Rodriguez et al v. Harris County, Texas et al
Filing
107
ORDER. (Signed by Judge Vanessa D Gilmore) Parties notified.(bthomas)
IN THE UNITED STATES DISTRICT COURT
FOR THE SOUTHERN DISTRICT OF TEXAS
HOUSTON DIVISION
JAMES RODRIGUEZ, et al.,
Plaintiffs,
v.
HARRIS COUNTY, TEXAS, et al.,
Defendants.
§
§
§
§
§
§
§
§
§
Civil Action No. 4:11-2907
ORDER
In this case, Plaintiffs Edward Gonzalez, Orlando Ybarra, Patricia Gonzales, Reynaldo Guerra,
Sandra Puente, Thomas Berg, and James Rodriguez (collectively, “Plaintiffs”) challenge the legality of
the redistricting plan adopted by the Harris County Commissioner’s Court on August 9, 2011, known as
Revised Plan A-1, which reapportions voters among Harris County’s four commissioner precincts (the
“County’s Plan”). Plaintiffs claim that the County’s Plan dilutes the voting strength of politically
cohesive Latinos in Harris County Commissioner’s Precinct 2 (“Precinct 2”), such that this population
has less opportunity than other members of the electorate to participate in the political process and to
elect representatives of their choice, resulting in a violation of Section 2 of the Voting Rights Act, as
amended, 42 U.S.C. § 1973 et seq. Plaintiffs further contend that the County’s Plan violates the Equal
Protection Clause of the Fourteenth Amendment of the United States Constitution, because it allegedly
was drawn (i) with the motive of intentional discrimination against Hispanics and (ii) with excessive and
unjustified use of race and racial data. Defendants Harris County, Texas, and Ed Emmett, in his capacity
as Harris County Judge (collectively, “Defendants”), contend that the County’s Plan neither violates the
Voting Rights Act nor does it offend the Constitution.
This matter was tried without a jury beginning on November 13, 2012. The Court, having heard
the arguments, read the submissions of counsel, reviewed the governing legal principles, and considered
the testimonial evidence and exhibits, submits the following findings of fact and conclusions of law.
1
I.
LEGAL FRAMEWORK: SECTION 2 OF THE VOTING RIGHTS ACT
Section 2 of the Voting Rights Act of 1965 (the “Voting Rights Act”), as amended in 1982,
provides that:
(a) No voting qualification or prerequisite to voting or standard, practice, or procedure
shall be imposed or applied by any State or political subdivision in a manner which
results in a denial or abridgement of the right of any citizen of the United States to vote
on account of race or color . . . .
(b) A violation of subsection (a) of this section is established if, based on the totality of
circumstances, it is shown that the political processes leading to nomination or election in
the State or political subdivision are not equally open to participation by members of a
class of citizens protected by subsection (a) of this section in that its members have less
opportunity than other members of the electorate to participate in the political process
and to elect representatives of their choice. The extent to which members of a protected
class have been elected to office in the State or political subdivision is one circumstance
which may be considered: Provided, that nothing in this section establishes a right to
have members of a protected class elected in numbers equal to their proportion in the
population.
42 U.S.C. § 1973. The Voting Rights Act was enacted to “give those who had been disenfranchised on
account of their race the opportunity to participate in the political process.” White v. Alabama, 74 F.3d
1058, 1069 (11th Cir. 1996). Section 2 proscribes practices that, while permitting a mechanical exercise
of the right to vote, “operate[] to cancel out or minimize [i.e. dilute] the voting strength of racial
groups,” such that members of the racial minority have less opportunity than other members of the
electorate to participate in the political process and to elect representatives of their choice. See
Thornburg v. Gingles, 478 U.S. 30, 87 (1986) (O’Connor, J., concurring); Rodriguez v. Bexar Cnty., 385
F.3d 853, 859 (5th Cir. 2004); see also White v. Regester, 412 U.S. 755, 765-66 (1973).
Ultimately, the right to undiluted voting strength, provided by Section 2, is a guarantee of equal
opportunity in voting, ensuring that a minority group is not denied, on account of race, color, or
language minority status, the opportunity to exercise an electoral power that is commensurate with its
population in the relevant jurisdiction. Johnson v. De Grandy, 512 U.S. 997, 1014 n.11 (1994) (“the
ultimate right of Section 2 is equality of opportunity”); Hall v. Virginia, 385 F.3d 421, 429 (4th Cir.
2004); see also Campos v. City of Houston, 113 F.3d 544, 546 (5th Cir. 1997). That said, the goal of
2
Section 2 is not to guarantee success at the polls for minority-preferred candidates but to provide
assurances of fairness in the electoral process. De Grandy, 512 U.S. at 1014. The promise of Section 2 is
the removal of all barriers to truly equal participation, but a necessary part of equal participation is the
possibility of a loss. United States v. Euclid City Sch. Board, 632 F. Supp. 2d 740, 752 (N.D. Ohio
2009). After all, Section 2 protects the right to meaningful participation, “not the right to vote for the
winning candidate.” Nevett v. Sides, 571 F. 2d 209, 236 (5th Cir. 1978).
A state or political subdivision thereof violates Section 2 whenever it adopts a practice that
dilutes the voting strength of a minority group, such that members of the minority group have less
opportunity for meaningful participation and franchise.1 Voinovich v. Quilter, 507 U.S. 146, 157 (1993);
Gingles, 478 U.S. at 44. Proof of vote dilution is assembled using the two-part framework set forth in
the seminal case of Thornburg v. Gingles, 478 U.S. 30 (1986). Growe v. Emison, 507 U.S. 25, 40-41
(1993) (vote dilution challenge to a single-member districting scheme is analyzed using the framework
set forth in Gingles); Fairley v. Hattiesburg, 584 F.3d 660, 667 (5th Cir. 2009).
Under Gingles, a plaintiff must first demonstrate that: (1) the affected minority group is
sufficiently large and geographically compact to constitute a majority in an additional single-member
district; (2) the affected minority group is politically cohesive; and (3) the white majority votes
sufficiently as a bloc to enable it—in the absence of special circumstances—usually to defeat the
minority group’s preferred candidates. League of United Latin Am. Citizens v. Perry, 548 U.S. 399, 425
(2006); Gingles, 478 U.S. at 50-51; Fairley, 584 F.3d at 667. “[The Fifth Circuit] has interpreted the
Gingles factors as a bright line test.” Valdespino v. Alamo Heights Indep. Sch. Dist., 168 F.3d 848, 852
(5th Cir. 1999). Each factor must be proved. Growe, 507 U.S. at 40-41; Fairley, 584 F.3d at 667
(quoting Sensley v. Albritton, 385 F.3d 591, 595 (5th Cir. 2004)). Failure to establish any one of the
three Gingles factors precludes a finding of vote dilution, because if these three preconditions are not
1
Vote dilution occurs either by fragmenting large concentrations of minority populations and dispersing them into
separate political districts, or by concentrating minorities into districts where they constitute super-majorities. See
Gingles, 478 U.S. at 46 n.11 (1986).
3
present then the plaintiffs cannot show that the challenged electoral practice or device impairs minority
voters’ ability to elect representatives of their choice. League of United Latin Am. Citizens No. 4552 v.
Roscoe Indep. Sch. Dist., 123 F.3d 843, 847 (5th Cir. 1997); League of United Latin Am. Citizens No.
4434 v. Clements, 986 F.2d 721, 743 (5th Cir. 1993) (“LULAC III”); Benavidez v. Irving Indep. Sch.
Dist., 690 F. Supp. 2d 451, 455 (N.D. Tex. 2010).
The satisfaction of these three preconditions is necessary but not sufficient to establish a Section
2 claim. De Grandy, 512 U.S. at 1012-13; see also Perry, 548 U.S. at 425-26. If these three
preconditions are met, the district court must then examine a variety of other factors to determine
whether, under the totality of the circumstances, the challenged practice impairs the ability of the
minority voters to participate equally in the political process and to elect a representative of their choice.
De Grandy, 512 U.S. at 1011; Fairley, 584 F.3d at 667 (citing Sensley, 385 F.3d at 595); see also
Bartlett v. Strickland, 556 U.S. 1, 11-12 (2009). This requires a “searching practical evaluation of the
past and present reality.” Gingles, 478 U.S. at 45 (quoting S. Rep. No. 417, 97th Cong., 2d. Sess., (1982)
(“Senate Report”) at 30, reprinted in 1982 U.S. Code Cong. & Admin. News (“U.S.C.C.A.N.”) 177 at
208).
In conducting this broader inquiry, the court should consider the objective factors set forth in the
Senate Report accompanying the 1982 amendments to the Voting Rights Act, including: (1) the history
of voting-related discrimination in the State or political subdivision; (2) the extent to which voting in the
elections of the State or political subdivision is racially polarized; (3) the extent to which the State or
political subdivision has used voting practices or procedures that tend to enhance the opportunity for
discrimination against the minority group; (4) the degree to which members of the minority group have
been denied access to the candidate slating process; (5) the extent to which minority group members
bear the effects of past discrimination in areas such as education, employment, and health, which hinder
their ability to participate effectively in the political process; (6) the use of overt or subtle racial appeals
in political campaigns; (7) the extent to which members of the minority group have been elected to
4
public office in the jurisdiction; (8) whether there is a significant lack of responsiveness on the part of
elected officials to the particularized needs of the members of the minority group; and (9) whether the
policy underlying the state or political subdivision’s use of such voting qualification, prerequisite to
voting, or standard, practice, or procedure is tenuous. Senate Report at 28-29, 1982 U.S.C.C.A.N. at
206-07; Perry, 548 U.S. at 426 (citing Gingles, 478 U.S. at 44-45).
Although the totality of the circumstance analysis is guided by the Senate factors, these factors
are neither “comprehensive nor exclusive.” Gingles, 478 U.S. at 45; Teague v. Attala Cnty., Miss., 92
F.3d 283, 292 (5th Cir. 1996) (noting that the district court must be “flexible in its totality inquiry and
guided by factors drawn from the Senate Judiciary Committee report on the 1982 amendments to the
Voting Rights Act.”). The Court must also consider “whether the number of districts in which the
minority group forms an effective majority is roughly proportional to its share of the population in the
relevant area,” Perry, 548 U.S. at 426, 436; De Grandy, 512 U.S. at 1000, as well as any other fact
which may bear on the question of dilution. See Gingles, 478 U.S. at 45. Ultimately, the totality of the
circumstances inquiry presents very complex political and legal issues whose resolution requires a
comprehensive canvassing of the relevant facts, De Grandy, 512 U.S. at 1011, and “an intensely local
appraisal of the design and impact of the contested electoral mechanisms.” Gingles, 478 U.S. at 79.
Given that the plaintiff’s ability to establish the three Gingles factors does not end the court’s
inquiry, proof of the Gingles factors does not always portend liability. De Grandy, 512 U.S. at 1012-13.
That said, “it will only be the very unusual case in which the plaintiffs can establish the existence of the
three Gingles factors but still have failed to establish a violation of § 2 under the totality of
circumstances.” Teague, 92 F.3d at 292; Clark v. Calhoun Cnty., Miss., 88 F.3d 1393, 1396 (5th Cir.
1996) (“Clark II”); Benavidez v. City of Irving, Tex., 638 F. Supp. 2d 709, 712-13 (N.D. Tex. 2009).
“Because the resolution of a voting dilution claim requires close analysis of unusually complex
factual patterns, and because the decision of such a case has the potential for serious interference with
state functions, . . . district courts [must] explain with particularity their reasoning and the subsidiary
5
factual conclusions underlying their reasoning.” Westwego Citizens for Better Gov’t v. City of
Westwego, 872 F.2d 1201, 1203 (5th Cir. 1989) (“Westwego I”); Roscoe, 123 F.3d at 846; see also
Fairley v. Hattiesburg, 584 F.3d 660, 668, 668 n.5 (5th Cir. 2009). Although the Court is expected to
consider all substantial evidence, Plaintiffs, at all times, bear the burden of proving, by a preponderance
of the evidence, that all of the Gingles preconditions have been met and that based on the totality of the
circumstances, the challenged election device dilutes the voting strength of the minority group. Roscoe,
123 F.3d at 846.
II. BACKGROUND2
A.
THE PARTIES & THEIR CONTENTIONS
Plaintiffs Edward Gonzalez, Orlando Ybarra, Patricia Gonzales, Reynaldo Guerra, and Sandra
Puente all are Hispanic registered voters residing in Commissioners Court Precinct 2 in Harris County,
Texas. (Instrument No. 91, at 14 ¶ 1). Plaintiffs Thomas Berg and James Rodriguez are not residents of
Precinct 2. (Instrument No. 91, at 14 ¶ 2). Plaintiffs allege that the County’s Plan violates Section 2 of
the Voting Rights Act and the Fourteenth Amendment. As for the Section 2 claims, Plaintiffs contend
that: (i) they can draw a precinct in which Latinos constitute a majority of the citizen voting age
population; (ii) Latinos are politically cohesive; (iii) Anglos vote, as a bloc, to defeat the Latinopreferred candidate; and (iv) under the totality of the circumstances, the County’s plan dilutes the voting
strength of the Latino residents of Precinct 2. As for the constitutional claims, Plaintiffs contend that the
County’s Plan was drawn with an excessive and unjustified use of racial data and was drawn with the
intent to discriminate against Latinos. Intervenor the League of United Latin American Citizens
(“LULAC”) agrees with Plaintiffs’ contentions.
Defendants Harris County, Texas, and Ed Emmett, in his capacity as Harris County Judge, have
appeared and answered herein. (Instrument No. 91, at 14 ¶ 4). Defendants claim that the County Plan
2
To the extent that a finding of fact constitutes a conclusion of law, it is adopted as such, and to the extent that a
conclusion of law constitutes a finding of fact, it is likewise adopted as such.
6
does not violate either Section 2 or the Fourteenth Amendment. Specifically, Defendants contend that
their Plan does not violate Section 2 because: (i) Plaintiffs cannot show that Latinos constitute a majority
of the citizen voting age population in a geographically compact district drawn in accordance with
traditional redistricting principles; (ii) Plaintiffs cannot establish the existence of racial bloc voting
because Plaintiffs cannot show that race, rather than some non-racial factor, such as partisanship,
accounts for the differences in the voting patterns of Anglos and Latinos; and (iii) under the totality of
the circumstances, Plaintiffs cannot show that the County’s Plan diminishes their opportunity to
participate in the political process and elect representatives of their choice. Defendants further contend
that the County’s Plan does not violate the Fourteenth Amendment because the Plan was not drawn with
the intention to discriminate nor was it drawn with undue attention to race or racial considerations.
The Precinct 1 Intervenors, Senfronia Thompson, Harold Dutton, Bruce Austin, Willie Bell
Boone, and Howard Middleton, are all African-American registered voters residing in Harris County
Commissioners Precinct 1. (Instrument No. 91, at 14, 15 ¶ 3, 7). Harris County’s Commissioner Precinct
1 is an effective African-American opportunity precinct created in 1980 and is protected by the Voting
Rights Act. (Instrument No. 91, at 14, 15 ¶ 3, 7). The Precinct 1 Intervenors also argue that the County’s
Plan violates neither Section 2 of the Voting Rights Act nor the Fourteenth Amendment.
B.
CURRENT METHOD OF ELECTING COUNTY COMMISSIONERS & THE
NEED FOR REDISTRICTING
Section 18 of Article V of the Texas Constitution provides, in relevant part, that:
Each county shall, in the manner provided for justice of the peace and constable
precincts, be divided into four commissioners precincts in each of which there shall be
elected by the qualified voters thereof one County Commissioner, who shall hold his
office for four years and until his successor shall be elected and qualified. The County
Commissioners so chosen, with the County Judge as presiding officer, shall compose the
County Commissioners Court, which shall exercise such powers and jurisdiction over all
county business, as is conferred by this Constitution and the laws of the State, or as may
be hereafter prescribed.
Tex. Const. art. V, § 18. In keeping with this constitutional dictate, Harris County government is
organized into a commissioners court; the Harris County Commissioners Court, comprised of the county
7
judge and four county commissioners, is tasked with the administration of the county’s business and
legislative affairs. (Instrument No. 91, at 14 ¶ 5); see also City of San Antonio v. City of Boerne, 111
S.W.3d 22, 27-28 (2003) (quoting Avery v. Midland Cnty., 406 S.W.2d 422, 426 (1966)); LouisianaPacific Corp. v. Newton Cnty., 149 S.W. 3d 262, 263 (Tex. App.—Eastland 2004, no pet.); Concerned
Citizens for Equality v. McDondald, 863 F. Supp. 393, 395 (E.D. Tex. 1994). The services provided by
Harris County Commissioners to their constituents include drainage services, unincorporated road
maintenance, emergency services, medical clinics, parks and recreational facilities, community centers,
education centers, libraries, and precinct-wide community programs. (Instrument No. 91, at 14 ¶ 6).
Harris County has a competitive multiparty political system dominated by the two major political
parties, the Democratic Party and the Republican Party. (Instrument No. 91, at 19 ¶ 35). Candidates for
County Commissioner run partisan campaigns. Democrat Sylvia Garcia, a Latina candidate, was first
elected as Precinct 2 County Commissioner in 2002. She was re-elected in 2006, but was narrowly
defeated in 2010 by Republican Jack Morman, an Anglo candidate. (Instrument No. 91, at 19 ¶ 36). Mr.
Morman is the current commissioner for Precinct 2. (Instrument No. 91, at 19 ¶ 38).
Prior to 2011, Harris County commissioner precincts were last reapportioned in 2001, resulting
in the 2001 map, depicted in the next diagram. (Instrument No. 91, at 15 ¶ 8).
8
Between 2001 and 2010, population changes and shifts occurred in Harris County that resulted in
substantial numerical variation in the population of the four commissioner precincts, as reflected by
2010 Census data. (Instrument No. 91, at 15 ¶ 9). Based on data from the 2010 Census, the populations
among the Commissioner Precincts, as drawn under the 2001 Map, ranged from a high in Precinct 3 of
1,148,890 to a low in Precinct 2 of 888,572, for a total top-to-bottom deviation of 25.44%. (Instrument
No. 91, at 15 ¶ 10).
Following the publication of the 2010 Census data, the Harris County Commissioners Court
announced its intent to redistrict the boundaries of its commissioner precincts. (Instrument No. 91, at 15
¶ 11). To assist the Commissioners Court in the evaluation of the new population statistics resulting
from the 2010 Census and in drafting the proposed plan, the County retained the services of Drs.
9
Richard Murray and David Branham and hired Andrews Kurth LLP as special counsel for the County on
redistricting (collectively, the “Consultants”). (Instrument No. 91, at 15 ¶ 12).
On June 21, 2011, the Consultants made a presentation at a public meeting of the Commissioners
Court concerning legal requirements for redistricting, the 2010 census data, and the need for adoption of
principles to guide the redistricting. An opportunity was provided for public comment at this meeting.
(Instrument No. 91, at 15 ¶ 13). On June 21, 2011, the Commissioners Court adopted an Order Adopting
Priorities and Principles for Redistricting of Harris County Commissioner Precincts (the “Priorities and
Principles”) to guide the formulation of the redistricting plan. (Instrument No. 91, at 15 ¶ 14). The
Commissioners Court adopted the following Priorities and Principles:
(a) the four commissioner precincts should be of substantially equal population and, in no
event, exceed a 10% top-to-bottom deviation;
(b) the four commissioner precincts should be contiguous and reasonably geographically
compact;
(c) to the extent possible, the redistricting plan should use identifiable geographic
boundaries as precinct boundaries, preserve natural historical boundaries, recognize
identifiable communities of interest in a single precinct and avoid splitting neighborhoods
when drawing precinct lines;
(d) to the extent possible, the four commissioner precincts should be based on existing
composition of the precincts;
(e) the redistricting plan should use whole county voting precincts to draw commissioner
precincts;
(f) the redistricting plan will adhere to (i) the Equal Protection Clause of the Fourteenth
Amendment to the United States Constitution and not use race as the predominate factor
in establishing commissioner precincts and (ii) the Voting Rights Act and avoid
retrogression in the position of racial, ethnic and language minorities with respect to the
effective exercise of their right to vote or electoral franchise. The redistricting plan
should not dilute voting strength of racial or language minority citizens and should not
either fragment minority communities or pack them into precincts in concentrations
greater than necessary to help them elect minority representation;
(g) recognizing the value of incumbent-constituency relations, the redistricting plan
should seek to keep (i) existing commissioners in their existing precincts and (ii) facilities
and service locations established by incumbent commissioners in the precincts of those
commissioners;
(h) the redistricting plan should recognize the duties and obligations imposed by law on
commissioners to provide services to the residents of precincts and the public investment
in facilities, service locations and personnel that has been made to provide such services.
(Defendants’ Ex. 3, at 4); (Instrument No. 91, at 15-17 ¶ 15).
10
On July 12, 2011, in a public meeting, the Commissioners Court adopted Plan A-1 for formal
consideration and for presentation to the public. An opportunity was provided for public comment at the
meeting. (Instrument No. 91, at 17 ¶ 16). In mid-July 2011, notice of the dates, times and locations for
public hearings was published in English, Spanish and Vietnamese in newspapers of general circulation
and in community newspapers. Notice was also provided in two separate locations on Harris County’s
website with Plan A-1 attached, along with the “Priorities and Principles of Redistricting” and census
information. (Instrument No. 91, at 17 ¶ 17).
From July 25 to August 1, 2011, Plan A-1 was presented and public comment was received at
four separate public hearings, one in each commissioner precinct. (Instrument No. 91, at 17 ¶ 18). At
each public hearing, Douglas Ray, Assistant County Attorney, presided and a PowerPoint presentation
was made that summarized the redistricting process, the legal considerations for redistricting, and the
Principles and Priorities. Defendants’ expert Dr. Richard Murray presented a summary of the
demographic considerations that went into the development of Plan A-1. (Instrument No. 91, at 17 ¶
19).3 Additionally, at each public hearing, 36-inch-by-48-inch mounted copies of maps of the existing
districts and Plan A-1 were displayed and printouts of the PowerPoint presentation as well as a chart
depicting the voting precinct changes were available in English, Spanish, and Vietnamese. The County
3
In 2011, the County hired Dr. Richard Murray, a seasoned voting rights academic, as a consultant on the
redistricting of the Harris County Commissioner’s Court Precincts. Dr. Murray is a political scientist at the
University of Houston. Dr. Murray has a bachelor’s and master’s degree in government, both from Louisiana
State University, and a PhD in political science from the University of Minnesota. (Tr. 3:170). Dr. Murray has
maintained professional affiliation with the University of Houston since 1966. (Tr. 3:170). He has been involved
in redistricting for a number of years with extensive experience studying voting and demographic trends in Harris
County. (Tr. 3:171). His involvement with redistricting dates back to 1971, with his introduction to this area
coming in the way of has been involved with the 1971 state-wide redistricting in Texas, the challenge to the
redistricting of the justice of the peace and constable districts, he was also involved with a challenge to the City of
Houston’s at large council system, he worked on the 1981 Harris County redistricting that led to the creation of
Precinct 1, and he has been involved in every single Harris County Commissioner’s Court redistricting since, i.e.
1981, 1991, 2001, and the present suit, and most recently was involved in the trials in San Antonio and
Washington, DC related to the Congressional and state house and senate plans. (Tr. 3:171-175). In his
redistricting efforts, he typically performs demographic analysis and draws maps. In his past redistricting
experience, Dr. Murray has served as a consultant for both the government and for plaintiffs. In this suit, Dr.
Murray served as a witness for the defendants.
11
also engaged Spanish and Vietnamese interpreters for each public hearing. (Instrument No. 91, at 17 ¶
20).
Members of the public were given an opportunity to make comments at the public hearings, and
many did, including elected officials and members of the African-American community in Precinct 1
and the Hispanic community in Precinct 2. (Instrument No. 91, at 17 ¶ 21). At each of the public
hearings, certain members of the public expressed opposition to the proposed plan in light of the impact
the plan would have on the Hispanic community in Precinct 2. Other members of the public, including
certain residents of Precinct 1, expressed support for the County’s Plan. (Instrument No. 91, at 17-18 ¶
22). Latino citizens, Dr. Reynaldo Guerra and Robert Jara, however, presented a proposed alternative
map (the “Guerra-Jara plan”) at the Precinct 3 public hearing. (Instrument No. 91, at 18 ¶ 23).
Following the first four public hearings, the Consultants revised draft Plan A-1 in response to
some of the concerns raised by members of the Hispanic Community, resulting in Revised Plan A-1 (the
“Revised Plan A-1”), depicted in the next diagram.4
4
Source: Plaintiffs’ Ex. 7.
12
13
Specifically, Revised Plan A-1 removed from Precinct 2 and placed back into Precinct 4 two majorityAnglo voting precincts (758 and 760) located in the Kingwood, Texas area. Revised Plan A-1 also
moved to Precinct 2 from Precinct 1 five heavily Hispanic whole voting precincts (737, 397, 792, 106,
and 817) and five heavily Hispanic partial voting precincts (423B, 061B, 409B, 408B, 608B) in the
Aldine area. (Instrument No. 91, at 18 ¶ 24). The changes made by the Consultants resulted in an
increase in the Hispanic voting age population in Precinct 2 under Revised Plan A-1, as compared to
Plan A-1. (Instrument No. 91, at 18 ¶ 25).5
On August 5, 2011, notice of an additional public hearing and a copy of Revised Plan A-1 were
posted on the County’s website and the notice of public hearing was published in a newspaper of general
circulation. (Instrument No. 91, at 18 ¶ 27). On August 9, 2011, a final public hearing was held during
which the public was invited to comment on redistricting generally and the proposed changes
incorporated in Revised Plan A-1. (Instrument No. 91, at 18 ¶ 28). Following the final public hearing, on
August 9, 2011, the Commissioners Court adopted Revised Plan A-1 and authorized its submission to
the United States Department of Justice (the “Justice Department”) for preclearance through an Order
Establishing New Boundaries for Commissioner Precincts for Harris County, Texas. (Instrument No. 91,
at 18 ¶ 29).6
On August 5, 2011, Plaintiffs filed the instant suit in the United States District Court in and for
the Southern District of Texas. See (Instrument No. 1). In the suit, Plaintiffs claim that the County’s Plan
violated Section 5 of the Voting Rights Act because the County was attempting to employ a new map
without first obtaining pre-clearance from the Justice Department. (Instrument No. 19). Plaintiffs further
claim that the County’s Plan dilutes the voting strength of politically cohesive Latinos in Precinct 2, in
5
The various plans considered by the County (and the plan adopted) did not show the Latino citizen voting age
population data but rather showed the Hispanic voting age population. (Instrument No. 91, at 18 ¶ 26).
