] the more 28 E.g., Majority Op. at 137 (“race as a proxy for partisanship”). 166 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 167 of 171 heavily African-American segment [of the precinct] within [the challenged district].” Id. at 248. The Court accepted, with apparent approval, the testimony of a state expert “that African-Americans are more reliable Democrats than whites.” Id. at 262. And in Bethune-Hill, the Court most recently reminds us that instead of focusing on a minuscule change at a precinct boundary, “[t]he ultimate object of the inquiry . . . is the legislature’s predominant motive for the design of the district as a whole . . . . [A]ny explanation for a particular portion of the lines, moreover, must take account of the districtwide context.” Bethune-Hill, 137 S. Ct. at 800. In Cromartie, “the [Democrat-dominated] legislature drew boundaries that, in general, placed more-reliably Democratic voters inside the district, while placing less-reliable Democratic voters outside the district.” Cromartie, 532 U.S. at 252. The three-judge district court, agreeing with plaintiffs’ expert, declared that “race is the predominant factor.” Id. at 249. “But this statement of the conclusion is no stronger than the evidence that underlies it.” Id. “Conclud[ing] that the District Court’s . . . findings are clearly erroneous,” the Supreme Court reversed. Id. at 258. The same should obtain here. IV. The majority’s flawed reasoning. Faced with its perceived task of deciding whether race predominated in the 2011 Legislature’s redistricting decisions, the majority embarks, as I have said, on an impressive recitation of the record. But it gives inadequate nod to legitimate political aims and traditional redistricting principles and, instead, rests much of its decree on finding that various state witnesses presented testimony that it calls “not credible.”29 29 The majority says “not credible” twenty-seven times (in its Congressional and House opinions and findings/conclusions), an astonishing number of times for a court to find that key witnesses misrepresented material facts under oath. And it finds what it calls “bad faith” nine times. 167 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 168 of 171 The seriousness of the majority’s implied allegation cannot be overstated. In its impressively detailed explanation, it weaves a complex, widespread conspiracy of scheming and plotting, by various legislators and staff, carefully designed to obscure the alleged race-based motive in an effort to achieve their objectives and¯by necessary implication¯intentionally30 to deprive minority citizens of their political and civil rights.31 A representative example, though by far not the only one, is the majority’s critique of the redistricting of House districts touching Nueces County. Here is the plot that the majority uncovered: Because Nueces County was less than 50% SSVR, it was mathematically impossible to draw two SSVR-majority districts within the County. However, mapdrawers knew that SSVR was typically lower, than HCVAP, that Nueces County had been over 50% HCVAP in 2000, and that Hispanic population had increased while Anglo population had declined, meaning they had every reason to believe that the HCVAP of Nueces County was over 50% even without more recent HCVAP data. Moreover, Defendants’ assertion that HCVAP data was not available is simply untrue. The HCVAP data was available to mapdrawers by April 21, 2011, before the map went to the floor, and would have shown that Nueces County HCVAP was over 50% . . . . Nevertheless, Interiano . . . and other[s] . . . continued to focus solely on SSVR . . . . [M]apdrawers did not explore whether to include more than two districts in Nueces County . . . . The . . . facts . . . concerning the decision to eliminate HD33 in Nueces County demonstrate intentional vote dilution. . . . Mapdrawers intentionally did not consider whether § 2 required two districts in Nueces County, and used a false reliance on SSVR and the County Line Rule to justify their refusal to maintain two Latino opportunity districts in Nueces County. Further, they knew that HCVAP was the proper measure for a Latino opportunity district and had HCVAP data available before the map went to the floor, but insisted on using SSVR as the measure because SSVR was lower (and lower than 50%) and it allowed them to argue that mathematically, two SSVR-majority districts could not be drawn. Although Nueces County was majority HCVAP and had been since at 30 The majority uses the word “intentionally” approximately one hundred times in the Congressional and House opinions and findings/conclusions. 31 In its description of the allegedly nefarious enterprise, the majority uses forms of the word “manipulate” about forty-five times. 