6
Under Section 5 of the Voting Rights Act, jurisdictions with a history of racial discrimination in voting cannot
implement any change in voting practice, including the implementation of a redistricting plan, without first
seeking approval (i.e. “preclearance”) from the Justice Department. See 42 U.S.C. § 1973c.
14
violation of Section 2 of the Voting Rights Act. Finally, Plaintiffs contend that the County’s Plan
violates the Equal Protection Clause of the Fourteenth Amendment to the United States Constitution,
because it allegedly was drawn (i) with the motive of intentional discrimination against Hispanics and
(ii) with excessive and unjustified use of race and racial data. See (Instrument No. 19).
On October 19, 2011, a three-judge panel, composed of Judge Jennifer Elrod of the United States
Court of Appeals for the Fifth Circuit, United States District Judge Sim Lake of the Southern District of
Texas, and United States District Judge Vanessa Gilmore of the Southern District of Texas, was
convened to consider Plaintiffs’ Section 5 claims. On November 7, 2011, the three-judge panel
determined the County was seeking preclearance of the Revised Plan A-1 and did not intend to
implement the Revised Plan A-1 unless it obtained preclearance from the Justice Department.
Accordingly, the three-judge panel concluded that the Section 5 claims should be held in abeyance
unless the County attempted to implement a non-precleared plan.
On November 14, 2011, this Court, acting of its own accord without the three-judge panel, held
trial on the constitutionality of conducting the 2012 elections using the 2001 map, which was, at the
time, the last precleared plan. At the end of trial, the Court found that the use of the existing plan would
violate the constitution’s guarantee of one man, one vote and therefore concluded that the continued use
of the existing plan was unconstitutional. Because the Revised Plan A-1 had not at that time obtained
preclearance from the Department of Justice, the County could not use that plan to conduct the 2012
election. Accordingly, the Court fashioned an interim map that was used to conduct the 2012 election
(the “Interim Map”). See (Instrument No. 66). After slight modifications, neither party objected to the
Court’s interim map. The interim map is depicted in the next diagram.7
7
Source: Plaintiffs’ Ex. 60.
15
16
On December 30, 2011, the Justice Department issued a letter stating that it would not interpose
any objection to Revised Plan A-1 under its authority granted under Section 5 of the Voting Rights Act.
The Justice Department stated no opinions under Section 2 of the Voting Rights Act. (Instrument No.
91, at 19 ¶ 31). In the correspondence exchanged between Harris County and the Justice Department, the
County took the position that Precinct 2 was not an effective opportunity district but was instead a
Hispanic influence district. An opportunity district is a district in which members of a minority group
(i.e. Latinos), alone, are able to elect candidates of their choice, whereas an influence district is a district
in which members of a minority group (i.e. Latinos) are a minority of the voters, but “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.”
Bartlett v. Strickland, 556 U.S. 1, 13 (2009). The County further argued that the Revised Plan A-1
preserved Precinct 2’s status as an influence district. (Instrument No. 91, at 19 ¶ 32). The Justice
Department preclearance letter stated in part, “the failure of the Attorney General to object does not bar
subsequent litigation to enjoin enforcement of the change.” (Instrument No. 91, at 19 ¶ 33).
After the Justice Department pre-cleared the map, the three-judge panel dismissed Plaintiffs’
Section 5 claims, leaving only the aforementioned Section 2 claims and the constitutional claims at issue
in this lawsuit. Moreover, given that the County had obtained preclearance, it was free to implement the
Revised Plan A-1. Nevertheless, because the County obtained preclearance so late in the electoral
calendar, the County conducted the 2012 elections using the Court’s interim map, with the intention of
conducting all future elections under the Revised Plan A-1. In the instant case, Plaintiffs challenge the
Revised Plan A-1 on the basis that the Plan dilutes the Latino vote in Precinct 2, in violation of Section 2
of the Voting Rights Act, and the Plan was drawn with discriminatory intent, in violation of the
Fourteenth Amendment to the United States Constitution. Ultimately, Plaintiffs seek to establish
Precinct 2 as a Latino opportunity district under the Voting Rights Act. (Instrument No. 91, at 19 ¶ 34).
17
C.
POPULATION CHANGES AND DEMOGRAPHIC INFORMATION
In 2000, there were 3,400,578 people living in Harris County. (Plaintiffs’ Ex. 49, at 1, 2). Of that
number, 1,432,264 (42.1%) were Anglo, 1,119,791 (32.9%) were Latino, 619,694 (18.2%) were
African-American, and 228,869 (6.8%) were Asian and/or other racial background. Over the course of
the decade, the total population of Harris County increased by 691,881 individuals. See (Plaintiffs’ Ex.
49, at 2); (Tr. 2:7).8 During that time, the county-wide Anglo population shrunk by some 82,618
persons; the Latino population grew by 551,789 individuals; the African-American population grew by
134,564 persons; and the Asian and other population grew by 88,146 persons. (Plaintiffs’ Ex. 49, at 2);
(Tr. 2:7). At the end of the decade, Anglos accounted for 1,349,646 (33.0%) of the total population of
Harris County; there were 1,671,540 (40.8%) Latinos; 754,258 were (18.4%) African American; and
317,015 (7.7%) were people of Asian or other descent.
TOTAL POPULATION AND VOTING AGE POPULATION (VAP) CHANGE WITHIN
HARRIS
COUNTY
COMMISSIONER
PRECINCTS
BETWEEN
2000-2010
9
BYRACIAL/ETHNIC GROUP
Precinct 1
2000
2010
Change
827,341
100%
930,613
100%
103,272
+12.48%
Total Pop.
188,270
22.76%
175,127
18.8%
-13,143
-6.98%
Anglos
363,515
43.94%
346,810
37.3%
-16,705
-4.6%
Blacks
263,792
28.62%
357,504
38.4%
+93,712
+35.52%
Latinos
38,764
4.68%
51,172
5.5%
+12,408
+32.01%
Asians/Others
Voting
Age
Pop.
Anglos
Blacks
Latinos
Asians/Others
592,896
100%
680,238
100%
+87,342
+14.73%
158,278
253,932
150,706
29,980
26.70%
42.83%
25.42%
5.06%
151,788
256,488
230,504
41,458
22.3%
37.7%
33.9%
6.1%
-6,490
+2,556
+79,798
+11, 478
-4.1%
+1.01%
+52.95%
+32.89%
8
In composing this opinion, the Court had to rely on an unofficial transcript. Citations to the unofficial transcript
are denoted as (Tr. Day of Testimony: Page of Transcript).
9
Source: Plaintiffs’ Ex. 24.
18
Precinct 2
Total Pop.
Anglos
Blacks
Latinos
Asians/Others
Voting
Age
Pop.
Anglos
Blacks
Latinos
Asians/Others
2000
847,078
312,717
66,237
441,508
26,616
100%
36.92%
7.82%
52.12%
3.14%
2010
888,572
248,481
77,933
533,812
28,346
100%
28.0%
8.8%
60.1%
3.2%
Change
+41,494
+4.90%
-64,236
-20.54%
+11,696
+17.66%
+92,304
+20.91%
+1,730
+6.5%
584,279
100%
626,274
100%
+41,995
+7.19%
242,420
44,442
278,054
19,363
41.49%
7.61%
47.59%
3.32%
203,286
55,820
345,683
21,485
32.5%
8.9%
55.2%
3.4%
-39,134
+11,378
+67,629
+2,122
-16.14%
+25.60%
+24.32%
+10.96%
100%
38.1%
14.9%
35.3%
11.7%
Change
+259,244
+29.14%
-8,695
-1.95%
+62,610
+57.72%
+168,687
+71.12%
+36,642
+37.53%
Precinct 3
Total Pop.
Anglos
Blacks
Latinos
Asians/Others
Voting
Age
Pop.
Anglos
Blacks
Latinos
Asians/Others
2000
889,646
446,368
108,465
237,189
97,624
100%
50.17%
12.19%
26.66%
10.97%
2010
1,148,890
437,673
171,075
405,876
134,266
644,038
100%
830,382
100%
+186,344
+28.93%
342,879
71,611
156,444
73,104
53.24%
11.12%
24.29%
11.35%
344,183
117,703
266,428
102,068
41.4%
14.2%
32.1%
12.3%
+1,304
+46,092
+109,984
+28,964
+0.38%
+64.36%
70.3%
39.62%
100%
43.4%
16.0%
33.3%
7.3%
Change
+287,871
+34.41%
+3,456
+0.71%
+85,964
+91.88%
+170,086
+83.27%
+28,365
+52.74%
Precinct 4
Total Pop.
Anglos
Blacks
Latinos
Asians/Others
Voting
Age
Pop.
Anglos
Blacks
Latinos
Asians/Others
2000
836,513
484,909
93,563
204,262
53,779
100%
57.97%
11.18%
24.42%
6.42%
2010
1,124,384
488,365
179,527
374,348
82,144
594,809
100%
807,730
100%
+212,921
+35.8%
364,501
60,803
130,907
35,598
61.28%
10.22%
22.01%
6.49%
386,373
120,970
239,955
60,432
47.8%
15.0%
29.7%
7.5%
+21,872
+60,167
+109,048
+24,834
+6.0%
+98.95%
+83.3%
+69.76%
19
Over the course of the decade, the racial composition of the four commissioner’s precincts,
individually, as well as the racial composition of the County, as a whole, changed dramatically. See (Tr.
3: 185-87). Between 2000 and 2010, approximately 13,143 Anglo residents moved out of Precinct 1,
reducing the Anglo population from 188,270 in 2000 to 175,127 in 2010. (Plaintiffs’ Ex. 24, at 1). The
175,127 Anglo residents who remained comprised 18.8% of the total population of Precinct 1 in 2010.
(Plaintiffs’ Ex. 24, at 1); (Tr. 3:185). During that same period, the African-American population of
Precinct 1 contracted by 16,705 residents (or 6.98%), decreasing the African-American population from
363,515 in 2000 to 346,810 in 2010. In 2010, the 346,810 African-American residents who remained
accounted for 37.3% of the total population of Precinct 1. (Plaintiffs’ Ex. 24, at 1); (Tr. 3:185). Over the
course of the same period, the Latino population of Precinct 1 grew considerably. Between 2000 and
2010, an additional 93,712 Latinos took residence in Precinct 1, increasing Precinct 1’s Latino
population from 263,792, in 2000 to 357,504 in 2010. By 2010, the 357,504 Latinos taking residence in
Precinct 1 accounted for 38.4% of the entire population of Precinct 1. (Plaintiffs’ Ex. 24, at 1); (Tr.
3:185).
Over the same period, Precinct 2’s Anglo population declined precipitously; some 63,236 Anglo
residents left Precinct 2, reducing the Anglo population from 312,717 in 2000 to 248,481 in 2010.
(Plaintiffs’ Ex. 24, at 2); see also (Tr. 3:187). As of 2010, the 248,481 Anglo residents who remained
comprised 28.0% of the total population of Precinct 2. (Plaintiffs’ Ex. 24, at 2). From 2000 to 2010, the
African-American population of Precinct 2 grew by 11,696 residents, increasing the African-American
population from 66,237 in 2000 to 77,933, in 2010. (Plaintiffs’ Ex. 24, at 2). In 2010, the 77,933
African-American residents of Precinct 2 accounted for 8.8% of the total population of Precinct 2.
(Plaintiffs’ Ex. 24, at 2). The Latino population of Precinct 2 blossomed as the decade unfolded.
Between 2000 and 2010, an additional 92,304 Latinos took residence in Precinct 2, increasing Precinct
2’s Latino population from 441,508 in 2000 to 533,812 in 2010. (Plaintiffs’ Ex. 24, at 2); see also (Tr. 3:
20
186). By 2010, the 553,812 Latinos residing in Precinct 2 accounted for 60.1% of the entire population
of Precinct 2. (Plaintiffs’ Ex. 24, at 2).
In Precinct 3, approximately 8,695 Anglo residents moved out of Precinct 3 over the course of
the decade, reducing the Anglo population from 446,368 in 2000 to 437,673 in 2010. (Plaintiffs’ Ex. 24,
at 3). In 2010, the 437,673 Anglo residents who remained comprised 38.1% of the total population of
Precinct 3. (Plaintiffs’ Ex. 24, at 3). During that same period, the African-American population of
Precinct 3 grew by 62,610 residents, increasing the African-American population from 108,465 in 2000
to 171,075 in 2010. (Plaintiffs’ Ex. 24, at 3). In 2010, the 171,075 African-American residents of
Precinct 3 accounted for 14.9% of the total population of Precinct 3. (Plaintiffs’ Ex. 24, at 3). Over the
course of the same period, the Latino population of Precinct 3 swelled. Between 2000 and 2010, an
additional 168,687 Latinos took residence in Precinct 3, increasing Precinct 3’s Latino population from
237,189 in 2000 to 405,876 in 2010. (Plaintiffs’ Ex. 24, at 3); see also (Tr. 3:187). By 2010, the 405,876
Latinos residing in Precinct 3 accounted for 35.3% of the entire population of Precinct 3. (Plaintiffs’ Ex.
24, at 3).
Between 2000 and 2010, 3,456 Anglo residents moved to Precinct 4, increasing the Anglo
population of Precinct 4 from 484,909 in 2000 to 488,365 in 2010. (Plaintiffs’ Ex. 24, at 4). By 2010,
Anglo residents comprised 43.4% of the total population of Precinct 4. (Plaintiffs’ Ex. 24, at 4). During
that same decade, the African-American population of Precinct 4 grew by 85,964 residents, increasing
the African-American population from 93,563 in 2000 to 179,527 in 2010. (Plaintiffs’ Ex. 24, at 4). In
2010, the 179,527 African-American residents of Precinct 4 accounted for 16.0% of the total population
of Precinct 4. (Plaintiffs’ Ex. 24, at 4). Over the course of the same period, an additional 168,687
Latinos took residence in Precinct 4, increasing Precinct 4’s Latino population from 204,262 in 2000 to
374,348 in 2010. (Plaintiffs’ Ex. 24, at 4); see also (Tr. 3:188). By 2010, the 374,348 Latinos residing in
Precinct 4 accounted for 35.3% of the entire population of Precinct 4. (Plaintiffs’ Ex. 24, at 4).
21
As Dr. Murray explained, over the course of the ten-year period beginning in 2000 and ending in
2010, Harris County experienced exponential growth. See (Tr. 3:188). That growth was fueled, in part,
by robust growth in the Latino population. Although each of the four precincts experienced strong
Latino growth, the most significant growth among the Latino population occurred in the western and
northern portions of the County, in Precincts 3 and 4, respectively. (Tr. 3:188); see also (Plaintiffs’ Ex.
24). Although Precinct 2 had the largest total Latino population, it had the smallest growth among
Latinos. (Tr. 3:187-88). In fact, Precinct 2’s Latino growth rate of 20.91%, which resulted in an
additional 92,304 Latino residents in Precinct 2 by 2010, paled in comparison to the Latino growth rate
in: (i) Precinct 1, which experienced a 35.52% increase in Latino population over the decade, resulting
in an additional 93,712 Latino residents in Precinct 1, (ii) Precinct 3, which experienced a 71.12%
increase in Latino population over the course of the decade, resulting in an additional 168,687 Latino
residents in Precinct 3, and (iii) Precinct 4, which experienced an 83.27% increase in Latino population,
resulting in an additional 170,086 Latino residents in Precinct 4. See (Plaintiffs’ Ex. 24, at 1-4). The
graphs that follow depict the changes in the Latino distribution:
22
LATINO POPULATION GROWTH10
The western and central-western portions of the County experienced major growth in their
Latino population, such that by the end of the decade all four precincts boasted sizable Latino
populations.
10
Source: Defendants’ Ex.14, at 12, Fig. 12.
23
LATINO DISTRIBUTION IN HARRIS COUNTY11
Of course, this surge in Latino population in the northern and western portions of the County did
not merely affect the distribution of the County’s Latino residents, but it also impacted the county-wide
population trends. The county-wide population trends tracked the Latino population trends, such that the
county-wide population, just like the Latino population, experienced its most robust growth in the
western and northern portions of the County, in what was then commissioner’s Precincts 3 and 4,
respectively. (Tr. 3: 230). This boom in population left Precincts 3 and 4 considerably over-populated
while leaving Precincts 1 and 2 significantly under-populated by the end of the decade. (Tr. 3:181-82).
11
Source: Defendants’ Ex. 14, at 13, Fig. 1.
24
POPULATION CHANGE IN HARRIS COUNTY COMMISSIONER PRECINCTS: 2000- 201012
Precinct
Population in
Population in
Change in Population
Deviation from 2010
2000
2010
Mean
1
827,341
930,613
+ 103,272 (+12.48%)
- 92,502 (- 9.04%)
2
847,078
888,572
+ 41,494 (+ 4.90%)
-134,543 (- 13.15%)
3
889,646
1,148,890
+ 259,244 (+ 29.14%)
+125,775 (+12.29%)
4
836,513
1,124,384
+ 287,871 (+ 34.41%)
+101,269 (+9.9%)
Between 2000 and 2010, Precinct 4 experienced a population increase of almost 288,000 people
resulting in a 34% increase from the 2000 population. (Tr. 3:184); (Plaintiffs’ Ex. 23). Over the same
period, Precinct 3, like Precinct 4, also experienced a dramatic population increase of almost 259,000
people, resulting in a 29% population increase between 2000 and 2010. (Tr. 3:184); (Plaintiffs’ Ex. 23).
Precincts 1 and 2 grew, but their growth did not rival the robust growth of Precincts 3 and 4. In Precinct
1, for instance, the population grew by 12.8% between 2000 and 2010, resulting in a population increase
of 103,000 persons over the course of the decade. (Tr. 3:184). Although Precinct 2 grew, its population
grew at a much slower rate than the populations of either Precincts 1, 3, or 4. Over the course of the
decade, Precinct 2 only experienced a 4.9% growth in population, which resulted in an additional 41,000
persons living in Precinct 2 at the end of the decade. (Tr. 3:184); (Plaintiffs’ Ex. 23).
Under the one-man, one-vote principle, the County’s population should be equally divided
amongst the four commissioner precincts. See (Tr. 3:183). According to the 2010 census, there were
4,092,000 people living in the County. (Tr. 3: 184). In light of these population equality principles, each
precinct should have a population mean of 1,023,000 people. As the chart above indicates, Precincts 1
and 2 were well below the ideal population mean. In Precinct 1, the County would have to increase the
population of the precinct by approximately 92,500 people in order attain the population mean. (Tr. 3:
184); see also (Plaintiffs’ Ex. 23). In order for Precinct 2 to attain the population mean, the County
would have to increase the population by more than 134,000 people. (Plaintiffs’ Ex. 23); (Tr. 3: 184).
12
Source: Plaintiffs’ Ex. 24. Note that Plaintiffs’ Ex. 23 assumes a population mean of 1,023,114.75; for
simplicity the Court has simply used a population mean of 1,023,115.
25
Conversely, Precincts 3 and 4 were over-populated. In order for the County to attain the population
mean in Precinct 4, the County would have to remove more than 125,000 people from the precinct and
relocate them to another precinct. In order for Precinct 4 to attain the population mean, the County
would have to remove 101,000 people from the precinct and relocate them to another precinct. See
(Plaintiffs’ Ex. 23).
In addition to this surge in the overall population, the Latino citizen-voting age population also
grew dramatically.
Summary of the Changes in Latino Citizen Voting Age Population in Harris County
2000
2009
Change
Percentage Change
Harris Co. CVAP
1,964,970
2,281,093
316,123
16%
Latino CVAP
374,227
541,411
167,184
44%
%
Latinos
Among
19%
24%
4.7%
25%
CVAPs
CVAP Growth in Harris Co.
Total
316,123
Latino
167,184
% Latino of Growth
53%
See (Plaintiffs’ Ex. 39, at 2). From 2000 to 2009, the Latino citizen voting age population grew by
167,184 additional Latino adult citizens, growing from 374,227 persons in 2000 to 541,411 persons in
2009, resulting in a 44% increase in the Latino citizen voting age population over the course of the
decade. During that same time the total citizen voting age population of Harris County increased from
1,964,970 in 2000 to 2,281,093 through the end of 2009, representing an estimated total growth of
316,213, or 16%. Of those 316,123 individuals, 167,184 were Latinos. Thus, Latinos accounted for 53%
of Harris County’s citizen voting age population growth during the first decade of the new millennium.
In examining how the Latino population changed over the course of the decade, Plaintiffs’ expert
Dr. Matt A. Barreto13 explained that, when compared to the 2001 Map, the Revised Plan A-1 boundaries
13
Plaintiffs offered the testimony of Dr. Matt A. Barreto as an expert witness on the Gingles factors. Dr. Barreto
received a B.S. in political science from Eastern New Mexico University, an M.S. in social science from the
University of California—Irvine, and a Ph. D. in political science from the University of California—Irvine. Since
2009, Dr. Barreto has served as an associate professor of political science at the University of Washington. Dr.
Barreto also serves as the director of the Washington Institute for the Study of Ethnicity and Race (WISER) and is
affiliated with the Center for Statistics and Social Science at the University of Washington. See (Plaintiffs’ Ex.
26
reduce the overall Latino population and the Latino voting age population in Precinct 2. If the 2001 Map
remained in place, then by 2010, 60.1% of Precinct 2’s population would be Latino and 55.2% of the
precinct’s voting age population would be Latino. Under the Revised Plan A-1, however, only 58.2% of
Precinct 2’s population is Latino and 53.4% of the precinct’s voting age population is Latino. Further,
Dr. Barreto notes that county-wide the percentage of the population comprised by Latino residents
increased by 19%, from 32.9% in 2000 to 40.8% in 2010. However, under the Revised Plan A-1,
Precinct 2’s Latino fraction of the population only increased by 10.5%. (Plaintiffs’ Ex. 39, at 3).
Moreover, Dr. Barreto explained that the Revised Plan A-1 will also reduce the number of
registered voters in the precinct. Under the 2001 map, 34.9% of Precinct 2’s citizen voting age
population is Latino and 29.7% of registered voters are Spanish surnamed individuals (“SSRV”).14
Under the Revised Plan A-1, only 33.8% of Precinct 2’s citizen voting age population is Latino and
28.8% of registered voters are Spanish surnamed individuals. (Plaintiffs’ Ex. 39, at 3).
D.
REDISTRCTING: PRINCIPLES & PRIORITIES
1.
Principles & Priorities
The first re-districting challenge to the Harris County Commissioner’s Court occurred in 1981.
The litigation waged in connection with the 1981 redistricting ultimately resulted in the creation of
Precinct 1 as an African-American opportunity district. (Tr. 3:176). In 1991, Dr. Murray assisted the
County in making minor changes to the 1981 map to accommodate the minor population disparities that
had formed since the authoring of the map. (Tr. 3:176). The changes, though, were relatively modest and
were intended to preserve Precinct 1’s status as an opportunity district. (Tr. 3:176). In 2001, Dr. Murray,
once again, served as a general consultant in the redistricting process, advising the County on the ways
11). Dr. Barreto’s research focuses on the voting patterns of Latinos in the United States, the ways in which nonLatinos perceive Latinos, and how groups interact in society. Dr. Barreto has performed extensive research on the
voting patterns of Latinos as relate to the Voting Rights Act. In fact, Dr. Barreto has authored a book examining,
in part, the political cohesion of the Latino community in Harris County. Dr. Barreto has testified as an expert in
various other voting rights cases.
14
SSRV is a proxy for the number of Latino registered voters.
27
in which to draw a map that complied with the constitution, the Voting Rights Act, and the County’s
Priorities and Principles. See (Tr. 3:177). In the 2001 round of redistricting, the County’s main goal was
to cure the population imbalance; in 2001, as was the case in the most recent round of redistricting, the
population growth in the eastern portion of the County lagged behind the growth numbers in the countywide growth, leaving Precinct 2 under-populated. (Tr. 3:178). To remedy the under-population
problems, the County extended the Precinct northward above downtown into the heavily populated area
along Hardy Road, Eastex Freeway. (Tr. 3:179). In 2001, the map-drawing process was guided by a
statement of priorities and principles for redistricting of Harris County Commissioner’s precincts. (Tr.
3:179-180).
After the 2010 census, it was clear that the County would have to engage in another round of
redistricting to remedy the population imbalance between the precincts that had occurred over the course
of the previous decade.
On June 21, 2011, the Commissioner’s Court issued an order pronouncing the priorities and
principles to be applied in the redistricting of the Harris County Commissioner’s Precincts (the
“Priorities and Principles”). (Defendants’ Ex. 3). These Priorities and Principles were adopted at the
inception of the redistricting process; the County did not initiate in any map drawing until after the
adoption of the Priorities and Principles. (Tr. 3:188-89).15
The first priority and principle requires the four commissioner precincts be of substantially equal
population and, in no event, exceed a 10% top-to bottom deviation. (Defendants’ Ex. 3, at 4). Dr.
Murray testified that the County implemented this principle to ensure that the County’s map did not
violate the Constitution’s one man, one vote principle, which requires population equality. See (Tr. 3:
190).
15
Dr. Murray explained that the 2011 priorities and principles were essentially the same priorities and principles
adopted for the 2001 process. (Tr. 3:195); compare (Defendants’ Ex. 3, at 4) with (Defendants’ Ex. 4, at 1).
28
The second priority and principle requires that the four commissioner precincts be contiguous
and reasonably geographically compact. (Defendants’ Ex. 3, at 4). Dr. Murray testified that the second
priority, like the first, is designed to ensure that the County’s map does not violate constitutional
redistricting safeguards. See (Tr. 3:190).
The third priority and principle states that, to the extent possible, the redistricting plan will use
identifiable geographic boundaries as precinct boundaries, preserve natural historical boundaries,
recognize identifiable communities of interest in a single precinct and avoid splitting neighborhoods
when drawing precinct lines. (Defendants’ Ex. 3, at 4). According to Dr. Murray, the third priority
recognizes that redistricting should be conducted in a manner that respects the conveniences and the
communal integrity of the people being governed. (Tr. 3:191). In Dr. Murray’s view, the third priority’s
admonishment against splitting communities of interest and heeding natural and/or historic borders
accomplishes that goal of respect for the communal integrity of the people being governed. (Tr. 3:191).
The fourth priority and principle provides that to the extent possible, the four commissioner
precincts should be based on existing composition of the precincts. (Defendants’ Ex. 3, at 4). Dr. Murray
explained that the County’s adoption of this principle encourages the parties tasked with redistricting to
use the 2001 map as a starting point for the redistricting process. (Tr. 3:191). Dr. Murray further
explained that, in the absence of gerrymandering, the most recent previously employed map is often
used as the basis for the redistricting process with modifications made to make the old map legal, in
light of the intervening population changes in the years since that map was first adopted. See (Tr. 3:19192).
The fifth priority and principle requires the redistricting plan use whole county voting precincts
to draw commissioner’s precincts. (Defendants’ Ex. 3, at 4). Dr. Murray explained that when a voting
precinct is split, it is divided amongst two commissioner’s precincts. Therefore, this priority is designed
to minimize the impact of redistricting on the individual voting precincts by preserving the voting
precincts as whole units, thereby limiting the need for realignment at the local voting place. (Tr. 3:192).
29
The sixth priority and principle requires that the redistricting plan adhere to (i) the Equal
Protection Clause of the Fourteenth Amendment to the United States Constitution and not use race as the
predominate factor in establishing commissioner precincts and (ii) the Voting Rights Act and avoid (a)
retrogression in the position of racial, ethnic and language minorities with respect to the effective
exercise of their right to vote or electoral franchise and (b) dilution of the voting strength of racial or
language minority citizens. (Defendants’ Ex. 3, at 4). Dr. Murray explained that this priority describes
the delicate balancing of competing legal principles involved in the redistricting process. On the one
hand, the Fourteenth Amendment prohibits the County from drawing a map solely on the basis of race,
but on the other hand, compliance with the Voting Rights Act requires the County to take heed of the
racial composition of the map in order to ensure that the map is neither retrogressive nor dilutive. See
(Tr. 3:193).
The seventh priority and principle states the redistricting plan should, in recognition of the value
of incumbent-constituent services, seek to keep (i) existing commissioners in their existing precincts and
(ii) facilities and service locations established by incumbent commissioners in the precincts of those
commissioners. (Defendants’ Ex. 3, at 4). Dr. Murray testified that this precept recognizes that in areas
with both a sizeable urbanized population and a sizeable unincorporated area, like Harris County,
service delivery and facilities takes on greater importance than it might in more urbanized counties or in
purely legislative bodies. (Tr. 3:193-94). Harris County is unique in that it is one of the only counties in
the nation that features a highly urbanized population that does not reside in incorporated cities. (Tr.
3:182). In Harris County, a sizeable faction of the population lives in unincorporated areas—i.e. territory
within the bounds of Harris County that is not within the boundaries of a municipality. (Tr. 3:181-82). In
2010, approximately 1.6 million residents of Harris County lived in unincorporated areas. (Tr. 3:182). In
unincorporated areas, basic services, such as road maintenance and facilities management, which
ordinarily would be performed by a municipal government, are instead provided by the Harris County
Commissioner’s Court. (Tr. 3:183). According to Dr. Murray, this responsibility makes the delivery of
30
constituent services more important in Harris County than it might be in other counties or purely
legislative bodies where representatives are not responsible for the delivery of services. (Tr. 3:194).
The eighth priority and principle states that the redistricting plan should recognize the duties and
obligations imposed by law on commissioners to provide services to the residents of precincts and the
public investment in facilities, service locations and personnel that has been made to provide such
services. (Defendants’ Ex. 3, at 4). Dr. Murray testified that the County and its residents have made
large public investments in the County facilities and services and the redistricting plan should honor that
investment. (Tr. 3:194).
Dr. Murray testified that the eight priorities and principles are not always harmonious in that
adherence to one priority often necessitates the subordination of another priority. See (Tr. 3:192). Dr.
Murray explained that these eight principles outline the goals of the redistricting process, but the
application of these priorities and principles always involves a give and take. See (Tr. 3:192). The
significance accorded to any single priority must, of course, be balanced against the importance of the
other priorities. (Tr. 3:192). Thus, except for a few absolutes, such as the prohibition against a
population deviation in excess of 10%, there is always an interplay between the principles and which
priorities will take precedence in any given scenario. See (Tr. 3:192). Inevitably, sometimes the County
must temper its adherence to one principle in order to preserve the integrity of the map as a whole. (Tr.
4: 40).
2.
Opportunity Districts and Influence Districts
Dr. Murray described Precinct 1 as an opportunity district but described Precinct 2 as an
influence district. (Tr. 3:196). Dr. Murray explained that an opportunity district is a district where a
protected minority can elect a candidate of their choice, although not necessarily of their race or
ethnicity. In his view, a functioning opportunity district requires a sufficient population, number of
registered voters, voter turnout, and cohesiveness. Dr. Murray stated that Precinct 1 has met this
standard in every election since 1981. (Tr. 3:196). That said, Dr. Murray recognized that the population
31
changes that had occurred over the course of the last decade threatened Precinct 1’s ability to perform as
an opportunity district. In 2000, almost 44% of the population of Precinct 1 was African-American.
(Plaintiffs’ Ex. 24, at 1); (Tr. 3:234). At the end of the decade, African-Americans only accounted for
37.3% of Precinct 1’s population. (Plaintiffs’ Ex. 24, at 1); (Tr. 3:235). The evidence showed that the
African-American population in Precinct 1 was on a downward trend. (Tr. 3:235). In order to preserve
Precinct 1’s status as a functioning opportunity district, the County would have to augment the AfricanAmerican population of Precinct 1 to preserve this opportunity district.
On the other hand, Dr. Murray testified that Precinct 2 is a district that does not have a protected
minority with sufficient population, registered voter population, voter turnout, and cohesion to elect a
candidate of choice. (Tr. 3:196-97). Dr. Murray explained, for instance, that at the time of the 2001
redistricting, he did not believe Precinct 2 was a Latino opportunity district even though it experienced
significant Latino growth over the course of the 1990s. (Tr. 3:178). Today, as was the case in 2001, Dr.
Murray persists in his assessment that Precinct 2 is not an effective Latino opportunity district. See (Tr.
3:226).
In a memorandum authored by Dr. Murray, titled “Is Harris County Commissioner Precinct 2 an
‘Effective Latino Opportunity District’?” Dr. Murray examined the results of the 2002, 2006, and 2010
general elections to evaluate whether Precinct 2 was a Latino opportunity district. (Defendants’ Ex. 15,
at 3).16 Dr. Murray explained that the data from these three election cycles demonstrated that the Latinopreferred candidate, who in every analyzed election was also the Democratic nominee, did not perform
well in the general election. In 2002, “with the Latino turnout boost provided by Tony Sanchez’s wellfunded gubernatorial campaign, the [Latino]-backed candidates prevailed in 23 of 58 contests in Precinct
2.” (Defendants’ Ex. 15, at 4); (Tr. 3:229). In 2006, Latino-preferred candidates only won six of the 27
contests in Precinct 2. (Defendants’ Ex. 15, at 4); (Tr. 3:229). Finally, in 2010, only one Latino16
Dr. Murray chose these elections because they were conducted under the 2001 map and the ’02, ’06, and ’10
election cycles are the elections in which voters in Precinct 2 elect their county commissioner. (Defendants’ Ex.
15, at 3).
32
preferred candidate, former Houston mayor Bill White, prevailed out of the 81 contested elections.
(Defendants’ Ex. 15, at 4): (Tr. 3:229).
PERFORMANCE OF LATINO-PREFERRED CANDIDATES IN PRECINCT 2 (AS
FORMULATED UNDER THE 2001 MAP) IN THE 2002, 2006, and 2010 ELECTION CYCLES17
Election
Number of
Number of LatinoNumber of Latino% of LatinoYear
Contested
Preferred
Preferred
Preferred
Elections in
Candidates that
Candidates Defeated
Candidates
Precinct 2
Prevailed
Prevailing
2002
58
23
35
39.7%
2006
27
6
21
22.2%
2010
81
1
80
1.2%
TOTAL
166
30
136
18.1%
According to Dr. Murray, this data clearly shows that Precinct 2 is not a Latino opportunity
district because the Latino share of Precinct 2’s citizen voting age population and the percentage of
Spanish-surname registered voters in the precinct are too low to enable Latino voters to elect the
candidates of their choice. (Defendants’ Ex. 15, at 4). Dr. Murray testified that the low turnout amongst
Latinos makes it “impossible for any candidate to win in Precinct 2 by [only] mobilizing Latinos.”
(Defendants’ Ex. 15, at 4). Dr. Murray thus concluded that “even if Latino voters were extremely
cohesive and had very high turnout rates vis-à-vis other voters in Precinct 2, it would be extremely
difficult for Latinos alone, to elect candidates of their choice given such low citizen voting age
population and SSRV percentages.” (Defendants’ Ex. 15, at 1); see also (Tr. 3:226-27).
Occasionally, however, Latino-preferred candidates can win in Precinct 2, but they win as a part
of broader coalition that includes African-Americans, cross-over Anglos, and occasionally a few AsianAmerican voters. (Tr. 3:227); (Defendants’ Ex. 15, at 4). According to Dr. Murray, Senator Garcia’s
2002 electoral victory was the result of a coalition amongst Latinos, African-Americans, and inner-city
Anglos in Precinct 2. See (Defendants’ Ex. 15, at 4); (Tr. 3:227). And even in that 2002 contest, where
Senator Garcia received 52.8% of the vote, less than half of her support came from Latinos.
17
Source: Defendants’ Ex. 15, at 4.
33
(Defendants’ Ex. 15, at 4). In 2002, for instance, Latinos only made up 25% of the total vote in Precinct
2. (Defendants’ Ex. 15, at 2). These coalitions, though occasionally effective, are also quite rare. In more
than 80% of the contested elections in 2002, 2006, and 2010, the Latino-preferred candidate could not
secure a majority of the vote in Precinct 2. (Defendants’ Ex. 15, at 4).
Nevertheless, the Latino population is sizeable, and this sizeable number of minority voters can
be influential in the election of a commissioner. (Tr. 3:197). Accordingly, Dr. Murray concluded that
Precinct 2 is currently a Latino influence district rather than an opportunity district.
Dr. Murray further opined that transforming Precinct 2 into an performing opportunity district
would be difficult because: (i) Harris County’s Latino population is geographically dispersed throughout
the County and (ii) Latinos, generally, have low rates of political participation making it difficult for
them to elect their preferred candidates. See (Tr. 3: 227-231).
With respect to geographic dispersion, Dr. Murray explained that Harris County’s Latino is
spread out over the four commissioner’s precincts, making it difficult to fashion a compact district in
which Latinos constitute a majority of the citizen voting age population. (Tr. 3:230-31). Thus, for
example, the distance between the growing Latino populations in the northern and western portion of the
County and the established Latino population in the eastern portion of the County, makes it difficult to
draw a compact and contiguous map that joins both populations. (Tr. 3:231). In his opinion, the
geographic dispersion of Harris County’s Latino community makes it nearly impossible to create a
geographically compact district, drawn in accordance with traditional redistricting principles, where
Latinos constitute a majority of the citizen voting age population. (Tr. 3:233).
Second, he claims that Latinos in Precinct 2 have low voter turnout, which precludes them from
electing their candidates of choice. See (Defendants’ Ex. 15, at 1); (Tr. 3:227). Dr. Murray explained
that in 2010, before the adoption of the Revised A-1, 60.1% of the population in Precinct 2 was Latino,
but only 34.9% of the citizen voting age population of Precinct 2 was Latino, (Plaintiffs’ Ex. 39, at 4);
see also (Defendants’ Ex. 15, at 1), and only 29.7% of Spanish-surname individuals in Precinct 2 were
34
registered to vote. (Tr. 3:227); see also (Plaintiffs’ Ex. 39, at 3). The low percentage of Spanish surname
registered voters and the relatively small number of Latino adult citizens in Precinct 2 limits the ability
of Latinos to elect, by themselves, a candidate of their choice. (Tr. 3:227).
3.
The County’s Approach to the 2011 Redistricting
In Dr. Murray’s view, there were several issues facing the County during this redistricting
process. See (Tr. 3:195-96). First, the County had to equalize the distribution of the population across
the commissioner’s precincts. At the end of the previous decade, there was too much deviation in
population across the four precincts, so this time redistricting would require redistribution of a much
larger share of the population than in prior redistrictings. (Tr. 3:195). Precincts 1 and 2 were underpopulated, whereas Precincts 3 and 4 were over-populated. Under the 2001 map, Precinct 2 shared most
of its border with Precinct 1. The County therefore anticipated the possibility of removing population
from either Precinct 1 or Precinct 2 and adding it to the other precinct, which would of course worsen
the under-population issues of the donating precinct. (Tr. 3:203). Dr. Murray further explained that
Precinct 3 does not share a border with Precinct 2 and intimated that the Latino population in Precinct 3
was too geographically remote to remedy Precinct 2’s under-population problem. See (Tr. 3:187). This
left Precinct 4, the only over-populated precinct adjacent to Precinct 2, as the most practical source of
population to cure Precinct 2’s under-population problem.
Second, the County had to preserve Precinct 1’s status as an effective minority-opportunity
district. This would be a challenge because the black population in Precinct 1 was declining, while the
Latino population in the precinct was growing. (Tr. 3:196). According to Dr. Murray, Precinct 1’s
negative African-American growth rate combined with the dramatic Latino growth rate made the task of
maintaining Precinct 1’s status as an effective opportunity district more challenging than in prior redistrictings. (Tr. 3:186). In 2000, for instance, African-Americans accounted for 44% of the population
of Precinct 1, but after the intervening decade they now only accounted for 37.3% of the population.
35
Dr. Murray testified that the historic census data and population trends indicated that Latinos
were moving into areas that were predominately African-American. (Tr. 4:35). In keeping with this
trend, Latinos have migrated southward over the course of the last few decades, displacing traditionally
African-American pockets of Harris County. (Tr. 4:35). This trend of Latino settlement in the southern
portion of the County has changed the racial composition and complexion of the County, such that the
Alief, Sunny Side, South Union, and South Park areas, areas which were historically African-American,
are trending more multicultural and increasingly Latino, with a decreasing share of the population being
African-American. (Tr. 4:36). Similarly, Latinos have moved into the area between I-45 and I-59, which
had been referred to as “Redneck Alley,” in the north central part of the County, displacing the formerly
Anglo-populated area. (Tr. 4:35).
Dr. Murray noted that in his years of experience, once an area begins to trend Latino it does not
reverse course; he has never observed an area that was predominately non-Latino and then became
predominately Latino revert back to being predominately non-Latino. (Tr. 4:36). Thus, if an area, like
Alief, is currently trending Latino, it is improbable that there will be an influx of non-Latinos (AfricanAmericans or otherwise) over the next 10 to 20 years to reverse the trend of Latino settlement. (Tr.
4:36). In preserving Precinct 1 as an opportunity district, the County had to remain cognizant of these
historic population patterns.
Third, with respect to Precinct 2, the County had to address the under-population problem
plaguing the precinct, which meant that the County would have to find some way to increase the
population of Precinct 2 by at least 100,000 people, (Tr. 3:198), while also trying to preserve Latino
influence (Tr. 3:197). Dr. Murray appreciated that the mechanics of preserving Latino influence in
Precinct 2 might be different under the new Republican Precinct 2 Commissioner Jack Morman than
they were in the past under the former Democratic Precinct 2 commissioners. Dr. Murray suggested that
under the new commissioner, the incumbency-protection considerations were probably different than
they were under the former Democratic commissioners. (Tr. 3:197; Tr. 4:43).
36
It was clear that the four commissioners all had an interest in the outcome of the redistricting
process, with each commissioner taking a keen interest in the way in which the redistricting would alter
their precincts. (Tr. 3:198). Given that and the major issues related to population distribution, the
preservation of Precinct 1’s status as an opportunity district, and the preservation of minority influence
in Precinct 2, the County understood the redistricting process would involve balancing the Priorities and
Principles with the desires of the commissioners. (Tr. 3:198-99).
4.
Performing the Redistricting
The County hired Dr. Murray to assist it with the redistricting process. Dr. Murray explained that
he began by first considering how to equalize the distribution of the population across the four
commissioner’s precincts. (Tr. 3:199). To that end, Dr. Murray drafted an email on June 18, 2011,
identifying eight areas, mostly along the borders of the precincts, that were population rich and therefore
could be appended to the adjacent precinct, to help correct the population imbalance. (Defendants’ Ex.
66); see also (Tr. 3:199).
a.
Proposed Changes
The first locale Dr. Murray considered was the area east of Lake Houston in northeast Harris
County. The six precincts comprising the area east of Lake Houston had a total population of 32,000.
(Defendants’ Ex. 66, at 1). Moreover, this area was in Precinct 4, which was overpopulated, and shared
a border with Precinct 2, which was under-populated. (Tr. 3:200). Thus, this area east of Lake Houston
presented an opportunity for a simple swap between Precinct 4 and Precinct 2.18
The second area he considered was the Atascosita area, west of Lake Houston, because this area
was in Precinct 4 and bordered the under-populated Precinct 2. The six voting precincts comprising the
Atascosita area had a total population of approximately 24,000. See (Defendants’ Ex. 66, at 2); see also
(Tr. 3:201).
18
In this section, whenever the Court refers to the population of one of the eight areas, the Court is referring to the
population in that area as of 2010.
37
Third, Dr. Murray considered the Kingwood area, in north Harris County, because the twelve
voting precincts comprising the Kingwood area had a population of 56,000 and this area was close to
both Precincts 1 and 2. (Defendants’ Ex. 66, at 3); (Tr. 3:201).
Fourth, Dr. Murray area considered the area located in the far southeast section of Harris County
west of the Gulf Freeway. The eight voting precincts comprising this area have a total population of
28,229 residents. At the time, this area was located in the under-populated Precinct 1, but Dr. Murray
thought these eight precincts could be moved to Precinct 2, thereby mitigating Precinct 2’s underpopulation problem. See (Tr. 3:201); (Defendants’ Ex. 66, at 4).
The fifth area considered was comprised of thirteen precincts in north central Harris County. The
area, then located in Precinct 1, was close to the northern neck of Precinct 2 and boasted a population of
approximately 78,400 people. (Defendants’ Ex. 66, at 5). If added to Precinct 2, this population-rich area
would go a long way toward remedying Precinct 2’s under-population problem. (Tr. 3:202).
The sixth location considered was the Wallisville area north of Interstate 10. The nine voting
precincts comprising the Wallisville area had a total population of 61,769. (Defendants’ Ex. 66, at 6).
This area, then located in Precinct 2, lies along the border of Precinct 1; Wallisville has a growing black
population and most residents are homeowners, rather than renters. (Tr. 3:204). Dr. Murray considered
removing this area from Precinct 2 and adding it to Precinct 1 in an effort to temper Precinct 1’s underpopulation problem. (Tr. 3:204).
The seventh area under consideration was the diverse and densely populated area of Alief. (Tr.
3:207). This area, then located in Precinct 3, was comprised of thirteen voting precincts that collectively
boasted a population of almost 115,000. (Defendants’ Ex. 66, at 7). Dr. Murray testified that he
discussed moving some of the population from Precinct 3’s Alief area to Precinct 1, as he did during the
2001 redistricting. (Tr. 3: 206). In the 2011 redistricting, Commissioner Radack, the commissioner of
Precinct 3, had recommended that the population imbalance between Precinct 1 and 3 be addressed
38
using the same method the County employed in 2001, taking population from the Alief area and adding
it to Precinct 1. (Tr. 3: 206).
The eighth and final area under consideration was comprised of nine voting precincts in the
Greenspoint/ Bush Airport area in north central Harris County. (Tr. 3:207); see also (Defendants’ Ex.
66, at 7). This area, then located in Precinct 4, had a population of approximately 85,300 and had a
sizeable African-American population. (Tr. 3:207); see also (Defendants’ Ex. 66, at 7). This area,
comprised almost entirely of home dwellers, had the fastest-growing African-American population in
Harris County. (Tr. 3:207-08). Given that this area was located in Precinct 4, which was over-populated,
the Bush Airport area was a natural contender for relocation to the Precinct 1, which was underpopulated. See (Tr. 3:207).
b.
Implemented Changes
After conducting this population analysis, Dr. Murray then considered the eight Priorities and
Principles and with these principles and population statistics in mind, Dr. Murray drafted approximately
six preliminary maps for discussion with the attorneys for the County. (Tr. 3:205). Dr. Murray testified
that he mainly interacted with the attorneys but also had some limited contact with the commissioners.
Dr. Murray explained that this is a political process with four commissioners each with his own
concerns. (Tr. 4:23). Dr. Murray admitted that the map would have come out differently if Sylvia Garcia
was the Precinct 2 County commissioner, for she would have encouraged efforts to increase the number
of Democratic voters in Precinct 2. (Tr. 4:43).
Despite these political considerations, Dr. Murray contends that the Revised Plan A-1 respects
the County’s Redistricting Priorities and Principles in that it: (i) substantially equalizes the population,
well within the permissible levels of deviation; (ii) creates contiguous commissioner precincts; (iii) was
developed based on the existing precincts as they were fashioned in the 2001 map; (iv) minimally splits
voting precincts, splitting only a dozen precincts; and (v) minimizes the number of people changing
39
precincts, in that most residents are in the same precinct, under the Revised Plan A-1, as they were in
under the 2001 map. (Tr. 3: 210-12).
In creating the Revised Plan A-1, the County made substantial changes to both Precinct 1 and
Precinct 2.
First, Precinct 1 is extended northward to Farm to Market Road 1960, picking up the sizeable
Black homeowner population living near the Bush Airport. (Tr. 3:212). This area has experienced
substantial African-American growth and the homeowners who populate this area are a very important
voting constituency. (Tr. 3:212).
Second, Precinct 1 also is extended eastward, picking up a sizeable group of voters from the
Wallisville area, another area with a large number of Black homeowners. (Tr. 3:213).
Third, a small portion of the Alief area, which was previously in Precinct 3, was added to the
southwest side of Precinct 1. However, the Revised A-1 did not move all thirteen voting precincts of the
Alief area. (Tr. 3: 212). Dr. Murray explained that even though the County wanted to preserve Precinct
1’s status as an opportunity district, which, of course, required the County to be mindful of the AfricanAmerican population in Precinct 1, the County did not believe that bolstering Precinct 1’s population
with residents of Alief would best preserve Precinct 1’s status as an opportunity district. (Tr. 3:213-14).
According to Dr. Murray, the Alief area was rapidly built in the 1970s and was predominately Anglo,
but after the oil boom bust the Anglos left. In the 1980s and 1990s, this was a major area of Black
growth. As the millennium approached, the Black population in Alief began to move across the county
line into Fort Bend County, but the Latino population in Alief flourished. (Tr. 3:214). Today, the
African-Americans who remain are mostly apartment-dwellers, because this area does not have many
single family dwellings. (Tr. 3:215).
Given these demographic trends, Harris County chose not to go further west into Alief to remedy
Precinct 1’s under-population, but rather it chose to go further north into the Bush Airport area and the
Wallisville area to pick up home-owning African-American populations with strong voter turnout. (Tr.
40
3:216). Again, Dr. Murray stressed that the Bush Airport area and Wallisville area are areas of AfricanAmerican growth and both areas are populated with middle class home-owning voters. Therefore, the
Bush Airport area and the Wallisville area were much better sources of voter population than Alief. (Tr.
3:216).
Fourth, there were also some small swaps in the Hobby Airport area, with one voting precinct,
Precinct 285, moving from Precinct 2 to Precinct 1, and another precinct, Precinct 36, moving from
Precinct 1 to Precinct 2. (Tr. 3:212). Fifth, the Revised Plan A-1 ensures that Precinct 1’s significant
service facilities remain in Precinct 1. Dr. Murray explained that the Revised A-1 map keeps the
Challenger Seven Memorial Park in the southeast portion of Precinct 1, the Deussen Park in the eastern
part of the precinct, and the El Franco Lee Community Center and the Hardy Road facility in Precinct 1,
as well. (Tr. 3:217).
Ultimately, under the Revised A-1, the African-American population of Precinct 1 increased
from 37.3% to 39%, mainly by the addition of the growth area in north central Harris County, near Bush
Airport area. (Tr. 3:236). According to Dr. Murray, the addition of this area will compensate for the
declining African-American population elsewhere and may reverse the trend of declining Black
population in Precinct 1. (Tr. 3:236).