168 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 169 of 171 least 2000, they refused to consider whether two HCVAP-majority districts could be drawn wholly within the County. In addition, [they] steadfastly refused to consider whether the County Line Rule might have to yield to maintain two Latino opportunity districts in the area . . . [They] attempted to use § 5 as a shield for their failure to attempt compliance . . . . Mapdrawers chose two districts that were already performing reliably for Latino voters despite being under 50% SSVR . . . and raised their SSVR over 50% . . . . Mapdrawers and redistricting leadership relied in bad faith on a 50% SSVR standard . . . . [M]apdrawers acted in bad faith to thwart such claims by artificially inflating the number of HCVAP-majority districts . . . . [T]hey intentionally used [race] to undermine Latino voting opportunity. [M]apdrawers were likely using race to assign population [using ten precinct splits]. [They] intentionally packed Hispanic voters into HD32 to minimize their number and influence in HD34 and protect Hunter. [T]hey targeted low turnout minority areas and intentionally drew out potential Hispanic rivals . . ., again to protect the Anglo incumbent . . . [and] intentionally overpopulated HD32 (and underpopulated . . . HD34), without a legitimate justification. [T]he Court finds that redistricters intentionally diluted Latino voting strength by eliminating HD33 in Nueces County. Mapdrawers’ use of race in HD90 and HD148 to “offset” this loss was not to comply with the VRA but to intentionally dilute Latino voting strength by (1) allowing redistricting leadership to claim that they were complying with § 5 despite eliminating HD33 and creating no new ability districts; (2) thwarting arguments that § 2 might require additional opportunity districts by artificially inflating the number of Latino-majority districts in the plan; and (3) racially gerrymandering and utilizing population deviations to further dilute the Latino vote in the remaining two Nueces County districts. Majority opinion at 33–40 (some paragraph breaks omitted). Indeed, this complex conspiracy for Nueces County alone, affecting districts spanning hundreds of miles from Corpus Christi to Dallas/Fort Worth, would make Bernie Madoff proud. To the contrary, however, if ever traditional districting principles should be respected¯as countless Supreme Court decisions say they must¯it is in Nueces County. As the majority admits, the County Line Rule and Nueces County are a perfect fit. Because the county proportionally lost population in relation to the rest of the state, it no longer qualifies for 2-plus districts but, by simple math, is entitled to 2.03 representatives. That is as close as it gets. The Legislature accordingly 169 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 170 of 171 assigned Nueces County exactly two whole seats, and HD33 was sent packing to North Texas. The only remaining question should be where, within the county, to draw the line between the two so as to comply with one-person-one vote. The majority goes well beyond that by complaining of the removal of HD33 and other ills. An ultimate decision to return CD33 to Nueces County would violate the most perfect fit of that rule in the entire state and would require, at the remedy stage, changes in myriad districts between Nueces County and the new HD33. And that is just Nueces County. The same apparent perspective¯that legislative leaders and staff plotted tirelessly, throughout the legislative session, to obscure the facts and violate the law¯infects most of the majority’s fulsome explanation. V. Summary. The law is, in some respects, unsure. For example, the Supreme Court will need to resolve the circuit split on whether to require coalition districts. And it will have opportunity to tackle the issue of partisan gerrymandering and to explain the limits, if any, to the exercise of partisan advantage. More refinement is needed to inform legislatures how to deal with the Scylla and Charybdis of meeting the requirements of Section 2 of the VRA without undermining equal protection or unduly violating traditional redistricting principles. And some higher court will need to tell this panel whether the 2011 proceedings are moot. Despite these uncertainties, one thing is certain: Despite its heartfelt efforts, this panel majority has badly overreached in finding that Texas used race, instead of partisan advantage, as the predominant factor in the 2011 redistricting. A finding that racial considerations were “dominant and controlling”32 defies everything about this record. 32 Miller v. Johnson, 515 U.S. at 913. 170 Case 5:11-cv-00360-OLG-JES-XR Document 1365 Filed 04/20/17 Page 171 of 171 Is there anything to show that, as to any district whatsoever, “race was the criterion that, in the State’s view, could not be compromised”?33 Nary a word. I respectfully dissent. Stay tuned. ______________/s__________________ JERRY E. SMITH UNITED STATES CIRCUIT JUDGE 33 Bethune-Hill, 137 S. Ct. at 798. 171