As for Precinct 2, the County had to remedy the precinct’s under-population problem and
preserve Latino influence in the precinct. See (Tr. 3:197-98). The County addressed these issues by
extending the eastern portion of the precinct northward into the existing Precinct 4 and adding thirteen
precincts east and west of Lake Houston, to Precinct 2, and second, the County extended the neck of the
precinct north of downtown. (Tr. 3:218).19
19
As discussed in the preceding paragraphs, the County also transferred some voting precincts from Precinct 2 to
Precinct 1, such as the African-American dominated voting districts near Hobby Airport moved from Precinct 2 to
Precinct 1 and five voting districts in the Wallisville area. (Tr. 3:218).
41
5.
Public Review of the Map: Opportunity for Public Comment
As part of the redistricting process, the County held numerous public hearings before adopting
the Revised Plan A-1. (Tr. 3:220). The County held a hearing in each of the four precincts and then also
held a meeting downtown at the County courthouse, for a total of five. (Tr. 3:220). The first of these five
meetings was in Precinct 2. At that meeting, Dr. Murray made a presentation on the original A-1 map.
(Tr. 3:221). According to Dr. Murray, this meeting was well-attended and the public comments were
polarized, with some attendees applauding the proposed map and others criticizing the County’s
proposal. (Tr. 3:221). This was in sharp contrast to the public hearing that occurred in Precinct 1, in that
most of the attendees at the Precinct 1 meeting favored the proposed changes to Precinct 1. (Tr. 3:22122).
The County was responsive to the concerns expressed by the community and considered the
concerns raised by the County’s residents. For instance, the County considered alternative plans that
were proposed by citizens, such as the Guerra-Jara plan, which was proposed by members of Precinct
2’s Latino community. The County ultimately rejected this plan because the County determined that the
Guerra-Jara plan was inconsistent with the priorities and principles the County had adopted and resulted
in major service disruptions in Precinct 1, which would lose all four major service facilities. (Tr. 3:224).
Although some concerns, like the Guerra-Jara plan, were rejected, others were implemented. For
example, the original A-1 map took two voting precincts from the Anglo populated Kingwood area, in
Precinct 4, and added them to Precinct 2. (Tr. 3:222). After the hearings, these Republican-leaning
Kingwood districts were left in Precinct 4. (Tr. 3:223). The Revised Plan A-1 is the product of those
revisions to the original A-1 map.
6.
Plaintiffs’ Criticism of the Revised Plan A-1
According to Plaintiffs, the Revised Plan A-1 splits five communities of interest, which are
census-designated places: Aldine, Sheldon, Humble, Channelview, and Atastocita. (Tr. 4:26). In
response to these concerns, Dr. Murray first noted that Aldine, Sheldon, Channelview, and Atastocita
42
are not cities, although he conceded that each is, to a limited extent, a community of interest. See (Tr.
4:25-27). Dr. Murray seems to intimate that even if these four locales are census-designated places, they
are not municipalities and therefore splitting them does not carry the same import as splitting a
municipality. See (Tr. 4:27). In his view, these are not well-defined communities. Second, even if
Humble is a city and not merely a census designated place, any fracturing of Humble occasioned by the
Revised Plan A-1 is insignificant because Humble is a small city with a small population with limited
prospects for growth given that the city is circumscribed by the City of Houston. (Tr. 4:25, Tr. 4:41).
Third, Dr. Murray testified that had the County brought in all of Atastocita in the Revised Plan A-1,
rather than split it, that would have driven the Latino numbers down in Precinct 2, as this area is
principally Anglo. (Tr. 4:42). Dr. Murray further testified that the same is true of Humble; if the entirety
of Humble was added to Precinct 2 this would have impacted the Latino numbers, but the effect would
have been minimal given that Humble is not a very large city. (Tr. 4:41-42).
E.
SUMMARY OF PLAINTIFFS’ CLAIMS
In summary, Plaintiffs allege that the County’s Plan violates Section 2 because the Revised Plan
A-1 dilutes the voting strength of Precinct 2’s Latino residents. According to Plaintiffs, Precinct 2
should be a Latino opportunity district. Plaintiffs also claim that the Revised Plan A-1 is unconstitutional
because it was designed with the purpose of discriminating against Latinos and because it was drawn
with an excessive and unjustified use of racial data.
III.
LAW & ANALYSIS: SECTION 2 CLAIMS
Plaintiffs first claim that the Revised Plan A-1 dilutes their voting strength in violation of Section
2 of the Voting Rights Act. In order to prevail on this claim, Plaintiffs must show that: (1) Latinos are a
sufficiently large and geographically compact group to constitute a majority in an additional singlemember district; (2) the affected minority group is politically cohesive; and (3) the white majority votes
sufficiently as a bloc to enable it—in the absence of special circumstances—usually to defeat the
minority group’s preferred candidates. Perry, 548 U.S. at 425; Gingles, 478 U.S. at 50-51; Fairley, 584
43
F.3d at 667. If these three preconditions are met, the district court must then examine a variety of other
factors to determine whether, under the totality of the circumstances, the Revised Plan A-1 impairs the
ability of the minority voters to participate equally in the political process and to elect a representative of
their choice. De Grandy, 512 U.S. at 1011; Fairley, 584 F.3d at 667.
A.
GINGLES ONE: IS THE LATINO POPULATION OF PRECINCT 2
SUFFICIENTLY LARGE AND GEOGRPAHICALLY COMPACT TO CONSTITUTE A
MAJORITY IN A SINGLE MEMBER DISTRICT?
The first Gingles precondition requires the plaintiff to show that the minority group is
sufficiently large and geographically compact to constitute a majority in a single-member district.
LULAC, 548 U.S. at 425; see also Growe, 507 U.S. at 40; Gingles, 478 U.S. at 50-51. When the voting
potential of a minority group that is large enough to form a majority in a district has been thwarted by
the manipulation of district lines, minorities may justly claim that their “ability to elect” candidates has
been diluted in violation of Section 2. Hall v. Virginia, 385 F.3d 421, 429 (4th Cir. 2004).
In order to show that the minority group is sufficiently large and geographically compact to
constitute a majority in a single member district, the plaintiff must demonstrate “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.” De Grandy, 517 U.S. at 1008; accord Perry, 548 U.S. at
429. The first Gingles precondition thus consists of a numerosity requirement and a compactness
requirement. De Grandy 512 U.S. at 1011. Together, the numerosity and compactness components of
the first Gingles precondition ensure that the affected minority group possesses the potential to elect
representatives of its choice, in the absence of the dilutive election practice. Gingles, 478 U.S. at 50
n.17; LULAC III, 986 F.2d at 743; Westwego Citizens for Better Government v. Westwego, 946 F.2d
1109, 1117, 1117 n.9 (5th Cir. 1991) (“Westwego III”).
Because the very concept of vote dilution implies—and, indeed, necessitates—the existence of
an undiluted practice against which the fact of dilution may be measured, a Section 2 Plaintiff will
usually postulate a reasonable alternative voting practice to serve as the benchmark undiluted practice.
44
Reno v. Bossier Parish Sch. Bd., 520 U.S. 471, 480 (1997) (quoting Holder v. Hall, 512 U.S. 874, 881
(1994)); Gingles, 478 U.S. at 88 (O’Connor, J., concurring) (“The phrase vote dilution itself suggests a
norm with respect to which the fact of dilution may be ascertained . . . . In order to decide whether an
electoral system has made it harder for minority voters to elect the candidates they prefer, a court must
have an idea in mind of how hard it should be for minority voters to elect their preferred candidates
under an acceptable system.”) (internal quotation marks omitted). Generally, the plaintiff establishes this
undiluted benchmark practice by proposing hypothetical redistricting schemes and presenting them to
the district court in the form of illustrative plans. Fairley, 584 F.3d at 669.
Where the plaintiff employs the use of illustrative maps to establish the first Gingles factor, the
plaintiff must introduce a demonstration map that creates a geographically compact single-member
district in which the majority of the citizen voting age population is Latino. Valdespino, 168 F.3d at 853.
“Where the minority group meets these requirements, the representatives that it could elect in the
hypothetical district . . . in which it constitutes a majority will serve as the measure of its undiluted
voting strength.” Gingles, 478 U.S. at 90-91 (O’Connor, J., concurring). If the minority group is not
sufficiently large and compact to allow an additional district to be drawn, then the Court has no means
by which to construct a measure of the minority group’s undiluted voting strength. This, of course,
precludes the court from determining whether the challenged practice makes it harder for minority
voters to elect the candidate of their choice. See Gingles, 478 U.S. at 88 (O’Connor, J., concurring).
Accordingly, when the plaintiff fails to produce a measure of the undiluted practice, the plaintiff cannot
show that the challenged electoral practice is actually responsible for the minority voters’ inability to
elect their candidate of choice, defeating any claim of vote dilution. See Gingles, 478 U.S. at 50 n.17
(“Unless minority voters possess the potential to elect representatives in the absence of the challenged
structure or practice, they cannot claim to have been injured by that structure or practice.”); see also
Valdespino, 168 F.3d at 852-53; LULAC IV, 986 F.2d at 743.
45
Plaintiffs have proposed seven Gingles plans which purport to demonstrate the ability of the
Precinct 2’s Latino community to form a precinct where Latinos are geographically compact and
constitute a majority of the citizen-age voting population of the precinct. Of the seven proposed plans,
Plaintiffs’ expert, Mr. George Korbel,20 authored four: Korbel 118, Korbel 257, Korbel 318, and Korbel
325.21
1.
Numerosity
To satisfy the numerosity requirement, Plaintiffs must demonstrate that it is possible to draw an
additional single member election district in which the majority of the citizen voting age population is
Latino. Valdespino, 168 F.3d at 853; Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368, 372 (5th Cir.
1999); Campos, 113 F.3d at 548.
This numerosity requirement relies on an objective, numerical test: Do Latinos make up more
than 50 percent of the citizen voting age population in the relevant geographic area? Valdespino, 168
F.3d at 853.
a.
Measuring the Citizen-Voting Age Population
i.
Source of the Data
Historically, the citizen voting age of the proposed district was calculated as part of the Census.
See, e.g., Valdespino, 168 F.3d at 853-54. Until recently, the Census consisted of a “Short Form,” which
was received by every household in the United States, and a “Long Form,” sent to approximately one in
every six households. The Short Form collected basic information, such as age, sex, race, and Hispanic
20
Plaintiffs offered the testimony of Mr. George Korbel as an expert witness on the Gingles factors and the Senate
factors. Mr. Korbel possesses a B.A. in government and philosophy and a J.D. from the University of Minnesota.
Mr. Korbel has been involved with redistricting for over forty years, dating back to his involvement in the
Supreme Court case of White v. Regester, 412 U.S. 755 (1973). Since 1971, he has been involved in every major
state-wide redistricting case in Texas. See (Tr. 2:1-3). He has also been intimately involved in local redistricting
in Texas, having worked on redistricting for the City of Houston, the State Senate districts, and community
college districts in Texas. (Tr. 2:4). Plaintiffs have retained Mr. Korbel to draw Gingles maps in this case. (Tr.
2:6). Mr. Korbel focuses on non-statistical analysis of elections and the factors that influence their outcomes. His
expertise has been relied upon in countless cases.
21
All of Mr. Korbel’s Gingles plans were drawn in the summer of 2012. (Tr. 2:33).
46
origin, while the Long Form asked more detailed questions on topics such as citizenship and
socioeconomic status. Dr. Barreto explained that although the Long-Form questionnaire was only
distributed to a portion of the population, rendering the long-form a sample rather than an actual
enumeration, the Long-Form data reliably estimated the citizenship data at various levels of geographic
specificity (i.e. statewide, county-wide, by zip code, by census tract, and by census block).22, 23 Recently,
the U.S. Census Bureau (“Census Bureau”) announced that it would not use the Long Form
Questionnaire for the 2010 Census. Benavidez v. Irving Indep. Sch. Dist., 690 F. Supp. 2d 451, 454
(N.D. Tex. 2010).
In its place, the Census has developed the American Community Survey (“ACS”) to estimate the
demographic composition of the United States. The ACS, like the Census Long-Form, is not an actual
population count; instead, the ACS is an annual nationwide survey conducted by the Census Bureau
wherein the Census Bureau randomly samples a large portion of the total population. Those sampled are
asked a series of questions designed to capture demographic information, including socio-economic
background, citizenship, educational attainment, and other information, much of which was formerly
included on the Long Form questionnaire. (Tr. 1:48-49). Thus, the ACS data, like the Long Form
Questionnaire preceding it, provides an estimate of citizen voting age population of a community based
on population samples, rather than an actual enumeration. Benavidez, 690 F. Supp. 2d at 454. According
to Dr. Barreto, the ACS is the only reliable source for citizen voting age population data. (See 1:74).
22
Dr. Barreto explained that census tracts and census blocks are small subdivisions of a county that are generally
coterminous with neighborhoods. Although both census tracts and census blocks only cover a small geography,
census blocks are comprised of blocks of a neighborhood, whereas census tracts correspond to a slightly larger
area. The geographic area bound by the census tracts and census blocks is then aggregated to build political
districts.
23
The boundaries of voting precincts are not always coterminous with census-tract or census-block boundaries.
Voting district boundaries (“VTDs”) are created and drawn by counties and states, whereas census-tract and
census-block boundaries are drawn by the Census. Although states (and their political subdivisions) attempt to
draw VTDs whose boundaries are conterminous with the census-tract or census-block boundaries, often states
will have to create voting districts that divide a census-tract or census-block. (1:54-55).
47
Although ACS data are released annually, the Census Bureau recommends using the three-year
or five-year aggregations of ACS data when working with smaller populations due to relatively small
number of households surveyed. See generally U.S. Census Bureau, A Compass for Understanding and
Using American Community Survey Data: What General Data Users Need to Know (Oct. 2008),
http://www.census.gov/acs/www/guidance_for_data_users/estimates/.
The Fifth Circuit has not decided whether the five-year aggregation of ACS data is always
sufficient to establish the citizenship voting age requirement. However, the Fifth Circuit has authorized
the use of non-census data when that data is sufficiently probative on the issue of citizen voting age
population. See Valdespino, 168 F.3d at 852; Westwego Citizens for Better Gov’t v. City of Westwego,
906 F.2d 1042, 1045 n.3 (5th Cir. 1990) (“Westwego II”). In Westwego II, for instance, the Fifth Circuit
reasoned that the plaintiffs could establish the first Gingles precondition with non-census data, where the
census data available did not include information on the minority voting-age population. Westwego II,
906 F.2d at 1045 n.3.
In this case, the Court finds the ACS data sufficiently probative on the issue of citizen voting age
population. First, as explained earlier, the Census Bureau no longer collects citizen voting-age
population data on the official Census forms. Therefore, the citizenship data from the 2000 Census is the
only official Census citizenship data available. However, this information is out-dated and, given the
surge in population growth Harris County has experienced over the last ten years, therefore inaccurate
and not probative on the question of citizenship status. Second, the ACS data is perhaps the best
measure of citizen voting age data currently available; it is collected by the Census Bureau and the
Census Bureau’s publication of and reliance on ACS data “suggests that the Bureau considers ACS data
reliable and intends for it to be relied upon in decisions such as Voting Rights Act compliance.”
Benavidez v. City of Irving, Tex., 638 F. Supp. 2d 709, 721 (N.D. Tex. 2009). Third, Plaintiffs’ expert,
Dr. Barreto, offered un-rebutted testimony that the five-year aggregation provides a reliable measure of
the citizen voting age population in Harris County because the ACS data proved to be a reliable measure
48
of voting age population. Dr. Barreto testified that he compared the voting age population estimates on
the 2005-2009 five year aggregation of the ACS (the “2005-2009 ACS”) to the voting age population
tallies on the 2010 Census and found that the ACS data accurately approximated the Census tallies.
Because the sampling methodology employed by ACS yielded reliable measures of voting age
population, Dr. Barreto deduced that the ACS data would also yield reliable measures of other
demographic data, such as citizen voting age population. This testimony was not challenged at trial.
Fourth, although Defendants’ expert, Dr. Alford, challenges the method by which Dr. Barreto has used
the ACS data to extrapolate information about the wider population, see (Defendants’ Ex. 49, at 2-4),
both parties have used the ACS citizenship data to approximate the citizen voting age population in
Harris County. On this record, the Court concludes that the five-year aggregated ACS citizenship data is
sufficiently probative on the issue of citizen voting age population and Plaintiffs may rely upon this data
in establishing the first Gingles precondition.
Dr. Barreto, however, states that the five-year ACS data does not always yield adequate
approximations of the number of Latino citizens, because the data is a composite of several years rather
than a snap shot of any single year. For instance, Dr. Barreto explained that in the last several years there
has been a sizeable increase in Harris County’s Latino citizen voting age population. Given the
accelerated rate at which the Latino citizen voting age population is growing, Dr. Barreto contends that
the ACS’ five-year aggregations include historic data that significantly underestimates Harris County’s
Latino citizen voting age population. Given the dynamism of population trends in Harris County, Dr.
Barreto suggests that the five year aggregates capture the population trends that occurred in the midpoint of the data set. Thus, for example, Dr. Barreto contends that the 2005-2009 ACS data estimates is
a reflection of the Latino citizen voting age population in 2007, the mid-point of the 2005-2009 data set,
but does not accurately approximate the Latino citizen voting age in 2009, the end point of the data set.
49
In fact, Dr. Barreto hypothesizes that the 2005-2009 ACS data underestimates the Latino citizen voting
age population in 2009 by one to two percent percentage points. (Tr. 1:57).24
Although this testimony was un-rebutted at trial, the Court’s reliance on the ACS data will not
be disturbed by the fact that this data is static and therefore fails to fully capture the dynamism of Harris
County’s population changes. The ACS is at present the only reliable source of citizen-voting age
population data. And although the five-year arc of the ACS might diminish its ability to be perfectly
reflective of current trends, this five-year aggregation enables the survey to, as Plaintiffs explained,
reliably approximate Census tallies. The Court has determined that the ACS data can be used to
demonstrate citizenship voting age population data because of ACS’ ability to consistently approximate
Census data. While Plaintiffs may be correct that the ACS data is not entirely accurate, it is the best
alternative the parties have proposed, and Plaintiffs have not offered “concrete evidence” of the undercount sufficient to overcome the presumption of accuracy of the ACS. See Perez, 958 F. Supp. at 1211
(noting that claims of Latino undercount in census had to be substantiated by concrete evidence before
the court would entertain the possibility that the census data was inaccurate.) Inaccuracies, alone, will
not render the ACS data inherently unreliable. See Campos v. City of Houston, 894 F. Supp. 1062, 1064
(internal citations omitted) (noting that although inaccuracies are inherent in census data, census data
was not inherently unreliable). Accordingly, the Court finds that the ACS data is the only reliable indicia
of citizen voting age population in the record.
ii.
How Should the ACS Be Used
Having determined that the five-year aggregation of citizenship data provided by the ACS is a
reliable indicator of the citizen voting-age population, the Court must now determine how the five-year
aggregation should be considered. Again, the ACS produces estimates, rather than actual counts, of
citizen voting age population. The estimates are denoted with a numeric estimate and then an
24
In preparing his expert report, Dr. Barreto used the 2005-2009 ACS data, because that was the most recent data
then available. Since that time, the ACS has also released five year aggregates of the 2006-2010 data and 20072011 data. Dr. Barreto’s results, however, are based only on the 2005-2009 ACS data.
50
accompanying margin of error. For example, a five-year ACS survey may estimate that the AfricanAmerican population of a city is 1,000,000 (±15,000). Accordingly, the ACS provides an estimate of
1,000,000 with a margin of error of 15,000 persons.
Plaintiffs’ expert, Dr. Barreto, has used the ACS numeric estimate for citizen voting age
population, without the accompanying margin of error, to approximate the actual count of the citizen
voting age population of Harris County. Defendants’ expert, Dr. John Alford,25 questions Dr. Barreto’s
use of the ACS numeric estimate. According to Dr. Alford, it is statistically improper for Dr. Barreto to
use the ACS numeric estimate of citizen voting age population to approximate the raw number of Latino
adult citizens residing in Harris County. According to Dr. Alford, the ACS guidelines and the census
handbook suggest that demographers and statisticians should not be making extrapolations to a raw
number from the ACS data. (Defendants’ Ex. 49, at 2-4).
Dr. Barreto testified, however, that the ACS data can be used to approximate raw number
estimates. Thus, for instance, the ACS’ citizen voting age population data, for Harris County, could be
used to estimate the actual number of Latino citizens of voting age in Harris County. According to Dr.
Barreto, the ACS data lends itself to this kind of extrapolation because the ACS data is collected from a
large sample of the population. Although Dr. Barreto concedes that the ACS cautions against
manipulating the data to produce raw number estimates, Dr. Barreto states that this warning is directed
to a lay user, who lacks familiarity with proper extrapolation techniques and may be tempted to use ACS
25
Dr. John Alford holds a Bachelor of Science in political science, a master’s degree in public administration, a
master’s degrees in political science and a Ph. D. in political science. (Tr. 4:44). Dr. Alford is a tenured associate
professor of political science at Rice University. (Tr. 4:44).25 He has been affiliated with Rice University for the
twenty-six years. (Tr. 4:44). He teaches classes on American politics and elections behavior. (Tr. 4:45). Outside
the classroom, he has assisted jurisdictions in drawing single-member plans and helping them secure preclearance for maps; he has also served as a Voting Rights expert and statistical expert in Section 5 and Section 2
litigation regarding redistricting. (Tr. 4:45). He has done significant voting rights work in Harris County, having
worked for Harris County Community College, City of Baytown, Pasadena Independent School District, and
other political subdivisions in the surrounding counties. (Tr. 4:45). Dr. Alford also testified that he has experience
as a statistician, but admitted that most of his statistical analysis has been related to political behavior and has not
been related to voting rights. (Tr. 4:46). Nevertheless, the defendants offered him as an expert in redistricting and
statistical analysis, without objection from Plaintiffs. (Tr. 4:46-47).
51
data to extrapolate population-wide data even when the sample from which the ACS data was collected
is small. As Dr. Barreto explained, when the sample from which the ACS data is collected is small—as
occurs when the ACS demographer is only able to survey a small number of people in the community—
then the sample data will not yield accurate approximations of the population-wide demographics. Thus,
for example, it would be inappropriate to use the ACS citizenship data from Polk County, Texas to
extrapolate the number of Latino citizens in Polk County, because for small counties, like Polk, the
number of people from whom the ACS collects data is very small. Accordingly, the data collected from
this small sample cannot be used to determine population-wide (i.e. county-wide) trends. However, in
large counties, such as Harris County, the ACS collects data from a large segment of the population. In
these counties, the ACS staff conducts many interviews, thereby increasing the reliability of the ACS
data. In these large counties, the sample size from which the ACS data is collected is large enough that
the sample data can produce reliable raw number estimates. Therefore, in large counties, ACS data
describing the percentage of Latino citizens in the county could be used to determine the number of
Latino citizens in that county. See (Tr. 1:135-36).
According to Dr. Barreto, demographers and others versed in surveying techniques can readily
discern the difference between sample data and geography levels that lend themselves to extrapolation
and sample data or geographies that are too small to yield reliable approximations of population-wide
statistics. Moreover, the sample size used in compiling the ACS data for Harris County is large enough
that the ACS data can yield very accurate approximations of population-wide data. Dr. Barreto thus
concludes that it is not improper to use ACS sample data from Harris County to extrapolate raw number
estimates. In fact, Dr. Barreto testified that he is confident that ACS citizenship data could yield very
accurate estimates of the county-wide citizenship numbers, because the ACS sample data provide
reliable estimates of other population-wide demographics. Upon comparing the voting age population,
racial and ethnic group counts, and total population estimates for Harris County from the 2005-2009
ACS with the voting age population counts, racial and ethnic group counts, and total population data
52
from the 2010 Census, Dr. Barreto found that the ACS data accurately approximated the Census tallies.
Thus, for instance, in Harris County the ACS will publish an estimate of the number of Latino adult
citizens residing in Harris County and alongside this number the ACS will also publish a margin of error
for the estimate. According to Dr. Barreto, in counties as populous as Harris County, the margin of error
for the estimate is small. Thus, for example, the ACS might estimate that there are 1,000,000 Latino
voting age adults and the margin of error for that estimate may only be +/- 10,000. (Tr. 1:136-37). When
those ACS estimates for Harris County are compared to actual Census data, the ACS data is incredibly
accurate. In 2010, the ACS estimated that there were 1,092,000 Latino adults in Harris County and the
Census estimated that there were 1,082,000 Latino adults in Harris County. This demonstrates that the
Latino voting age count extrapolated from the ACS sample accurately approximates the Latino voting
age actual enumeration collected by the Census. (Tr. 1:137).
Because the sampling methodology employed by ACS yielded reliable measures of county-wide
voting age population, county-wide racial and ethnic tallies, and the county-wide total population, as
reflected in the Census tallies, Dr. Barreto reasoned that the ACS data would also yield reliable
measures of other demographic data, such as citizen voting age population. (Tr 1:138-39). According to
Dr. Barreto, the ACS is using the same principles to collect and organize data for voting age population
as they would for other demographic markers. Dr. Barreto concludes that in large counties like Harris
County, the ACS data is extremely accurate, because the sample size from which the data is collected is
large and the point estimates supplied by the ACS closely approximate the actual enumerations gathered
in the decennial census. (Tr. 1:139).26
26
For smaller geographies, like a census tract or a census block, the ACS does not provide single-year estimates
of demographic information because the survey will only sample a very small number of people within the census
tract or block for any one year. So, for example, the ACS will not estimate the percentage of voting age citizens
living in a particular census tract in 2010. To address the problems created by the limited sample size in smaller
geographies, the ACS aggregates five consecutive years’ worth of sampling data. This creates a larger sample size
from which the ACS can produce reliable census tract or census block-level demographic information. Although
the sample size created by this aggregation is larger than the sample size from any one year, the aggregated
sample is still relatively small. As such, the raw counts extrapolated from census tract (or block) level data are
53
On review, the Court finds that trained professionals may use the numeric estimates to
extrapolate raw number counts. The concerns pertaining to extrapolation raised by Dr. Alford, though
serious, do not apply to the case at hand. The weight of the testimony at trial indicated that the sample
size from which the ACS draws its data is large enough that the sample data reliably estimates
population-wide trends. As Dr. Barreto explained, the margin of error of the ACS’ citizen voting age
population statistics is small. The evidence further shows that the admonishments of the ACS relating to
the use of its estimates—admonishments echoed by Dr. Alford—are not really directed at experienced
demographers and academics, who have the ability to determine the appropriateness of any particular
data manipulation techniques. Accordingly, the Court finds that the numeric estimate may be used to
approximate the raw count totals.
iii.
Which Five Year Aggregation Should Be Used
Having determined that the five-year aggregation of citizenship data provided by the ACS is a
reliable indicator of the citizen voting-age population, the Court must now determine which five-year
aggregation should be considered. The bulk of Plaintiffs’ evidence challenging the Revised A-1 relies on
the 2005-2009 ACS. Plaintiffs testified, and Defendants have not rebutted, that this 2005-2009 ACS data
was the only ACS data publicly available at the time the Revised Plan A-1 was adopted in August 2011.
After the adoption of the Revised Plan A-1, but before the commencement of trial, the 2006-2010 ACS
was released to the public. There is some tension amongst the parties as to which set of ACS data should
be considered. See (Tr. 2:64) (At trial Plaintiffs commented that it is “not clear, when a court considers a
less accurate, i.e. will have a higher margin of error or a wider confidence band, than raw counts extrapolated
from county-wide data, where the sample is larger.
The estimation error associated with these smaller geographies can have a meaningful impact on the
voting rights inquiry. According to Dr. Barreto, when electioneers or social scientist build election maps, they
compile a number of census tracts and census blocks into voting districts and precincts. To determine the
demographic information, such as the number of citizens, for these voting precincts, the social scientists simply
aggregate the citizenship information of all of the individual census tracts and blocks within the precinct. Thus,
the citizenship data for a voting precinct is simply the sum of all the citizenship totals of the census blocks and
tracts within the voting precinct. Therefore, the estimation errors that occur when citizenship totals are
extrapolated from the census block (or tract) level will also be present in the voting precinct level data.
54
plan should it be looking at the ACS data office holders had at the time they drafted the map or can they
look at newer data?”).
In light of the changing demographics of the population and the specialized need for fact
findings that pay special heed to the local history and context, courts require Plaintiffs to provide
reliable proof of a Section 2 violation, including reliable proof that the minority group is sufficiently
geographically compact and numerous to constitute a majority in a single member district. See Perez,
958 F. Supp. at 1210. The degree to which data may be relied upon to support a vote dilution is always
dependent on the reliability of the data itself. See Westwego I, 906 F.2d at 1045-46 n.3; Perez, 958 F.
Supp. at 1210-14. Historically, our courts have allowed Plaintiffs to mount a Section 2 challenge using
dated census data that was not entirely reflective of then-existing population trends. See e.g. Perez, 958
F. Supp. at 1213. Campos v. City of Houston, 894 F. Supp. 1065 (S. D. Tex. 1995). Dr. Barreto’s
testimony suggests, as common sense would predict, that the further away in time one moves from the
terminal year of the ACS aggregation period, the less reliable that data becomes. However, the mere fact
that the static data, such as the 2005-2009 ACS, is, as a general proposition, less reliable as time
progresses does not address whether the use of that data today is proper. But our past practice with the
census is instructive. Historically, census data was presumed reliable unless there was evidence that the
population dynamics had changed so dramatically and the census data was so outdated that it could no
longer be deemed reflective of the population. See Perez, 958 F. Supp. at 1210. “Proof of changed
figures must be thoroughly documented, have a high degree of accuracy, and be clear, cogent, and
convincing.” See Perez, 958 F. Supp. at 1210. Under this formulation, the publication of subsequent
ACS data could be proof of changed figures that would allow the Court to consider the later-published
ACS data.27 The Court thus finds that Plaintiffs may support their case with 2006-2010 ACS figures.
27
Obviously if the five-year aggregation of data is considered a reliable indicator of the population, then a later
amended five-year aggregation must also be a reliable indicator. If, as the Court has already ruled, five-year ACS
aggregations can be used to prove a Section 2 claim, then later published ACS data would, necessarily, be
55
Although Plaintiffs may support their case with the 2006-2010 ACS data, they are not bound to do such.
Plaintiffs may also elect to support their case with the 2005-2009 ACS, as long as there is no proof that
the older and inherently less reliable 2005-09 data has become so outdated that it is no longer probative
on questions of demography. In this case, there has been no such showing. Accordingly, the Court finds
that Plaintiffs may rely on either the 2005-09 ACS data or the 2006-2010 data.28
b.
Numerosity: Do Plaintiffs’ Gingles Maps Demonstrate a Latino
Majority?
Plaintiffs offered into evidence seven demonstration Gingles maps. Only three maps were
discussed at trial: Korbel 257, Korbel 325, and Map 3. The other four maps—Map 1, Map 2, Korbel
118, and Korbel 318—were merely included in the admitted exhibits. However, at the close of trial,
counsel for Plaintiffs indicated that he wished to have the Court consider all of Plaintiffs’ demonstration
plans.
i
Analysis of the Maps Discussed at Trial
Korbel 257
The first plan offered at trial was Korbel 257. The citizen-voting age population data associated
with this map is depicted in the next chart.
CITIZEN-VOTING AGE POPULATION DATA FOR PLAINTIFFS’ KORBEL 257 GINGLES
MAP29
2010 Census Data
2005-2009 ACS Data
2006-2010 ACS Data
Pre.
Total
VAP
CVAP
% Latino
CVAP
% Latino
CVAP
CVAP
1
1,045,108 778,692 574,695 (±7,510)
16.9%
562,800 (±6,715)
17.7%
2
952,194
645,124 410,950 (±6,017)
52 %
412,570 (±5,904)
53.8%
3
1,047,714 773,766 605,935 (±6,104)
14.3%
621,710 (±5,978)
15.4%
4
1,047,443 747,042 603,880 (±6,323)
16.1%
633,415 (±6,366)
17.9%
sufficiently accurate, clear, cogent, and convincing to prove an intervening change between an earlier five-year
aggregation and a later five-year aggregation.
28
As a practical matter, the cost of litigation would increase significantly if plaintiffs were expected to produce
the most recent ACS data to challenge the defendant’s map, when, as was the case here, the most recent ACS data
when Defendants’ maps were adopted differs from the ACS data published after adoption, but before trial.
29
Source: Plaintiffs’ Ex. 46. at 4.
56
The 2005-2009 ACS data shows that, under Korbel 257, 52% of the citizen voting age population in
Precinct 2 is Latino. See (Plaintiffs’ Ex. 46, at 5); (Tr. 2:49). The 2006-2010 ACS data, however, shows
that 53.8% of the citizen voting age population in Precinct 2 is Latino, under Korbel 257. See (Plaintiffs’
Ex. 46, at 5); (Tr. 2:55). Under both sets of data, the majority of Precinct 2’s citizen voting age
population is Latino. Thus, Plaintiffs’ Korbel 257 map satisfies the numerosity component of the first
Gingles factor.
Korbel 325
The second map offered at trial was Korbel 325. The citizen-voting age population data
associated with Korbel 325 is depicted in the next chart.
CITIZEN-VOTING AGE POPULATION DATA FOR PLAINTIFFS’ KORBEL 325 GINGLES
MAP30
2010 Census Data
2005-2009 ACS Data
2006-2010 ACS Data
Pre.
Total
VAP
CVAP
% Latino
CVAP
% Latino
CVAP
CVAP
1
996,174
737,277 542,400 (±7,384)
17.0%
529,000 (±6,566)
17.8%
2
994,188
674,439 437,800 (±6,001)
50.1%
443,960 (±6,010)
51.6%
3
1, 051,668 771,108 603,030 (±6, 085)
13.0%
627,375 (±5,882)
14.4%
4
1, 050,429 761,800 612,230 (±6,502)
17.1%
630,160 (±6,510)
18.5%
The 2005-2009 ACS data shows that, under Korbel 325, 50.1% of the citizen voting age
population in Precinct 2 is Latino. See (Plaintiffs’ Ex. 57, at 8). However, the 2006-2010 ACS data
shows that 51.6% of the citizen voting age population in Precinct 2 is Latino under Korbel 325. See
(Plaintiffs’ Ex. 57, at 6); (Tr. 2:62). Under both sets of data, the majority of Precinct 2’s citizen voting
age population is Latino. Thus, Plaintiffs’ Korbel 325 map satisfies the numerosity component of the
first Gingles factor.
30
Source: Plaintiffs’ Ex. 57, at 8 (chart labeled SBOE Districts-GLAME 325/TLC2E179). On the day this
evidence was presented at trial, Mr. Korbel explained that he did not include the 2005-09 ACS data in his exhibit.
(Tr. 2:63). However, the next day, Plaintiffs supplemented the exhibit to include the data from the 2005-09 ACS,
as well. See (Tr. 3:1).
57
Map 3
The third plan offered at trial was Map 3. Mr. Korbel did not author this plan. See (Tr. 2:68); see
also (Plaintiffs’ Ex. 3, at 1). According to Mr. Korbel, this plan was drawn by some community
members in Houston at the time the County Commissioner’s map was being debated, in August 2011.
See (Tr. 2:69). The citizen voting age population data accompanying this map is reproduced in the next
chart.
CITIZEN-VOTING AGE POPULATION DATA FOR PLAINTIFFS’ MAP 3 GINGLES
MAP31
District
Total
VAP
CVAP
% Latino CVAP
1
1,026,179
749,368
549,705 (±7,545)
16.5%
2
960,495
655,752
415,010 (±5,887)
50.9%
3
1,054,727
775,283
542,790(±6,169)
16.6%
4
1,050,959
764,148
612,055 (±5,876)
13.6%
Plaintiffs have not offered any evidence (either testimonial or documentary) as to the source of
this data; it is not clear if this data was supplied by the ACS or whether it was provided by some other
source. Plaintiffs have not provided any means by which the Court could assess the reliability of the data
offered and therefore the Court finds that this evidence is not probative on the citizen voting age
population data for Plaintiffs’ third Gingles map. Given that this is the only citizen voting age data
Plaintiffs have submitted in connection with Plaintiffs’ third Gingles map, the Court finds that Plaintiffs
have not submitted reliable evidence of the citizen voting age population in Precinct 2 under Map 3.
Accordingly, Plaintiffs have failed to demonstrate that Latinos constitute a majority of the citizen voting
age population of Precinct 2 under the third Gingles map and thus have failed to establish the
numerosity requirement of the first Gingles factor. Accordingly, the Court finds that the Map 3 Gingles
plan cannot give rise to a claim of vote dilution.
31
Source: Plaintiffs’ Ex. 3, at 2.
58
ii.
Analysis of the Other Maps
In addition to the plans discussed at trial, Plaintiffs also submitted four additional plans into
evidence: Map 1, Map 2, Korbel 118, and Korbel 318.
Map 1
Attached to Plaintiffs’ Map 1 was a chart detailing the 2005-2009 ACS citizen voting age
population data for each of the commissioner’s precincts under Plaintiffs’ Map 1. That data is
reproduced in the chart that follows.
2005- 2009 CITIZEN-VOTING AGE POPULATION DATA FOR PLAINTIFF’S MAP 1
GINGLES MAP32
District
Total
VAP
CVAP
% Latino CVAP
1
1,023,902
745,328
584,490 (±7,663)
17.7%
2
1,017,706
708,120
506,600 (±6,389)
42.6%
3
1,030,421
747,631
515,190 (±6,118)
17.9%
4
1,020,430
743,545
589,180 (±5,733)
14.2%
Under Map 1, 42.6% of Precinct 2’s citizen voting age population is Latino; Latinos do not
constitute a majority of the citizen-voting age population of the precinct. Therefore, Plaintiffs have
failed to establish that Map 1 satisfies the numerosity requirement of the first Gingles factor.
Valdespino, 168 F.3d at 853; Campos, 113 F.3d at 548. Accordingly, the Court finds that the Map 1
Gingles plan cannot give rise to a claim of vote dilution.
Map 2
Attached to Plaintiffs’ Map 2 was a chart detailing the 2005-2009 ACS citizen voting age
population data for each of the commissioner’s precincts under Plaintiffs’ Map 2. That data is
reproduced in the chart that follows.
32
Source: Plaintiffs’ Ex. 1, at 6.
59
2005-2009 ACS CITIZEN-VOTING AGE POPULATION DATA FOR MAP 233
District
Total
VAP
CVAP
% Latino CVAP
1
1,019,643
741,936
583,290 (±7,658)
17.7%
2
1,017,706
708,120
506,600 (±6,389)
42.6%
3
1,030,365
751,044
523,970 (±6,171)
17.5%
4
1,024,745
743,524
581,600 (±5,682)
14.4%
Under Map 2, 42.6% of Precinct 2’s citizen voting age population is Latino; Latinos do not
constitute a majority of the citizen-voting age population of the precinct. Therefore, Plaintiffs have
failed to establish that Map 2 satisfies the numerosity requirement of the first Gingles factor.
Valdespino, 168 F.3d at 853; Campos, 113 F.3d at 548. Accordingly, the Court finds that the Map 2
Gingles plan cannot give rise to a claim of vote dilution.
Korbel 118
The next map is Korbel 118. The citizen-voting age population data associated with Korbel 118
is depicted in the next chart.
CITIZEN-VOTING AGE POPULATION DATA FOR PLAINTIFF’S KORBEL 118 GINGLES
MAP34
2010 Census Data
2005-2009 ACS Data
2006-2010 ACS Data
Pre.
Total
VAP
CVAP
% Latino
CVAP
% Latino
CVAP
CVAP
1
1,039,016
772,266 573,465 (±7,726)
16.5%
561,270 (±6,861)
17.6%
2
967,592
651,959 414,800 (±5,870)
52.3%
420,485 (±5,893)
53.7%
3
1,044,999
771,077 600,850 (±6,273)
15.0%
614,385 (±5,995)
16.0%
4
1,040,852
749,322 606,345 (±6,028)
15.3%
634,355 (±6,203)
17.0%
The 2005-2009 ACS data shows that, under Korbel 118, 52.3% of the citizen voting age
population in Precinct 2 is Latino. See (Plaintiffs’ Ex. 45, at 5). The 2006-2010 ACS data shows that
53.7% of the citizen voting age population in Precinct 2 is Latino, under Korbel 118. See (Plaintiffs’ Ex.
45, at 5). Under both sets of data, the majority of Precinct 2’s citizen voting age population is Latino.
Thus, Plaintiffs’ Korbel 118 map satisfies the numerosity component of the first Gingles factor.
33
34
Source: Plaintiffs’ Ex. 2, at 6.
Source: Plaintiffs’ Ex. 45, at 5.
60
Korbel 318
The last map is Korbel 318. This exhibit does not include any citizen-voting age population data,
but rather includes a chart detailing the voting age population data. The voting age data associated with
Korbel 318 is depicted in the next chart.
District
1
2
3
4
VOTING AGE POPULATION DATA FOR KORBEL 31835
Total
VAP
Latino VAP
% Latino VAP
1,031,414
767,310
259,887
37.7%
984,225
668,428
461,140
69.0%
1,037,940
767,424
175,229
2208%
1,038,880
741,462
186,314
14.4%
As a threshold matter, the Court notes that Plaintiffs’ evidence does not conclusively establish
that Latinos in Precinct 2 constitute a majority of the citizen voting age population under Korbel 318,
because the data submitted only analyzes voting age population data, rather than citizen voting age
population data. That said, this map has a Spanish Surname registration rate of 47.1% and Latinos
constitute 69% of the voting age population. This latter point is noteworthy because in Korbel 325,
Latinos only constituted 68.2% of the voting age population of Precinct 2, but still constituted a majority
of the citizen voting age population of Precinct 2. Given that, the fact that 69% of the voting age
population is Latino, here, lends credence to the inference that Latinos constitute a majority of the
citizen voting age population under Korbel 118. Nevertheless, our case law requires plaintiffs to
establish each Gingles precondition and these inferences are insufficient to satisfy that burden,
especially given the fact that the use of SSRV data is often a crude and imprecise measure of the Latino
registered voter population, that tends “to misidentify Hispanic persons as non-Hispanic and viceversa…[Therefore], without a strict showing of its probativeness, Spanish-surname data are disfavored”
in the Section 2 analysis. Rodriguez v. Bexar Cnty., 385 F.3d 853, 867 n.18 (5th Cir. 2004). Accordingly,
the Court finds that Korbel 318 cannot give rise to a claim of vote dilution. See Roscoe, 123 F.3d at 846
(noting that Section 2 plaintiffs must “prove by a preponderance of the evidence that all of the Gingles
35
Source: Plaintiffs’ Ex. 47, at 2.
61
preconditions were satisfied...[,] therefore [a]ny lack of evidence in the record regarding a violation of
the Voting Rights Act of 1965 must be attributed to [the plaintiffs]”).
iii.
Summary
In summary, Plaintiffs submitted into evidence seven demonstration maps: (i) Korbel 257; (ii)
Korbel 325, (iii) Map 3, (iv) Map 1, (v) Map 2, (vi) Korbel 118, and (vii) Korbel 318. Of these maps, the
Court finds that Plaintiffs have not demonstrated that Latinos constitute a majority of Precinct 2’s citizen
voting age population under Map 3, Map 1, Map 2 and Korbel 318.
Although Plaintiffs submitted data in Map 3 showing that the Latinos constituted a majority of
Precinct 2’s citizen voting age population, Plaintiffs did not provide any means for the Court to ascertain
the reliability of this data. The source of the data is unspecified and therefore the Court cannot attest to
its reliability. Accordingly, the Court finds that Plaintiffs failed to demonstrate that Latinos constituted a
majority of Precinct 2’s citizen voting age population in Map 3, using reliable data. As for Map 1 and
Map 2, the data showed that Latinos constituted only a plurality, rather than a majority, of Precinct 2’s
citizen voting age population. Finally, as for Korbel 318, Plaintiffs failed to offer any statistical evidence
on the citizen voting age population. Although Plaintiffs offered crude proxies for citizen voting age
population, they failed to submit any evidence directly on point.
The Court did, however, find that Latinos constitute a majority of Precinct 2’s citizen voting age
population under Korbel 257, Korbel 325, and Korbel 118. Accordingly, these three maps are the only
bases for Plaintiffs’ claim of vote dilution.
2.
Compactness
To satisfy the compactness requirement, the plaintiff must show that the minority population is
geographically compact. Abrams v. Johnson, 521 U.S. 74, 91-92 (1997); accord Sensley v. Albritton,
385 F.3d 591, 596 (5th Cir. 2004). The compactness inquiry focuses on the “compactness of the
minority population, not [] the compactness of the contested district.” LULAC, 548 U.S. at 433. Stated
62
otherwise, the compactness inquiry focuses on the population dispersal of the minority community, not
on the geographic density of the proposed district. See Sensley 385 F.3d at 596.
The compactness requirement is necessary to show that the challenged electoral practice, rather
than the dispersion of the minority community, prevents the affected minority group from electing the
candidates of their choice. See Bush v. Vera, 517 U.S. 952, 979 (1996) (plurality) (if, “because of the
dispersion of the minority population, a reasonably compact majority-minority district cannot be created,
§ 2 does not require a majority-minority district.”); see also LULAC, 548 U.S. at 433, 435 (“[T]here is
no basis to believe a district that combines two farflung segments of a racial group with disparate
interests provides the opportunity the Section 2 requires . . . . The mathematical possibility of a racial
bloc does not make a district compact.”). A district is sufficiently compact if it allows for effective
representation. A district would not be sufficiently compact if it was so convoluted that there was no
sense of community, that is, if its members and its representative could not easily tell who actually lived
in the district. Clark v. Calhoun County, Miss., 21 F.3d 92, 96 (5th Cir. 1994) (Clark I); Dillard v.
Baldwin County, 686 F. Supp. 1459, 1466 (M.D. Ala. 1988).
While no precise rule has emerged delimiting the scope of the compactness inquiry, the court
must consider the shape of the plaintiff’s proposed map, Sensley, 385 F.3d at 596, and must also
consider the degree to which the plaintiff’s proposed demonstration maps comport with “traditional
districting principles such as maintaining communities of interest and traditional boundaries.” Sensley,
385 F.3d at 596; see also LULAC, 548 U.S. at 433 (when conducting the compactness analysis, a district
court must consider traditional districting principles); Fairley, 584 F.3d at 670 (in evaluating
redistricting plans, courts are expected to “take into account traditional districting principles such as
maintaining communities of interest and traditional boundaries”).
In this case, three maps, i.e. Korbel 257, Korbel 325, and Korbel 118, satisfied the numerosity
requirement. Therefore, the Court ordinarily would consider whether these three maps satisfied the
compactness requirement as well. However, at trial, Plaintiffs only presented compactness evidence
63
pertaining to Korbel 257 and Korbel 325. Plaintiffs did not present any compactness evidence on Korbel
118. To be sure, Plaintiffs have submitted evidence in support of Korbel 118, but none of the submitted
evidence explains or permits the Court to understand to what extent Korbel 118 map comports with
traditional redistricting principles. Without “sufficiently detailed evidence, a court is flatly unable to
evaluate whether a possible redistricting scheme would establish legally adequate districts consistent
with traditional districting principles such as compactness, contiguity, maintaining communities of
interest, and respect for incumbency.” Fairley, 584 F.3d at 669 (internal citations omitted). Plaintiffs,
here, never explained how this map comports with those traditional redistricting criteria, leaving this
Court “unable to determine whether [that compliance] would be met.” Fairley, 584 F.3d at 670. Because
Plaintiffs bear the burden of establishing all three Gingles preconditions, “[a]ny lack of evidence in the
record regarding a violation of the Voting Rights Act of 1965 must be attributed to [the plaintiffs].”
Roscoe, 123 F.3d at 846; accord Fairley, 584 F. 3d at 669. Moreover, because Plaintiffs have failed to
establish that Korbel 118 comports with traditional redistricting principles, Plaintiffs have failed to show
that Korbel 118 meets the first Gingles precondition. Accordingly, Korbel 118 cannot give rise to a
claim of vote dilution and the Court need not further consider this map.
a.
Shape
As for the shape assessment, the plaintiff’s burden is not onerous; this analysis does not require
“some aesthetic ideal of compactness,” but rather considers whether the minority population is
sufficiently compact to constitute a majority in a single-member district. Clark I, 21 F.3d at 95. By
compactness, Gingles does not mean that a proposed district must meet, or attempt to achieve, some
aesthetic absolute, such as symmetry or attractiveness. Geographical symmetry or attractiveness is a
desirable consideration for districting, but only to the extent it facilitates the political process. Dillard,
686 F. Supp. at 1466.
64
i.
Measuring the Shape: Geometric Test
There are many methods by which to assess the shape of the district. One recognized method is
simply to examine the physical boundaries of the maps and the proposed districts and, based on that
visual examination, determine if the district is strangely shaped. Sensley, 385 F.3d at 596. This
geometric test, though imprecise, offers some threshold assessment of compactness. The Court now
turns to the shape of the two remaining demonstration plans, Korbel 257 and Korbel 325.
Mr. Korbel testified that the black population and Latino population in Harris County sit like an
“X” and they pass each other right in downtown Houston. In order to fashion districts that preserve the
voting strength of one community, without diluting the voting strength of the other community,
legislatures or courts have traditionally drawn the Latino senatorial and congressional districts, which
occupy the central-southeast portion of the county, such that it snakes around the black population in
central-southwest Harris County. (Tr. 2:29). According to Mr. Korbel, the U.S. Congressional and Texas
state senatorial districts that overlap with Harris County Commissioner’s Precincts 1 and 2 have
“irregular shapes . . . that are being drawn to fit the minority population.” (Tr. 2:26). For instance, Mr.
Korbel examined the structure of Senate District 6, in effect from 2001 through 2010, and noted that
Senate District 6, a Latino opportunity district, dances through the central and southwest Harris County
going around the Senate District 13, the adjacent African-American opportunity district. See (Plaintiffs’
Ex. 50, at 4).
Mr. Korbel further explained that Korbel 257 and Korbel 325’s Precinct 2 was modeled after
Senate district 6; in both Korbel 257 and Korbel 325, Precinct 2 wraps around the two arms of Precinct
1, in exactly the same pattern as the congressional and senate plans. (Tr. 2:80, 83). Accordingly, Mr.
Korbel testified that this plan looks like the congressional and Senate plans, in that the black population
lives in two arms of the district that protrude from downtown and go east and westward, and then the
Latino opportunity district wraps around the black district. (Tr. 2:80).
65
Defendants claimed that, under Korbel 257 and Korbel 325, Precinct 2 extends as far east as
Baytown and its westernmost point ends in Spring Branch, thus Precinct 2 spans a distance of almost
forty miles east-to-west (Tr. 2:81). At first glance this seems alarming, however all the maps offered,
both Plaintiffs’ maps and Defendants’ Revised A-1 map, include precincts that traverse great expanses
of the County. This, of course, is unsurprising given the size of Harris County. On review, the Court
does not find Plaintiffs’ proposed maps particularly unusual; they are no more unusual than Defendants’
Revised Plan A-1. See Jeffers v. Clinton, 730 F. Supp. 196, 207 (E.D. Ark. 1989) (three-judge court)
(“[Plaintiffs’] alternative districts are not materially stranger in shape than at least some of the districts
contained in the present apportionment plan.”), aff’d mem., 498 U.S. 1019 (1991); Marylanders for Fair
Representation v. Schaefer, 849 F. Supp. 1022, 1053-54 (D. Md. 1994). Although, the shape of the
precincts contained in Korbel 257 and 325 are not Euclidean, the shapes are not so bizarre or irregular as
to render them objectionable. Visually the maps appear to incorporate natural geographic boundaries and
the maps replicate the historic patterns of the predominately Latino district sitting atop the
predominately African-American opportunity district in the center of the County. The Court thus finds
that Plaintiffs’ Korbel 257 and Korbel 325 are compactly shaped.
ii.
Measuring Shape: Statistical Measures of Compactness
Although the geometric assessment is valuable, its imprecision limits its utility. Knowing that,
Section 2 litigants whenever possible should try to “adduce evidence systematically comparing proposed
redistricting plans to either the plan being challenged or the predecessor plan in effect during the last
relevant election.” Schaefer, 849 F. Supp. at 1053. “Social scientists have developed a plethora of
methodologies for objectively measuring the compactness of individual districts and of entire plans.”
Schaefer, 849 F. Supp. at 1053; see also (Tr. 2:91).
Mr. Korbel testified that the compactness of his proposed Gingles map could be measured, in
part, by reference to the compactness metrics employed by the Texas Legislative Council. At trial, Mr.
Korbel testified that the Texas Legislative Council offers many different methods and statistics by which
66
to assess the compactness of a proposed map. Mr. Korbel has decided to evaluate the compactness of his
Gingles maps using three such compactness metrics: area rubber band, perimeter-to-area, and
population-rubber band. See (Plaintiffs’ Ex. 48, at 3); (Tr. 2:24). Each metric is measured on a scale
from 0 to 1, with 0 indicating a non-compact area and a score of 1 indicating a perfectly compact area.
(Tr. 2:94).
The area rubber band metric measures the “ratio of the area of the district to the area of the
smallest convex polygon enclosing the district,” (Plaintiffs’ Ex. 62, at 1 n.1), and symbolizes “how
tightly packed or spread out the geography of a district is.” Richard H. Pildes & Richard G. Niemi,
Expressive Harms, “Bizarre Districts,” and Voting Rights: Evaluating Election-District Appearances
After Shaw v. Reno, 92 Mich. L. Rev. 483, 554 (1993). Operationally, it involves taking the areas of the
district and of the smallest circle that completely encloses the district. The ratio of the former to the
latter yields the dispersion compactness score. Hence, a circular district is perfectly compact. A square
district is relatively compact because, when one draws a circle around the district, there is little area
inside the circle that is not also in the district. A long, narrow district, or one with “fingers” or other
extensions, is less compact because it takes a large circle to enclose the entire district, yet much of that
circle is empty.
Mr. Korbel testified that the perimeter-to-area metric compares the perimeter of a district to the
area that is contained within that perimeter; this metric captures the irregularity or jaggedness of a
district’s border by calculating the ratio of the district’s area to the square of its perimeter. (Tr. 2:25); see
also Perry, 548 U.S. at 455 n.2. However, in another exhibit submitted by Plaintiffs, Plaintiffs describe
the perimeter to area metric as “the ratio of the area of a circle with the same perimeter as the district to
the area of the district.” (Plaintiffs’ Ex. 62, at 1 n.2). The latter definition coincides with the definition
found in the case law and the definition proposed by commentators. See, e.g., Vieth v. Jubelirer, 541
U.S. 267, 349 n.3 (2004); Pildes & Niemi, Expressive Harms, 92 Mich. L. Rev. at 556; Daniel D. Polsby
& Robert D. Popper, The Third Criterion: Compactness as a Procedural Safeguard Against Partisan
67
Gerrymandering, 9 Yale L. & Pol’y Rev. 301, 345 (1991). In either event, “the intuitive justification for
this measure is that a given perimeter length will enclose the most area if the shape it surrounds is a
circle.” Pildes & Niemi, Expressive Harms, 92 Mich. L. Rev. at 555. In general, districts that have
smooth borders and relatively regular shapes will have shorter boundaries and enclose considerable area
given the boundary length. They therefore score high on this perimeter measure. Jagged district borders
substantially lengthen the boundary without enclosing more area and hence score low. Pildes & Niemi,
Expressive Harms, 92 Mich. L. Rev. at 556.
The population technique is designed to ascertain the distribution (or dispersion) of voters in the
district. See (Tr. 2:92-93). The case law has described this metric as measuring the degree to which the
population distribution deviates from the district’s population-weighted center of gravity. See (Tr. 2:25);
Karcher v. Daggett, 462 U.S. 725, 756 n.19 (1983) (Stevens, J.) (concurring); see also Polsby and
Popper, The Third Criterion, 9 Yale L. & Pol’y Rev. 301, 345 (1991) (explaining the mechanics of the
population rubber band method in the context of a constitutional compactness). In this measure, the
population of a district is compared to the population of nearby territory by utilizing a ratio where the
district’s population is the numerator and a “rubber band” area around the district is the denominator.
Pildes & Niemi, Expressive Harms, 92 Mich. L. Rev. at 557. The “rubber band” is “the area that would
be inside a rubber band stretched tightly around the district.” Pildes & Niemi, Expressive Harms, 92
Mich. L. Rev. at 556; Vieth, 541 U.S. at 349 n.3.
In Perry, the Supreme Court acknowledged that district courts could use the quantitative
measures of compactness to gauge the compactness of the proposed Gingles maps. See Perry, 548 U.S.
at 455 (describing the perimeter-to-area metric as a “standard measure” of compactness in Section 2
cases); see generally Vieth, 541 U.S. at 348 (acknowledging the value of considering quantitative
measures of compactness, like area rubber band, perimeter-to-area, and population rubber band, when
evaluating the compactness of a map); Pildes & Niemi, Expressive Harms, 92 Mich. L. Rev. at 553-59
(describing dispersion (area rubber band), perimeter (perimeter to area), and population (population
68
rubber band) as robust statistical measures of compactness). Accordingly, the Court finds that the three
quantitative metrics of compactness used by Plaintiffs are probative, but not themselves determinative,
on the question of compactness. Mr. Korbel testified these compactness scores were calculated by the
Texas Legislative Council. (Tr. 2:35). The Court thus finds that Plaintiffs’ compactness data is reliable.
Accordingly, the compactness of Plaintiffs’ maps will be assessed, in part, by evaluating the
compactness statistics supplied by the Texas Legislative Council.
In this case, Mr. Korbel compared the compactness metrics of Plaintiffs’ proposed plans to the
Revised A-1 map, and the 2001 map. The results are summarized in the next table.
COMPACTNESS: COMPARING THE COMPACTNESS METRICS OF PLAINTIFFS’
PROPOSED PLANS, THE REVISED A-1 MAP, and the 2001 MAP
Plan
Precinct
Area Rubber
Perimeter-toPopulation
Band
Area
Rubber Band
1
0.442
0.082
0.403
36
The 200l Plan
2
0.679
0.208
0.754
3
0.796
0.298
0.801
4
0.573
0.165
0.573
The Revised Plan
A-137
1
2
3
4
0.454
0.635
0.781
0.519
0.065
0.119
0.236
0.108
0.388
0.675
0.745
0.542
Korbel 257
(Gingles Plan
#1)38
1
2
3
4
0.572
0.513
0.729
0.507
0.099
0.051
0.178
0.084
0.577
0.522
0.587
0.387
Korbel 325
(Gingles Plan
#2)39
1
2
3
4
0.554
0.541
0.597
0.496
0.097
0.056
0.176
0.061
0.549
0.574
0.467
0.353
36
Source: Plaintiffs’ Ex. 48, at 3; see also Plaintiffs’ Ex. 58, at 1.
Source: Plaintiffs’ Ex. 48, at 6; see also Plaintiffs’ Ex. 58, at 1.
38
Source: Plaintiffs’ Ex. 46, at 3; see also Plaintiffs’ Ex. 58, at 1.
39
Source: Plaintiffs’ Ex. 57, at 5.
37
69
(a)
Korbel 257
Mr. Korbel testified that the compactness scores for Korbel 257 were better than the
compactness scores of either the 2001 map or the Revised A-1 map. (Tr. 2:32). The evidence shows that
the area rubber band, perimeter-to-area, and population rubber band scores for Precinct 1 are higher in
Korbel 257 than they are in both the 2001 Map and the Revised A-1 map, (Tr. 2:35), but the area rubber
band, perimeter-to-area, and population rubber band scores for precincts 2, 3, and 4 are worse under
Korbel 257 than they are under either the 2001 map or the Revised Plan A-1. (Tr. 2:32). Mr. Korbel,
however, stressed that the differences in the scores were slight; in his view, the area rubber band,
perimeter to area, and population rubber band scores of all four precincts under all three maps are “so
close that [he] wouldn’t judge them to be really very different.” (Tr. 2:47). Even though the 2001 Map
and the Revised Plan A-1 outperformed Korbel 257 in precincts 2, 3, and 4, Mr. Korbel contends that, in
the aggregate, the compactness scores for all three plans are comparable. (Tr. 2:32). Moreover, Mr.
Korbel explained that Korbel 257’s compactness scores were within the expected range for relatively
compact districts drawn in a large urban area. (Tr. 2:89).
In fact, Mr. Korbel testified that Korbel 257’s compactness scores were similar to the
compactness scores for the four Texas Senate districts covering Harris County. (Tr. 2:46). For instance,
as detailed in the chart below, Texas Senate District 13, which covers much of Precinct 1 as that precinct
is drawn in Revised A-1, has an area rubber band score of 0.529 and a perimeter-to-area score of 0.132.
See (Plaintiffs’ Ex. 62.1, at 3). Senate District 6, which snakes around Senate District 13 in much the
same way that Plaintiffs’ proposed Precinct 2 navigates around Plaintiffs’ proposed Precinct 1, has an
area rubber band score of 0.516 and a perimeter-to-area score of 0.072. See (Plaintiffs’ Ex. 62.1, at 3).
Senate District 15, which carves out the northeastern-north central portion of Harris County, in much the
same way as Plaintiffs’ proposed Precinct 4, has an area rubber band score of 0.510 and a perimeter-toarea score of 0.08. See (Plaintiffs’ Ex. 62.1, at 3). Finally, Senate District 7, which encompasses the
north-central to north-western and western portions of the county, much like Plaintiffs’ proposed
70
Precinct 3, has an area rubber band of 0.625 and a perimeter-to-area score of 0.187. See (Plaintiffs’ Ex.
62.1, at 3).
COMPACTNESS: COMPACTNESS METRICS OF THE TEXAS STATE CONGRESSIONAL
AND SENATORIAL DISTRICTS THAT ARE PRIMARILY IN HARRIS COUNTY40
Plan
District
Area Rubber Band
Perimeter-to-Area
6
0.516
0.072
2012 Texas Senate Districts41
7
0.625
0.187
13
0.529
0.132
15
0.510
0.08
2
7
9
18
29
2012 Texas Congressional
Districts42
0.421
0.554
0.607
0.587
0.600
0.074
0.136
0.156
0.080
0.086
Accordingly, Mr. Korbel concluded that Korbel 257 was compact because the map’s
compactness scores were comparable to the compactness scores for the County’s Plan and the
compactness scores for other legislative districts in Harris County.
(b)
Korbel 325
The compactness metrics for Korbel 325 are comparable to the compactness metrics for Korbel
257. For instance, under Korbel 325, Precinct 2 had an area rubber band of 0.541, a perimeter to area
ratio of 0.056, and a population rubber band ratio of 0.574. See (Plaintiffs’ Ex. 57, at 5). Under the
Revised A-1 map, Precinct 2 has an area rubber band ratio of 0.635, a perimeter to area ratio of 0.119,
and a population rubber band ratio of 0.675. (Plaintiffs’ Ex. 48, at 6). Mr. Korbel explained, however,
that the area rubber band, perimeter-to-area, and population rubber band measures under Plaintiffs’
demonstration map falls within the acceptable range of scores/ratios for election districts drawn for
Harris County. Mr. Korbel opined that Korbel 325 was more compact than the Revised Plan A-1,
40
Of the Senate districts listed, all the districts listed are contained entirely in Harris County, with the exception
of district 13, which extends into Fort Bend County. Of the congressional districts listed, all the districts are
contained entirely in Harris County, except for district 9, which extends into Fort Bend County.
41
Source: Plaintiffs’ Ex. 62, at 1; see also Plaintiffs’ Ex. 62. 1, at 4.
42
Source: Plaintiffs’ Ex. 62, at 2; see also Plaintiffs’ Ex. 62. 1, at 3.
71
because all three compactness metrics for Precinct 1 are higher under his plan than they are under the
Revised Plan A-1 and Precincts 2, 3, and 4 have approximately the same compactness ratios as the
Revised Plan A-1. See (Tr. 2:67); compare (Plaintiffs’ Ex. 48, at 6) with (Plaintiffs’ Ex. 57, at 5) and
(Plaintiffs’ Ex. 46, at 3). In his view, the plans are all essentially the same in terms of compactness. (Tr.
2:67).
b.
Compactness: Compliance With Traditional Redistricting Principles
As for the compliance with the traditional districting principles assessment, Plaintiffs and
Defendants offer two competing theories for the way in which this assessment should be conducted.
Under the case law, the “[section] 2 compactness inquiry should take into account ‘traditional districting
principles such as maintaining communities of interest and traditional boundaries.’” Abrams v. Johnson,
521 U.S. 74, 92 (1997) (quoting Bush v. Vera, 517 U.S. 952, 977 (1996)); Miller v. Johnson, 515 U.S.
900, 916 (1995) (recognizing “respect for political subdivisions or communities defined by actual shared
interests” as a traditional, race-neutral districting principle); Miller, 515 U.S. at 916 (identifying
compactness and contiguity as traditional redistricting principles); Shaw v. Reno, 509 U.S. 630, 651-52
(1993) (identifying population equality as a traditional districting principle).
Defendants, relying on Sensley v. Allbritton, 385 F.3d 591 (5th Cir. 2004), argue that the
compliance with traditional districting principles component of the compactness inquiry requires
Plaintiffs to show that the Latino community in Harris County is sufficiently large and geographically
compact to constitute a majority in a single commissioner precinct drawn in accordance with the
traditional redistricting principles employed by Harris County when the county is formulating an
electoral map. (Instrument No. 93, at 6-9; Instrument No. 92, at 11 ¶ 18; Instrument No. 104, at 8-18).
Plaintiffs, however, object to Defendants’ interpretation. According to Plaintiffs, the compliance with
traditional districting principles component of the compactness analysis is a malleable inquiry; therefore,
while the demonstration maps must respect the County’s traditional redistricting principles, the
demonstration maps do not have to comply with the traditional redistricting principles to the same
72
degree as an adopted map. (Instrument No. 95, at 6). Thus, the parties’ dispute centers on the degree to
which the demonstration maps must comport with the challenged jurisdiction’s traditional redistricting
principles.
Both parties partially predicate their claims on Sensley. Sensley simply stands for the proposition
that the first Gingles factor requires Plaintiffs to show that their map comports with traditional
redistricting principles. Sensley does not, however, address the degree to which the demonstration maps
must comply with these principles.
Political subdivisions, like Harris County, have many tools with which to craft a dilutive map.
See generally Voinovich, 507 U.S. at 153-54 (explaining how political subdivisions may manipulate
district lines to dilute the voting strength of a politically cohesive minority group). To be sure, this Court
does not mean to suggest that Harris County has intentionally wielded those tools of dilution to devise a
dilutive map. However, the Section 2 analysis focuses on effect, not intent. “The essence of a Section 2
claim is that a certain electoral law, practice, or structure interacts with social and historical conditions
to cause an inequality in the opportunities enjoyed by [Latino] and white voters to elect their preferred
representatives.” Gingles, 478 U.S. at 47; see also De Grandy, 512 U.S. at 1006-07.
In this case, Harris County has selected eight traditional redistricting principles to guide the
County in its redistricting efforts. See (Instrument No. 91, at 15-17 ¶ 14-15). Thus, any electoral map
drafted by the County is drafted in accordance with these eight redistricting principles. When a Section 2
plaintiff challenges the defendants’ redistricting scheme, the plaintiff is implicitly asserting that the
County’s application of these eight districting principles operates to deprive the plaintiff and other
members of her protected class the opportunity to elect candidates of their choice. In essence, the
plaintiff alleges that the County’s choice and application of redistricting principles interacts with social
and historical conditions to impair the ability of the protected class to equally participate in the political
process. See De Grandy, 512 U.S. at 1007.
73
Given that Plaintiffs contend that the selection and application of the traditional redistricting
principles chosen by the County dilutes their voting power, it would be unfair to require Plaintiffs to
draw maps in strict accordance with the County’s priorities. Under this scheme, the entire Section 2
analysis is infected by which traditional redistricting principles the County has prioritized, thereby
precluding any meaningful review of the dilutive effect, if any, of the County’s choice and application of
its chosen redistricting principles. If compliance with the traditional redistricting principles is defined as
rigidly as Defendants propose, then voting rights cases could be defeated at the outset by the very
barriers to political participation that Congress has sought to remove: legislative bodies could evade
compliance with the Voting Rights Act by carefully selecting an array of redistricting principles such
that it would be difficult for a plaintiff to draw a demonstration map that would both differ from the
defendant’s map and yet still comply with each of the defendant’s redistricting principles. Surely,
Congress did not intend for the broad remedial scope of Section 2 to be so easily evaded by a
defendants’ selection of redistricting principles. The scope of the statute must be construed to avoid this
anomalous result. See Chisom v. Roemer, 501 U.S. 380, 403-04 (1991); Allen v. State Board of
Elections, 393 U.S. 544, 567 (1969) (the Voting Rights Act should be interpreted in a manner that
provides “the broadest possible scope” in combating racial discrimination).
Moreover, the Court finds that Defendants’ rigid formulation of the traditional redistricting
principles is inconsistent with the litigation framework contemplated by Gingles and its progeny, in that
it conflates the liability analysis with the remedy analysis. In a Section 2 case, the plaintiffs must prove
that, under the totality of the circumstances, the challenged redistricting scheme dilutes the minority
group’s voting power. If the plaintiffs successfully prove dilution, then the defendant jurisdiction must
modify its map to cure the dilution; the defendant does not, however, have to adopt the maps proposed
by the plaintiffs. Clark II, 88 F.3d at 1407; Westwego III, 946 F. 2d at 1124. Thus, every Section 2 case
may be divided into two phases: a liability phase, where the Court determines whether the challenged
electoral device dilutes minority voting power, and a remedy phase, where the challenged jurisdiction
74
remedies the dilution. Under this scheme, the ultimate viability and effectiveness of a remedy is
considered at the remedial stage of litigation and not during analysis of the Gingles preconditions. Clark
II, 88 F.3d at 1407. According to the County, a rigid formulation of the compliance with traditional
redistricting principles inquiry ensures that the demonstration maps are viable and stable alternatives to
the defendant’s redistricting scheme. Although the logic of this contention is seductive, the argument
misconstrues the goal of the first Gingles precondition. The ultimate end of the first Gingles
precondition is to prove that a solution is possible, not necessarily to present the ultimate solution to the
problem. Gingles, 478 U.S. at 50 n.17; Houston v. Lafayette Cnty., 56 F.3d 606, 612 (5th Cir. 1995). In
“Gingles, the Court provided that it is sufficient that a plaintiff show that a workable plan for another
minority-controlled voting district is possible; the plaintiff’s plan need not be an ultimate solution.”
Fairley, 584 F.3d at 671. Defendants’ reading collapses the liability and remedy determinations into a
single threshold analysis. This is plainly at odds with the scope of the first Gingles precondition because
the first question asks only whether any violation has occurred at all, not whether a remedy is possible.
Aside from these theoretical and structural objections, the Court rejects Defendants’ reading
because it is impractical. Map-drawing rarely, if ever, involves rote application of the redistricting
principles chosen by the state or county. Even in a case like the one at hand, where the defendant’s
redistricting efforts are informed by a handful of traditional redistricting precepts, these principles are
applied dexterously, with the defendant balancing many competing considerations. A defendant does not
assign each of the selected redistricting principles the same importance; it prioritizes these principles,
such that some of these principles anchor the redistricting process whereas others are merely buoys
demarcating the outer bounds of the redistricting process. Despite this prioritization, occasionally, the
defendant’s redistricting principles will be in tension with one another, such that adherence to one
redistricting principle necessitates the subjugation of a competing principle. Given the nuance and
discretion involved in the redistricting calculus, a defendant would always be able to object to a
75
plaintiff’s demonstration map on the grounds that the plaintiff did not weigh competing considerations
in the same manner as the defendant would have.
While Plaintiffs are obligated to show that their maps comply with traditional redistricting
principles, Plaintiffs are not required to prioritize those principles in the same manner as Defendants.
Plaintiffs need only show a “well-developed, legally-adequate plan that can be adjusted during [the
remedial stage].” Fairley, 584 F.3d at 671 n.14.
i.
Preserving Communities of Interest
(a)
Korbel 257
In composing Korbel 257, Mr. Korbel’s main priorities were to: (1) draw a map with majority
Latino CVAP (2) without cutting or splitting any cities or census-designated places and (3) maintain a
top-to-bottom deviation of 10% or less. (Tr. 2:30-31). Mr. Korbel testified that Korbel 257 does not split
any voting precincts or voting tabulation districts, whereas the Revised Plan A-1 split fourteen districts.
(Tr. 2:34). According to Mr. Korbel, if he cut the same number of precincts in his map as the County did
in its Revised A-1 map, then he could have reduced the total population deviation to near zero. (Tr.
2:30).
Although the map does not split any voting districts, it does split two cities, Baytown and
Pasadena, and one census-designated place, Channelview. (Tr. 2:86). Mr. Korbel testified that the
condition that the map not split any cities or census-designated places makes it more difficult to reduce
the top-to-bottom deviation. According to Mr. Korbel, he split Baytown because if he left the city whole
then there was no way for him to draw a map that would get through to the rest of the less urban areas in
Harris County. (Tr. 2:87). Although Mr. Korbel didn’t explain the reason he split Pasadena, he testified
that after splitting the city, he did not think to reassemble the City because even with Pasadena split
between two commissioner’s precincts, Precinct 2 had enough population that it was no longer
unconstitutionally under-populated. See (Tr. 2:87). Although Mr. Korbel admitted that under both
Korbel 257 and Korbel 325 Precinct 2 is the most under-populated, see (Tr. 2:87); see also (Plaintiffs’
76
Ex. 46, at 6); (Plaintiffs’ Ex. 57, at 1), Mr. Korbel admitted that the reason he underpopulated Precinct 2
was so that he could maximize the Latino population. (Tr. 2:88).
Defendants, however, point out that under Korbel 257, two commissioners are co-located in the
same precinct and two commissioners have been moved out of their precincts. Under Korbel 257,
Commissioner Morman, who is the commissioner for Precinct 2, would reside in Precinct 4 and
Commissioner Cagle, the commissioner of Precinct 4, would reside in Precinct 3 with Commissioner
Radack, the commissioner for Precinct 3. (Tr. 3: 239); (Plaintiffs’ Ex. 16; Plaintiffs’ Ex. 46). Mr. Korbel
admitted that when he drew his two Gingles plans, he drew them without consideration of where the
incumbent commissioners resided. (Tr. 2:81). Mr. Korbel further testified that drawing compact
commissioner’s court precincts in Harris County is a challenge because three of the commissioners (the
commissioners for Precinct 1, 3, and 4) all live within close proximity to one another, see (Plaintiffs’ Ex.
52, at 6), making it difficult to draw a compact and normal shaped map in which each commissioner
lives in his or her respective commissioner precinct. (Tr. 2:18).
In Dr. Murray’s view, Korbel 257 reaches into the heart of Baytown to remove a sizeable Latino
population; divides Pasadena; then loops around to Redneck Alley; then moves west, before finally
moving south into Spring Branch in an effort to amass sizeable concentrations of Latino population. (Tr.
3:238). According to Dr. Murray, the impact of splitting Pasadena and Baytown cannot be overstated.
Dr. Murray testified that, historically, these cities have always been in Precinct 2 and thus have always
had one commissioner. If Korbel 257 was implemented, then part of both cities would be in Precinct 4
while the other portion of the cities would remain in Precinct 2. Coordination of service delivery is far
more difficult when a city is represented by two Commissioners rather than one. (Tr. 3:240).
In Dr. Murray’s view, this map was designed to maximize the Latino population in the district.
(Tr. 3:238). Dr. Murray believes this map would dramatically alter Precinct 2 from its present state.
77
(b)
Korbel 325
As for Plaintiffs’ second plan, Korbel 325, Plaintiffs contend that Korbel 325 is, in large part, the
same map as Korbel 257. (Tr. 2:58). However, Korbel 325 differs from Korbel 257 in that Korbel 325
cuts fewer cities and census designated places and it also ensures that none of the four commissioners
are co-located in the same precinct. See (Tr. 2:58). This map also differs from Korbel 257 in that Korbel
325 splits two voting tabulation districts (VTDs); under this map, VTD #220 is split between Precincts 2
and 4, and VTD #346 is split between Precincts 2 and 4. See (Plaintiffs’ Ex. 57, at 5); (Tr. 2:65). Mr.
Korbel explained that these two voting tabulation districts were split to maintain the boundary lines of a
city and another census-designated place. (Tr. 2:65). In order to keep the city together, he had to cut the
two voting tabulation districts (VTD 220 and VTD 346). In his view, it was an either-or scenario, where
Mr. Korbel was forced to choose between cutting a city and retaining the integrity (and unity) of all the
voting tabulation districts or cutting two voting tabulation precincts, but preserving the integrity (and
unity) of all the cities in Harris County.
Dr. Murray testified that while this map essentially realigns Precincts 3 and 4, the map is really
designed with Precincts 1 and 2 in mind. (Tr. 3:244). According to Dr. Murray, Korbel 325 is essentially
the same map as Korbel 257 and every criticism leveled against that map applies with equal force to this
map. (Tr. 3:245). The only real difference between Korbel 257 and Korbel 325 is that in Korbel 325,
Commissioners Cagle, Radack, and Morman are no longer co-located in Precinct 4. (Tr. 3:245). The
map does not, however, resolve any of the issues in Precinct 1; this map, like Korbel 257, continues to
divest Precinct 1 of all its important service facilities and adds population in Alief, which does not
comport well with Precinct 1’s status as a performing opportunity district. (Tr. 3: 245).
ii.
Incumbent-Constituency Relations
Mr. Korbel explained that when he draws a Gingles map he does not pay attention to the number
of constituents that are actually moved to a different precinct. (Tr. 2:73). In his view, Gingles maps are
not bound by the prior maps; thus, he when he drafts a map, he does not always start with the existing
78
map and then draw a map that cures the dilution existing therein. (Tr. 2:78). Therefore, when he drew
his Gingles maps, he did not use the existing County map as his template, but rather started whole-cloth,
because he thinks that is a better way to draw a Gingles map. (Tr. 2:79).
Dr. Alford testified that in reviewing Plaintiffs’ plans, he performed an analysis on the number of
people shifted to new commissioners court precincts under Plaintiffs’ plan. (Tr. 4:47). The results from
Dr. Alford’s analysis, which are included in Defendants’ Exhibit 67, are replicated in the table below.
According to Dr. Alford, the chart below is the end product of a standard analysis performed by TLC
that details the proportion of the population moved in a redistricting plan compared to a base plan. (Tr.
4:48). The table below describes the population moved out of each precinct under Plaintiffs’ proposed
plans and under the County’s Revised Plan A-1. (Tr. 4:48).
Dr. Alford explained that he undertook this analysis because there has been a great deal of
discussion concerning the disruption of services and population occasioned by the irregular nature of
Plaintiffs’ proposed districts. (Tr. 4:51). His analysis offers a means by which to gauge the real impact
of that disruption by examining the number of people the map moves. According to Dr. Alford,
evaluating the effect of the maps based on the number of people the map moves rather than the amount
of physical territory the map moves is a better measure of the real impact of the map. This is because
sometimes a map might incur major territorial disruptions while looking on a map, but those territorial
disruptions may not involve a lot of people. On the other hand, sometimes a small territorial disruption
may be densely populated, affecting a large share of the population. Therefore, merely examining the
territorial shifts occasioned by the map does not always offer insight into the scope and impact of the
proposed redistricting change. (Tr. 4:51).
79
VOTING AGE POPULATION MOVED TO A NEW COMMISSIONER’S PRECINCT FROM
THE BASE (THE 2001 MAP) OF THE EXISTING 2010 EXISTING PRECINCTS43
Korbel 257
Korbel 325
Revised Plan A-1
298,248 (40.5%)
Moved out of
292,586 (38.3%)
48, 799 (6.6%)
Precinct 1
Moved out
Precinct 2
Moved out
Precinct 3
of
Moved out
Precinct 4
236,463 (32.1%)
of
of
275,346 (36.0%)
29, 344 (4.0%)
327,113 (44.4%)
314,085 (41.1%)
Total
Moved
County-wide
65, 935 (8. 9%)
361,269 (49.0%)
373,015 (48.8%)
112,822 (15.2%)
1,255,032 (42.6%)
1,223,093 (41.5%)
256,900 (8.7%)
Dr. Alford explained that Plaintiffs’ Korbel 257 moves 1,255,032 people; this map would, if
implemented, move almost 43% of the voting age population of Harris County. (Tr. 4:49); (Defendants’
Ex. 67). Korbel 325 offers a slight improvement over Korbel 257, in that it only moves 1,233,093
people, or approximately 42% of the county-wide voting age population. (Tr. 4:49). Under either Korbel
257 or Korbel 325, Plaintiffs’ plan contemplates shifting almost two-fifths of the county-wide voting
age population. In contrast, the County’s Revised Plan A-1 only moves 256,900 adults. Thus, it only
moves approximately 9% of the county-wide adult population. (Tr. 4:51).
Defendants’ fact witness Mr. Frank King, the County director of budget and planning, examined
Plaintiffs’ Gingles maps and assessed the impact those maps would have on facilities and real estate in
the County.44 His results are encapsulated in the charts below.
43
Source: Defendants’ Ex. 67, at 1-2. For ease of reading, the Count’s chart condenses the two charts contained in
Defendants’ Ex. 67 into one chart.
44
At the time of trial, Mr. Frank King was the Director of Budget and Planning for Harris County. He has been
employed with Harris County for over thirteen years. (Tr. 4:101). His first five years of service were with the
Harris County tax office as the director of accounting; in 2004, he moved to the budget office, where he
eventually assumed his current position. (Tr. 4:102). Mr. King explained that he examined Plaintiffs’ maps and
assessed the impact those maps would have on facilities and real estate in the County.
80
FACILITIES AND REAL ESTATE: EFFECT OF REDISTRICTING ON COUNTY ASSETS
(NUMBER OF ASSETS CHANGED)45
Korbel 257
Korbel 325
Revised Plan A-1
Road Miles (Unincorporated Areas)
Bridge Miles46
Number of Maintenance Camps
Acres of Parks47
Number of Senior/Community Centers48
Libraries49
Physical Area in Square Miles50
2,207
8
5
6,869
5
10
629
2,489
9
6
9,749
21
12
698
1,035
3
2
520
3
4
232
The first chart shows the number of county assets—road miles, bridge miles, maintenance
camps, park acreage, community centers, libraries, and physical area—that move from one Precinct to
another Precinct under the different maps. At trial, Mr. King further explained the significance of the
road miles and the maintenance camps metrics.
King explained that the chart discusses the number of road miles in unincorporated areas because
the road miles within the limits of a municipality are maintained by that municipality. Thus, for
example, the road miles within the limits of the City of Houston are maintained by the City of Houston.
(Tr. 4:103). However, unincorporated miles are maintained by the County Commissioners. (Tr. 4:103).
There are a total of 6,902 total unincorporated miles. (Tr. 4: 106).
As for the number of maintenance camps, Mr. King explained that County-wide there are
thirteen road camps and each commissioner uses the road camps within his Precinct to maintain the
roads in the unincorporated areas of the Precinct. (Tr. 4:104-06).
45
Source: Defendants’ Ex. 69.
Bridge miles are simply the number of miles of bridges connected to the unincorporated roads. (Tr. 4:103).
County-wide there are 20 miles of bridges. (Tr. 4:106).
47
County-wide there are 26,601 acres of park (Tr. 4:106)
48
County-wide there are 35 community centers. (Tr. 4:104).
49
County-wide there are 28 libraries (Tr. 4:107).
50
Mr. King explained that Harris County encompasses 1749 square miles of land. (Tr. 4:104).
46
81
FACILITIES AND REAL ESTATE: EFFECT OF REDISTRICTING ON COUNTY ASSETS
(PERCENTAGES OF ASSETS CHANGED)51
Korbel 257
Korbel 325
Revised A-1
37%
41%
17%
Road Miles (Unincorporated Areas)
40%
46%
14%
Bridge Miles
38%
46%
15%
Number of Maintenance Camps
26%
37%
2%
Acres of Parks
14%
60%
9%
Number of Senior/Community Centers
36%
43%
14%
Libraries
Physical Area in Square Miles
36%
40%
13%
This second chart shows the percentage of county assets that change under the different maps.
(Tr. 4:104). Mr. King explained that Plaintiffs’ proposed maps move more county resources, regardless
of the metric used–be that road miles, bridge miles, number of maintenance camps, acres of park,
community centers, libraries, or physical area. Plaintiffs’ maps re-allocate more county resources than
the County’s Revised A-1 map. See (Tr. 4:106-107). For instance, Korbel 257 and Korbel 325 reallocate
36% and 40%, respectively, of the County’s land mass, whereas the Revised Plan A-1, only reallocates
13% of the County’s land mass.
Dr. Murray further testified that Korbel 257 and Korbel 325 occasion many service disruptions
and changes. Most of these changes are leveled against Precinct 1. Under both maps, Precinct 1 loses the
Challenger 7 Facility, the Deussen Park facility, and the El Franco Lee Center to Precinct 4. Moreover,
Precinct 1 losses the Hardy Road Service Facility to Precinct 2. (Tr. 3:242-43); see also (Defendants’
Ex. 33, at 1). So under these maps, Precinct 1 is divested of four of its most important service facilities.
Precinct 2 also suffers some service disruption, numerous road camps that are currently in Precinct 2 are
shifted to Precinct 4 and five community centers in Precinct 2 are also shifted over to Precinct 4. (Tr.
3:243). Moreover, Precinct 2 does not pick up many centers or facilities under Korbel 257, which
suggests that the map was not driven by considerations about service facilities and infrastructure. (Tr.
3:244).
51
Source: Defendants’ Ex. 69.
82
Ms. Helen Berrott-Timms, the Director of Community Programs and Facilities in Precinct 2, was
called as a witness by Defendants and testified that moving community centers from one precinct to
another precinct can affect the type of programming and services the community center delivers to its
constituents. Ms. Berrott-Timms explained that each precinct has a different method of funding the
community centers under its jurisdiction. Therefore, moving a community center from one precinct to
another could jeopardize the type and scope of funding that community center receives, which
ultimately jeopardizes the center’s ability to deliver the same type of programming. See (Tr. 4: 74-80,
85-87, 98). That said, Ms. Berrott-Timms also testified that ultimately, the four county commissioners
work together and employ the County assets, wherever they may be located, to service the needs of the
residents of Harris County. See (Tr. 4: 93).
On the whole, the Court finds that the movement of the assets as contemplated by Plaintiffs’
proposed maps would be extremely disruptive.
iii.
Traditional Redistricting Principles: The Preservation of
Precinct 1 as an Effective Minority Opportunity District
As part of the traditional redistricting principles inquiry, Plaintiffs must show that their map
preserves Precinct 1 as an effective minority opportunity district. In compliance with this principle, the
County has drawn Precinct 1 such that it includes the sizeable home-owning African-American
population in the Bush Airport area, thereby increasing the African-American share of Precinct 1’s
citizen voting age population. However, Plaintiffs’ plans do not include this area in Precinct 1. Instead,
Plaintiffs address Precinct 1’s opportunity district performance problems by extending Precinct 1 to
include areas of Southwest Houston. Under Plaintiffs’ maps, Precinct 1 includes the populous and
ethnically diverse area of Alief. According to Dr. Murray, implementing Korbel 257 or Korbel 325
would adversely impact Precinct 1, because under these maps deprives Precinct 1 of the home-owning
African-American population in the Bush Airport area, which was added in the Revised A-1 map. (Tr.
3:240).
83
At trial, Mr. Korbel testified that when crafting his proposed maps he studied or was conscious
of the African-American growth rates in Harris County. See (Tr. 2:84). According to Mr. Korbel, the
African-American population in Harris County is growing anemically. See (Tr. 2:84). For the past three
decades, the Latino population has been moving into areas that were formerly occupied by AfricanAmericans. As a result, the African-American community is retracting on the edges of the Latino
community.
Mr. Korbel testified that he has noticed a pattern in some less affluent communities in which
Anglos first occupy a community, then the Anglos move out and African-Americans and Latinos move
in. Next, there is a period of overlap, in which Blacks and Latinos constitute comparable shares of the
community’s population. Then, there is a tipping point at which the black population decreases and the
Latino population in the census tract increases, which results in the community becoming predominately
Latino (Tr.2:99-100).
The Alief area, which Mr. Korbel has included in Precinct 1, is in the middle of this demographic
lifecycle. Today, there is a sizeable African-American population in Alief, but the concentration of
blacks is decreasing. (Tr. 2:101). The Alief area has a high concentration of apartments. (Tr. 2:101). Mr.
Korbel admitted that people of moderate means, irrespective of race, generally live in apartments. (Tr.
2:105). And he predicted that if the current population trends of lower income Latinos moving into the
Alief area continues, the apartments in Alief that are currently in Plaintiffs’ proposed Precinct 1 will
become Latino-occupied in the next ten years. See (Tr. 2:105-06). But Mr. Korbel aptly noted that to the
extent that there have been dramatic population shifts in ten years, then there will also be the opportunity
for redistricting, which will address those shifts. (Tr. 2:106).
Plaintiffs’ witness Senator Sylvia Garcia,52 on cross-examination, admitted that given a choice
between a voting precinct with majority apartment-dwellers and a precinct where the majority of the
52
Sylvia Garcia, the former Precinct 2 County Commissioner, is currently a Texas state senator and is originally
from South Texas. (Tr. 3:4). She obtained her bachelor’s degree in social work and government from Texas
84
residents are home-dwellers, Garcia would prefer the precinct with the majority home-dwellers because
they are more likely to be remain in the community for a longer period of time; they are less transient.
(Tr. 3:77). As a general principle, she agreed that if established, well-developed and high performing
voting precincts are removed from a commissioner’s precinct and are replaced with voting precincts
composed mainly of apartment dwellers, who are less reliable voters, then you jeopardize the political
effectiveness of the precinct and threaten the ability of the precinct to deliver support for the traditional
Democratic candidate, be that candidate white, black, or Latino. (Tr. 3:78-80).
Plaintiffs’ expert Gerald Birnberg testified that, historically, 90% of Harris County’s AfricanAmerican voted straight Democratic Party, whereas only 60-65% of Harris County’s Latino voters vote
straight Democratic Party. See (Tr. 3:118). On cross-examination, he explained that removing a block of
voters who consistently vote straight Democratic party from the Precinct and replacing those voters with
Women’s University in 1972. After graduating, she obtained her J.D. from Thurgood Marshall School of Law at
Texas Southern University in 1978. (Tr. 3:2).
Upon arriving in Houston, in 1972, Senator Garcia immediately immersed herself in Harris County
politics. (Tr. 3:5, 6). She started as a low-level volunteer, serving, for example, as a block walker or a phone
banker. See (Tr. 3:5). In 1972, she worked for the Gary Hart campaign and later worked on the McGovern
campaign. Later in the decade, she worked on Leonel Castillo’s campaign for City Controller and many of former
state Representative Ben Reyes’ campaigns. (Tr. 3:6). After graduating from law school, Senator Garcia worked
as a legal aid lawyer for three years with the Gulf Coast Legal Aid Foundation in Houston. (Tr. 3:3).
In the 1980s, she began to take a more prominent role in the political structure of the Harris County
Democratic party, becoming a party delegate. (Tr. 3:6-7). During this time, she began to manage races and
coordinating campaigns and started receiving political appointments to local government. (Tr. 3:7). During this
time, Mayor Kathryn Whitmier asked Ms. Garcia to serve as Vice Chair of the City of Houston Appraisal Board.
(Tr. 3:3). In 1985, she served as an administrative law judge for the Equal Employment Opportunity Commission.
(Tr. 3:4). In 1986, the mayor asked Garcia to serve as a municipal court judge for the City of Houston. She
accepted and served as a municipal judge until 1987, when she was selected as chief of the municipal courts. (Tr.
3:4). In 1992, while still chief judge of the municipal courts, Ms. Garcia unsuccessfully ran for United States
Congress to serve as a congresswoman in the newly fashioned Latino opportunity district that covers the eastern
portion of Harris County; she was defeated in the Democratic primary. (Tr. 3:7). In 1996, while still chief judge of
the municipal courts, she ran for County Attorney. (Tr. 3:8). But, as Ms. Garcia explained “that was one of the
years we all experience here in Harris County where the presidential candidate did not [carry the County, so] there
was no coat tails and many of our Democratic judges and others lost during that time, including myself.” (Tr. 3:8).
She served as chief of the municipal courts for ten years, under both Mayor Whitmier and Mayor Bob Lanier,
until she ran for City Controller in 1997. (Tr. 3:4). She was elected Controller and was twice reelected to the
position, serving three consecutive terms. (Tr. 3:4). In the middle of her third term as Controller, she ran in the
2002 election for County Commissioner Court for Precinct 2. She won that race and was re-elected to the position
in 2006, before suffering defeat in the 2010 County Commissioner race. (Tr. 3:4). Subsequent to trial, Ms. Garcia
was elected the Texas Senate.
85
a population of voters that is less inclined to vote straight Democratic ticket impedes that Precinct’s
ability to elect a Democratic candidate. (Tr. 3:124 -25).
Mr. Korbel, however, resisted the suggestion that his maps threatened the continued vitality of
Precinct 1 as a functioning African-American opportunity district. First, he restated his belief that this
was a Gingles plan and was not intended to be an implemented plan. Second, the black voting age
population of Precinct 1 is higher, under his Gingles plan, than it is under the County’s Revised A-1
map. (Tr. 2:104). Under Korbel 257, African-Americans are 48.6% of Precinct 1’s voting age population
(using 2006-10 ACS data), see (Plaintiffs’ Ex. 46, at 5), but African-Americans are only 39.2% of the
voting age population in the Revised Plan A-1, see (Plaintiffs’ Ex. 7, at 1); see (Tr. 2:104). Thus, Mr.
Korbel argues that Plaintiffs’ map preserves and even augments African-American voting strength in
Precinct 1. Mr. Korbel further testified that his maps were drawn in much the same way that the
congressional and Senate districts in Harris County are drawn. (Tr. 2:110).
On balance, the evidence shows that Plaintiffs’ plan would threaten the continued existence of
Precinct 1 as an opportunity district. Plaintiffs’ maps achieve high Latino population in Precinct 2 by
taking population from Precinct 1 and then replacing that population with the highly mobile and
transient population of Alief. As Defendants explained, this is an area that is increasingly becoming
Latino. Although the inclusion of Alief in Precinct 1 increases the African-American population of
Precinct 1, this increase is largely comprised of highly-mobile apartment dwellers in an area whose
African-American population, though large, is beginning to dwindle and will most likely continue to
decline throughout the decade. Accordingly, this addition does not protect Precinct 1’s continued
effectiveness as an opportunity district.
iv.
Summary
On the whole, the Court finds that Plaintiffs’ maps do not respect traditional redistricting
principles. First, the evidence showed that Plaintiffs’ maps threaten the existing African-American
voting opportunity in Precinct 1 by wrapping Precinct 2 around the top of Precinct 1, cutting Precinct 1
86
off from the fast-growing African-American voting precincts in the Bush Airport area. Plaintiffs’ maps
further threaten the existing opportunity district in Precinct 1 by extending Precinct 1’s border too far
southwest into the Alief area, which has experienced high Latino growth and declining AfricanAmerican population. See (Tr. 3:240-41).
Second, Plaintiffs’ maps split the cities of Baytown and Pasadena between Precinct 2 and
Precinct 4, even though those cities have historically been a part of Precinct 2. Moreover, Mr. Korbel
did not articulate any reasoned justification for splitting the two cities.
Third, Plaintiffs’ maps would adversely impact the commissioners’ ability to deliver services to
their constituents. The testimony at trial indicated that Harris County is unique among urban counties in
that a sizeable portion of the County’s population resides in unincorporated areas. Because the residents
of these unincorporated areas do not have a municipal government, the commissioner’s court provides
basic services to the County’s unincorporated residents. Therefore, the provision of services takes on
greater import here in Harris County than it might otherwise. In fact, Dr. Murray testified that there was
a direct correlation between service facilities and incumbency-constituent services. The evidence
indicated that Plaintiffs’ maps impede the delivery of services by moving key Precinct 1 assets, such as
the Challenger 7 facility, Duessen Park, the El Franco Lee Community Center, and the Hardy Road
facility out of Precinct 1. Additionally, Plaintiffs’ plans move key Precinct 2 assets into Precinct 4.
Moreover, Plaintiffs’ maps shift approximately 40% of the total population to a new commissioner’s
precinct, which further disrupts incumbent-constituent relationships.
Although the Court remains steadfast in its conclusion that the compliance with traditional
redistricting principles inquiry should be applied flexibly, the Court finds that Plaintiffs, here, have
failed to demonstrate that their maps were drawn in compliance with traditional redistricting principles.
Plaintiffs’ maps jeopardize Precinct 1’s status as an opportunity district, the maps split two cities that
have long been in Precinct 2, and the maps compromise existing incumbent-constituent relationships by
shifting wide swaths of the population and reallocating County assets. Accordingly, the Court finds that
87
Plaintiffs have failed to establish that their map comports with traditional redistricting principles.
Therefore, Plaintiffs have failed to establish the first Gingles factor.
B.
GINGLES TWO: POLITICAL COHESION53
The second Gingles precondition, whether the minority group is politically cohesive, explores
the extent to which a group of voters shares the same beliefs, ideals, principles, agendas, and concerns,
“such that they generally unite behind or coalesce around particular candidates and issues.” LULAC III,
986 F.2d at 744. In other words, the political cohesiveness inquiry is designed to examine the extent to
which the minority group has distinctive minority group interests such that the affected minority group
has a preferred candidate. See Gingles, 478 U.S. at 56; Cousin v. Sundquist, 145 F.3d 818, 823 (6th Cir.
1998) (the political cohesiveness precondition requires plaintiffs to establish that “voters of the same
race tend to vote alike.”). If the minority group is not politically cohesive—i.e. does not have a
candidate of choice—then it cannot be said that the selection of the electoral structure employed by the
County thwarts distinctive minority group interests. Gingles, 478 U.S. at 51.
In order to satisfy the political cohesiveness precondition, the plaintiff must show that a
significant number of minority group members usually vote for the same candidate. Gingles, 478 U.S. at
57; Cousin v. Sundquist, 145 F.3d 818, 823 (6th Cir. 1998) (the political cohesiveness inquiry asks
whether “voters of the same race tend to vote alike.”). While it may be supposed that minority
candidates will be preferred by minority voters, plaintiffs must introduce evidence that the minority
candidate is actually minority-preferred. Gingles, 478 U.S. at 67-70; see also Growe v. Emison, 507 U.S.
25, 42 (1993) (Section 2 “does not assume the existence of racial bloc voting; plaintiffs must prove it”).
Although statistical evidence of racially polarized voting is frequently employed to demonstrate a
minority group’s political cohesiveness, see infra Part III.C other evidence may also establish this
second Gingles factor. LULAC III, 986 F.2d at 743; Monroe v. City of Woodville, 897 F.2d 763, 764 (5th
53
Although Plaintiffs have failed to establish the first Gingles factor, the Court has conducted the entire Gingles
analysis.
88
Cir. 1990) (“Statistical proof of political cohesion is likely to be the most persuasive form of evidence,
although other evidence may also establish this phenomenon.”). In particular, political cohesiveness may
be demonstrated by statistical evidence of racial bloc voting or testimony from persons familiar with the
community in question, provided that testimony is not rebutted. LULAC III, 986 F.2d at 743-44; see also
Gingles, 478 U.S. at 56; see, e.g., Westwego III, 946 F.2d at 1118 n.12; Brewer v. Ham, 876 F.2d 448,
453-54 (5th Cir. 1989).
In this case, there is no dispute that the Latino voters in Precinct 2 are politically cohesive. See
(Instrument No. 91, at 19 ¶ 37). At trial, fact witnesses for Plaintiffs, such as Sylvia Garcia, Edward
Gonzalez,54 James Rodriguez,55 and Gerry Birnberg, testified that Latinos tend to coalesce around
candidates who are sensitive to the needs of the Latino community. Councilman Gonzalez, for instance,
testified that based on his years of public service and experience with various political campaigns, he
felt that Latinos in Harris County were politically cohesive and voted as a bloc. (Tr. 2:121). In his view,
Latinos often prefer someone that looks like them, comes from the same community as them, and is
sensitive to their language and cultural needs and values. (Tr. 2:122).
54
Mr. Gonzalez is a graduate of the Houston Independent School District. He received his undergraduate degree
from the University of Houston-Downtown and a master’s degree from the University of St. Thomas. At the time
of trial, he was the district council member for District H for the City of Houston. Before his role as a councilman,
Councilman Gonzalez was an eighteen-year veteran of the Houston Police Department, where he served the City
as a police officer, retiring at the rank of sergeant. (Tr. 2:116-17). He currently lives in Precinct 2 and has lived in
Precinct 2 for most of his adult life. (Tr. 2:117). In 2009, Mr. Gonzalez was elected to City Council for the City
of Houston, where he served as a Councilman for District H, a single-member Latino opportunity district. (Tr. 2:
117). Before running for City Council in 2009, he assisted then-council member Adrian Garcia, now-Sheriff
Garcia, in his campaigns, as well as state Representative Jessica Farrar and state Representative Armando Walle
(Tr. 2:118).
55
Mr. James Rodriguez currently lives in Precinct 1 and at the time of trial had lived there for the past six years.
Before that, he had lived in Precinct 2 for thirty-one years. (Tr. 2:129). Currently, Mr. Rodriguez is serving his
third, and final, term as a councilman on the Houston City Council, where he represents District I. (Tr. 2:129).
According to Mr. Rodriguez, District I is predominately Latino and straddles both County Commissioner’s
Precincts 1 and 2, but the majority of District I lies in Precinct 2. (Tr. 2:133). Prior to his career as a councilman,
he was chief of staff to Councilwoman Carol Alvarado for four years. (Tr. 2:129-30). Councilman Rodriguez
stated that he has, personally, only run three campaigns, but he has been involved in numerous county-wide
campaigns, including Senator Garcia’s first campaign for Harris County Commissioner’s Court Precinct 2 in
2002.
89
Mr. Birnberg testified that Latinos exhibit a strong preference for candidates who are perceived
as being sensitive to the cultural needs of their community.56 For example, in the 2012 general election,
there were seven incumbent criminal district court judges who sought reelection: Judges Herb Ritchie,
Hazel Jones, Maria T. Jackson, Shawna Reagin, David Mendoza, Randy Roll and Rueben Guerrero. See
(Tr. 3:117); see also (Defendants’ Ex. 51, at 22-24) (summarizing election results for these seven
candidates). Of these seven, only three, Judge Jackson, Judge Mendoza, and Judge Guerrero, prevailed
in the general election. According to Mr. Birnberg, the victory of the three candidates with names that
connote affiliation with the Latino community57 shows that Latinos are not simply voting for the
Democratic candidate, but rather are voting for candidates based on a perceived affiliation with the
Latino community. See (Tr. 3:117). The candidates perceived to have some affiliation with the Latino
community prevail in the same group of races—e.g. criminal court races—even when other candidates
of the same political party lose their races. (Tr. 3:117). According to Mr. Birnberg, Latinos are voting as
a bloc to elect co-ethnic candidates.
This testimony was augmented by expert testimony from Dr. Barretto, who also testified that
Latinos tend to vote as a bloc. In his expert report, Dr. Barretto analyzed the voting trends of Latinos in
Precinct 2 in some twenty-six county and statewide elections. See (Plaintiffs’ Ex. 39, at 16). In the
endogenous elections involving County Commissioner in Precinct 2, for instance, Dr. Barreto’s
ecological regressions determined that 78.6% of Latinos voted for Sylvia Garcia in 2010, (Plaintiffs’ Ex.
39, at 53), and 94.4 % of Latinos voted for Sylvia Garcia in 2002. (Plaintiffs’ Ex. 39, at 17); see also
(Tr. 1:142). Based on these findings and the results from the other twenty-four exogenous elections
analyzed, Dr. Barreto concluded that “Latino voters demonstrate[] very strong vote cohesion.”
(Plaintiffs’ Ex. 39, at 16). This conclusion is further supported by the expert reports of Dr. Murray, who
56
For example, in 2008, 65% of Latinos voted a straight Democratic ticket, whereas almost 90% of AfricanAmericans voted a straight Democratic ticket. (Tr. 3:123).
57
Mr. Birnberg explained that although Maria Jackson is not Latina, she was perceived as being Latina because
her first name is Maria.
90
observed that Latino voters are reasonably cohesive in general elections, (Defendants’ Ex. 14, at 3), and
Mr. Korbel, who reported that Latinos demonstrate high levels of political cohesion in Harris County
Commissioner elections. (Plaintiffs’ Ex. 41, at 6).
In light of the evidence discussed above and the statistical evidence of racial bloc voting,
discussed infra, the Court finds that Latinos in Precinct 2 are politically cohesive.
C.
GINGLES THREE: THE EXISTENCE OF RACIAL BLOC VOTING
The third Gingles factor, the existence of racial bloc voting, is related to, but distinct from, the
concept of political cohesion. The term political cohesion contemplates that a specified group of voters
shares common beliefs, ideals, principles, agendas, concerns, and the like such that it generally unites
behind or coalesce around particular candidates and issues. In contrast, racial bloc voting, sometimes
referred to as racial polarized voting, “describes an electorate in which white voters favor and vote for
certain candidates or propositions, and minority voters vote for other candidates or propositions.”
LULAC III, 986 F.2d at 744.58 Racial bloc voting “exists where there is a consistent relationship between
the race of the voter and the way in which the voter votes, or to put it differently, where [Latino] voters
and [Anglo] voters vote differently.” Gingles, 478 U.S. at 53 n.21 (internal quotation marks and citations
omitted).
The correct question is not whether white voters demonstrate an unbending or unalterable
hostility to whoever may be the minority group’s representative of choice, but whether, as a practical
matter, such bloc voting is legally significant. LULAC IV, 999 F.2d at 850, LULAC III, 986 F.2d at 745;
see also Jenkins v. Red Clay Consol. Sch. Dist. Bd. of Educ., 4 F.3d 1103, 1123 (3d Cir. 1993). “Bloc
voting [is] a matter of degree, with a variable legal significance depending on other facts.” De Grandy,
512 U.S. at 1011. Accordingly, no simple doctrinal test applies to the third Gingles factor because racial
bloc voting can “vary according to a variety of factual circumstances.” Gingles, 478 U.S. at 58. That
58
Thus, a showing of racially polarized voting will frequently demonstrate that minority voters are politically
cohesive. LULAC III, 986 F.2d at 744.
91
said, generally, “a white bloc vote that normally will defeat the combined strength of minority support
plus white ‘crossover’ votes rises to the level of legally significant white bloc voting.’ Gingles, 478 U.S.
at 56; accord Perry, 548 U.S. at 425; De Grandy 512 U.S. at 1011; LULAC IV, 999 F.2d at 850. It is
“the usual predictability of the majority’s success,” which “distinguishes structural dilution from the
mere loss of an occasional election.” Gingles, 478 U.S. at 51.
The amount of white bloc voting that can generally “minimize or cancel” minority voters’ ability
to elect representatives of their choice, however, will vary from district to district according to a number
of factors, including the nature of the allegedly dilutive electoral mechanism; the percentage of
registered voters in the district who are members of the minority group; and the size of the district.
Gingles, 478 U.S. at 56 (footnote and citations omitted); see also Rangel v. Morales, 8 F.3d 242, 245
(5th Cir. 1993). Thus, for example, “if the minority group constitutes only a small fraction of the total
number of registered voters,” or the district is large “it may be, relatively speaking, easier for the
members of that group to establish their effective submergence in a white majority.” Rangel, 8 F.3d at
245 (discussing intersection of low voter registration and racial bloc voting inquiry); LULAC III, 986
F.2d at 749 (discussing effect the size of the precinct has on the existence of racial bloc voting).
Proving this factor typically requires statistical evaluation of elections. Campos, 840 F.2d at
1244-45. “The number of elections that must be studied in order to determine whether voting is
polarized will vary according to [the] pertinent circumstances.” Gingles, 478 U.S. at 57 n.25; see also
Gingles, 478 U.S. at 58-59, 80-82 (finding data from three election years, involving minority candidates,
sufficient to uphold district court’s vote-dilution ruling). Courts have relied on various statistical
methods. See, e.g., Houston, 56 F.3d at 611 (use of bivariate ecological regression and extreme case
analysis); Clark II, 88 F.3d at 1397 (expert employed regression and homogenous precinct analysis);
Benavidez, 638 F. Supp. 2d at 723-24 (ecological inference).
The first technique is bivariate ecological regression analysis (“BERA”). Ecological
“[r]egression is a mathematical technique used for estimating the best fitting straight line that could be
92
drawn to describe the relationship between two variables in a scatter plot.” Teague v. Attala Cnty., 92
F.3d 283, 289 (5th Cir. 1996). In a voting rights case, regression analysis “compare[s] the density of a
particular population group in a specified area with the percentage of votes received by a particular
candidate in that area.” Perez, 958 F. Supp. at 1215; Benavidez, 638 F. Supp. 2d at 723-24. In this case,
the regression analysis examines the relationship between the concentration of Latino voters in a
precinct and the electoral success of the Latino-preferred candidate.
The ecological regressions performed in this case are based on a scatter plot in which the
horizontal axis reflects the percentage of total voters who are Latino (the independent variable) and the
vertical axis shows the percent of the vote received by the Latino-preferred candidate (the dependent
variable). A straight line, called the line of best fit, is plotted through the data points with the least
possible deviation from the points. This line can be described by the equation y = mx + b, where y is the
dependent variable, m is the slope of the line, x is the independent variable, and b is the y-intercept. See
Benavidez, 638 F. Supp. 2d at 723. The coefficients for the y-intercept and for the slope form the
equation of the actual regression line. The y-intercept defines the point at which the regression line
crosses the vertical axis, giving the predicted value of the votes cast for the Latino-preferred candidate
(the dependent variable) when the percentage of the Latino voters (the independent variable) is zero.
“The slope indicates the rise of the regression line, or the predicted change in the votes cast for the
Latino-preferred candidate (the dependent variable) for a one-percentage-point change in the percentage
of Latino voters (the independent variable).” Perez, 958 F. Supp. at 1215.
In a two-variable regression, such as is the case here, the Court must also determine the degree
of linear association, or correlation, between the two variables examined. The correlation between the
two variables is described by the coefficient, r. Teague, 92 F.3d at 289. In this case, the correlation
coefficient, r, measures the linear relationship between the percentage of Latino voters in the precinct
(the independent variable) and the number of votes cast for the Latino-preferred candidate (the
dependent variable). The correlation coefficient, r, may range from 0 to 1.00, positive or negative: a
93
correlation coefficient of -1.00 indicates a perfectly inverse relationship between the independent and
dependent variables, such that for every change in the independent variable there is a negative change of
equal magnitude in the dependent variable. Teague, 92 F.3d at 289-90. A correlation coefficient of +1.00
indicates a perfectly positive relation relationship between the two variables, such that for every change
in the independent variable there is an equal and positive change in the dependent variable. Teague, 92
F.3d at 289-90. A high correlation coefficient justifies confidence in the underlying data. Teague, 92
F.3d at 290.
Although a high correlation coefficient indicates that the independent and dependent variables
are in fact linearly related, the Court must also examine how well the line of best fit explains the
variance in the actual data. The variance coefficient, r2, describes the proportion of variation in the
dependent variable that is explained or accounted for by the independent variable. Teague, 92 F.3d at
290. Stated otherwise, the r2 value estimates “the extent to which the percentage of voters supporting the
[Latino-preferred] candidate can be expected to change in response to changes in the [percentage of
Latino voters].” Citizens for a Better Gretna v. Gretna, 834 F.2d 496, 499 n.7 (5th Cir. 1987). The r2
ranges from 0 to 1.0 but is always positive. Teague, 92 F.3d at 290. The variance coefficient approaches
1.0 when the change in the support for the Latino-preferred candidate is entirely attributed to the change
in the percentage of Latino voters, but the coefficient approaches 0.0 when the change in support for the
Latino-preferred candidate is not attributable to a change in the percentage of Latino voters. See Gretna,
834 F.2d at 499 n.7; Teague, 92 F.3d at 290.
A second technique, ecological inference (EI), or the King method, is similar to ecological
regression but abandons the assumption of linearity underpinning the ecological regression method.
Benavidez, 638 F. Supp. at 724. Instead, ecological inference assumes that the actual votes of two
groups for two particular candidates are based on fixed underlying propensities but vary from precinct to
precinct in random ways. It estimates the underlying propensity of each group to turn out for an election
and to vote for a particular candidate using the estimation technique of maximum likelihood. Shirt v.
94
Hazeltine, 336 F. Supp. 2d 976, 995 (D.S.D. 2004). Ecological inference further differs from ecological
regression in that ecological inference also applies a principle called the method of bounds to constrain
estimates to real limits by not allowing estimates of the percentage of votes cast for the Latino-preferred
candidate to exceed 100 or fall below zero. Benavidez, 638 F. Supp. at 724.
Certain elections are more probative of unequal electoral opportunity than others. Temporally,
recent elections are more probative than elections in the distant past. See Uno v. City of Holyoke, 72 F.3d
973, 990 (1st Cir. 1995); Meek v. Metro. Dade Cnty., Fla., 985 F.2d 1471, 1482-83 (11th Cir. 1993).
Substantively, “the most probative elections are generally those in which a minority candidate runs
against a white candidate.” Magnolia Bar Ass’n, Inc. v. Lee, 994 F.2d 1143, 1149 (5th Cir. 1993)
(Magnolia Bar Ass’n II); see also Gingles, 478 U.S. 30 at 80-82 (relying exclusively on interracial
legislative contests to determine whether a legislative redistricting plan diluted the black vote); Jamison
v. Tupelo, 471 F. Supp. 2d 706, 710 (S.D. Miss. 2007). Endogenous elections, or contests within the
jurisdiction and for the particular office that is at issue, are more probative than exogenous elections. See
Magnolia Bar Ass’n II, 994 F.2d at 1149 (giving greater weight to endogenous elections); Clark I, 21
F.3d at 97. Nevertheless, exogenous elections—i.e. elections in a district for positions that are not
exclusively representative of that district, Jamison, 471 F. Supp. 2d at 710—are probative on the
question of vote dilution and should be considered by the district court. Rodriguez, 385 F.3d at 861 n.5
(endorsing the analysis of exogenous elections in Section 2 vote dilution cases); see also Rangel, 8 F.3d
at 247 (5th Cir. 1993). That said, exogenous elections are meant to supplement, not replace, endogenous
elections. See Westwego I, 872 F.2d at 1209 n.11 (stating that “evidence derived from exogenous
elections . . . must be evaluated according to its probative value”).
Because the Section 2 analysis is concerned with more than the results of any individual election,
the results of a couple of elections do not discount the presence of racial bloc voting: “[a] pattern of
racial bloc voting that extends over a period of time is more probative of a claim that a district
experiences legally significant polarization than are the results of a single election.” Gingles, 478 U.S. at
95
57; Teague, 92 F.3d at 288-89 (noting that vote dilution is a determination that must be made over time).
Similarly, proof “that bloc voting is not absolute does not preclude a finding of racial polarization.”
Instead, to be legally significant, racial polarization need only be the “usual[]” pattern for a significant
proportion of voters. Gingles, 478 U.S. at 57; Campos, 894 F. Supp. at 1066 (in analyzing the third
Gingles factor, “the import of the word ‘usually’ cannot be underestimated.”).
Although statistical evidence is the primary means by which plaintiffs attempt to establish the
third Gingles factor, the Court has a duty to ponder all available evidence concerning racially polarized
voting that promises to cast light on the factors at work in a particular electoral scheme. Whitfield v.
Democratic Party of Ark., 890 F.2d 1423, 1428 (8th Cir. 1989); Askew v. City of Rome, 127 F.3d 1355,
1377 (11th Cir. 1997) (relying on both empirical and anecdotal evidence of racial bloc voting).
Anecdotal evidence relevant to the existence of racial bloc voting includes evidence of common political
interests or racial attitudes held by members of the relevant groups in the jurisdiction. United States v.
City of Euclid, 580 F. Supp. 2d 584, 600 (N.D. Ohio 2008).
1.
Plaintiff’s Burden: Plaintiff Must Demonstrate the Existence of Racial Bloc
Voting
In the Fifth Circuit, the plaintiff must first demonstrate the existence of racially divergent voting
patterns between Latinos and Anglos. Teague, 92 F. 3d at 290; Houston, 56 F.3d at 612. If the plaintiff
makes this threshold showing, the burden then shifts to the defendant to show that partisan affiliation,
rather than race, best explains the divergent voting patterns among Latinos and Anglos. LULAC IV, 999
F.2d at 850, 859-60; Teague, 92 F. 3d at 290. After the defendant demonstrates that partisanship, rather
than race, explains the difference in the divergent voting patterns of Anglos and Latinos, the burden then
shifts back to the plaintiff to negate the role of partisanship. See LULAC IV, 999 F.2d at 859-60.
96
a.
Statistical Evidence of Racial Bloc Voting
i.
Regression Analysis
Using ecological regression and ecological inference, Dr. Barreto examined the voting patterns
of Latinos and Anglos in twenty-six recent electoral contests to determine whether Latinos vote
cohesively and whether racial bloc voting exists in Harris County. Dr. Barreto’s statistical analysis
examined the voting behavior of Latino and Anglo voters in twenty-six county and statewide elections
that occurred between 2002 and 2010. See (Plaintiffs’ Ex. 39, at 14). Included in this collection of
elections are two endogenous elections for Precinct 2 County Commissioner and twenty-four exogenous
elections. To maximize the probative value of the exogenous races included in the data set, Dr. Barreto
only selected races that were sufficiently similar to the endogenous elections in terms of the candidates,
the level of office, and the partisan nature of the race. Accordingly, the majority of the exogenous
elections were countywide partisan general election races involving an Anglo candidate and a Latino
candidate; the remaining exogenous elections were statewide partisan general election races involving
an Anglo candidate and a Latino candidate. (Tr. 1:163).59 Dr. Barreto’s regression and ecological
inference results are summarized below:
Contest
2002 County
Commissioner
Precinct 2
RACIAL BLOC VOTING: REGRESSION ANALYSIS
Cand.
Prec.2 Prec.2 Differ. Harris Harris Differ. Prec. Harris
Cty.
Est.
Est.
Prec. 2 County County Harris
2
Anglo Latino
Est.
Est.
Est.
County
Vote
Vote
Anglo
Anglo Latino
Est.
Varia Varia
(x(y)
and
Vote
Vote
Anglo
nce
nce
2
int.)
(r2)
Latino
(x(y)
and
(r )
Vote
interce
Latino
pt)
Vote
ENDOGENOUS ELECTIONS
Sylvia
15.4% 94.4% 79.1%
N/A
N/A
N/A
.758
N/A
Garcia
59
The Court finds that the exogenous elections chosen are sufficiently similar in character to the endogenous
elections that the racial bloc voting results of the exogenous elections would be probative regarding the existence
of racial bloc voting in Precinct 2.
97
2010 County
Commissioner
Precinct 2
Sylvia
Garcia
N/A
.660
N/A
EXOGENOUS ELECTIONS
2002 Elections
10.5% 94.5% 84.0%
11.3%
83.1%
71.8%
.799
.769
9.6%
95.9%
86.3%
11.4%
83.6%
72.2%
.797
.763
8.9%
94.1%
85.2%
10.0%
82.3%
72.3%
.792
.773
77.4%
2006 Elections
24.6% -52.8% 76.9%
30.4%
-46.5%
.802
.779
19.7%
79.0%
59.3%
20.3%
72.5%
52.2%
.802
.786
75.6%
26.5%
-49.0%
76.8%
30.9%
-45.8%
.829
.791
12.8%
73.9%
61.2%
15.3%
65.9%
50.6%
.789
.793
Dion
Ramos
13.0%
2008 Elections
76.5% 63.5%
16.6%
74.2%
57.6%
.894
.870
Robert
Hinojosa
13.7%
77.0%
63.4%
17.7%
74.4%
56.6%
.889
.859
Richard
J. (Rick)
Noriega
15.0%
76.1%
61.1%
18.5%
72.9%
54.4%
.881
.859
2002
Judge
Silvia
County
Pubchara
Criminal
Court No. 2
2002
Judge Blanca E.
County
Lopez
Criminal
Court No. 5
2002 County Richard
Treasurer
Garcia
2006 County Orlando
Treasurer60
Sanchez
2006 Judge,
Silvia
County
Pubchara
Criminal
Court No. 2
2006 District Vanessa
Judge,
183 Velasquez
Judicial Dist.
2006
Maria L.
Lieutenant
Alvarado
Governor
2008 District
Judge,
55th
Judicial
District
2008 Family
District Judge,
312th Judicial
District
2008
State
Senator
11.7%
78.6%
66.9%
60
N/A
N/A
The data submitted in connection with this contest is inconsistent: coefficients from the regression (i.e. the
percentage Latino and the constant (i.e. the percentage white)) differed from the aggregated vote estimates (which
simply sum the coefficients and the constant to estimate what fraction of a designated racial group voted for the
Latino-surnamed candidate). See (Plaintiffs’ Ex. 39, at 38). Where there is such a conflict, the Court relies on the
coefficients derived from the regression, rather than the aggregated vote estimates.
98
Orlando
Sanchez
92.9%
2010 Elections
32.8% -60.0% 88.8%
35.7%
-53.1%
.892
.861
2010 Judge,
Mark
County
Diaz
Criminal
Court No. 11
2010 Judge,
Shelia
County
Acosta
Criminal
Court No. 761
2010 Judge,62
Alfred
County
“Bud”
Criminal
Valdez
Court No. 5
2010 Judge, Alfred G.
County
“Al” Leal
Criminal
Court No. 4
2010 Judge,
Mary
County
Connealy
Criminal
Acosta
Court No. 2
David
2010 Family
District Judge, Longoria
314th Judicial
Dist
2010 Family
Robert
District Judge, Hinojosa
312th Judicial
District
3.3%
75.0%
71.7%
9.0%
70.1%
61.1%
.884
.859
3.7%
75.2%
71.5%
9.2%
70.4%
61.2%
.885
.859
2.7%
74.0%
71.4%
8.3%
69.4%
61.0%
.887
.864
5.5%
74.5%
69.0%
12.1%
69.4%
57.2%
.871
.832
5.4%
74.0%
68.5%
11.4%
69.3%
57.8%
.871
.842
5.9%
75.7%
69.8%
12.9%
70.4%
57.5%
.868
.830
5.3%
75.0%
69.7%
12.4%
69.9%
57.4%
.867
.829
2010 District Vanessa
Judge,
183 Velasquez
Judicial Dist.
95.7%
28.8%
-66.9%
90.5%
32.7%
-57.8%
.891
.862
2010 County
Treasurer
61
The data submitted in connection with this contest is inconsistent: coefficients from the regression (i.e. the
percentage Latino and the constant (i.e. the percentage white)) differed the aggregated vote estimates (which
simply sum the coefficients and the constant to estimate what fraction of a designated racial group voted for the
Latino-surnamed candidate). See (Plaintiffs’ Ex. 39, at 90). Where there is such a conflict, the Court relies on the
coefficients derived from the regression, rather than the aggregated vote estimates.
62
The data submitted in connection with this contest is inconsistent: coefficients from the regression (i.e. the
percentage Latino and the constant (i.e. the percentage white)) differed from the aggregated vote estimates (which
simply sum the coefficients and the constant to estimate what fraction of a designated racial group voted for the
Latino-surnamed candidate). See (Plaintiffs’ Ex. 39, at 87). Where there is such a conflict, the Court relies on the
coefficients derived from the regression, rather than the aggregated vote estimates.
99
2010 Family
Dion
District Judge, Ramos
55th Judicial
Dist.
2010 Justice, Michael
1st Court of Gomez
Appeals,
Place 4
2010 Justice,
Eva
Supreme
Guzman
Court, Place 9
2010
Hector
Commissioner
Uribe
of the General
Land Office
Linda
2010
ChavezLieutenant
Thompson
Governor
4.8%
74.7%
69.9%
11.8%
69.4%
57.6%
.868
.833
3.4%
74.7%
71.3%
9.4%
69.9%
60.5%
.882
.857
90.9%
32.3%
-58.6%
87.4%
34.7%
-52.7%
.891
.863
2.6%
72.7%
70.1%
8.6%
67.9%
59.3%
.879
.858
4.3%
69.7%
65.4%
9.6%
65.7%
56.1%
.878
.859
The regression analysis of the two endogenous elections both show that Latinos overwhelmingly
supported Sylvia Garcia.63 In Ms. Garcia’s first bid for county commissioner, in 2002, 94.4% of Latinos
supported Ms. Garcia and in 2010, 78.6% of Latinos supported Ms. Garcia. Anglos, however, did not
support her candidacy to the same degree: in 2002, for instance, only 15.4% of Anglos supported Ms.
Garcia, and in 2010, as an incumbent, only 11.7% of Anglos supported her candidacy. In both election
cycles, there is an appreciable difference in Latino and Anglo support for Ms. Garcia: in 2002, there was
a 79.1% difference between the Latino and Anglo support for Ms. Garcia and in 2010, there was a
66.9% difference. The evidence further demonstrated that the variance of the regression model, i.e. the
r2 values, were .758, in the 2002 election and .660 in the 2010 election, meaning that the independent
variable (i.e. the percentage of Latino voters) explained a high proportion of the variance in the
dependent variable (i.e. the votes cast for the Latino-preferred candidate).
63
Since 2002, there have been three endogenous elections for Precinct 2 County Commissioner: the 2002, 2006,
and 2010 races. Plaintiffs, however, did not analyze the results of the 2006 race for Precinct 2 because
Commissioner Garcia ran unopposed in the 2006 race. Given that the incumbent ran unopposed, the Court finds
that the exclusion of the 2006 race was proper because in the 2006 race, voters did not have the opportunity to
oppose the Latino-preferred candidate, therefore the 2006 race will have little probative value on the question of
racial bloc voting. See Gingles, 486 U.S. at 54.
100
The regression results of the endogenous elections indicate that Anglos voted as a bloc to defeat
the Latino-preferred candidate. This evidence is further supported by the regression analysis of the
exogenous elections.
Of the twenty-four exogenous elections analyzed, the data showed that Anglos and Latinos
supported different candidates, such that the Latino-preferred candidate did not win a single election. On
the whole, Anglo support in Precinct 2 for the Latino-preferred candidate hovered between the low
single digits and the low teens, ranging from a low of 2.6%, in the 2010 Hector Uribe contest, to a high
of 19.7%, in the 2006 Silvia Pubchara contest. In contrast, Latino support in Precinct 2 for the Latinopreferred candidate ranged from a low of 69.7%, in the 2010 Linda Chavez-Thompson contest, to a high
of 95.9%, in the 2002 Blanca E. Lopez contest. The evidence further indicated that the r2 values of the
exogenous elections ranged from 0.789 to 0.894, indicating that the percentage of Latino voters
explained a high proportion of the variance in the votes cast for the Latino-preferred candidate, lending
credence to the predicative value of the regression model.
However, it should be noted that in five of the exogenous elections examined, the 2006 and 2010
contests of Vanessa Velasquez, the 2010 contest of Eva Guzman, and the 2006 and 2010 contests of
Orlando Sanchez, the Spanish-surnamed candidate captured a substantial share of the Anglo vote,
resulting in electoral victory. Plaintiffs, however, maintain that these results do not counsel against a
finding of racial bloc voting. According to Dr. Barreto, in these contests, the Spanish-surnamed
candidate was not the Latino-preferred candidate and therefore their ability to attract support from Anglo
voters does not refute the existence of racial bloc voting. In the 2006 Sanchez contest, only 24.6% of
Latinos voters supported Mr. Sanchez, whereas 77.4% of Anglo voters supported his candidacy; in the
2006 Velasquez contest, only 26.5% of Precinct 2’s Latino voters supported Ms. Velasquez, whereas
75.6% of the precinct’s Anglo voters supported Ms. Velasquez; in the 2010 Velasquez contest, only
28.8% of Precinct 2’s Latino voters supported Ms. Velasquez, whereas 95.7% of the precinct’s Anglo
voters supported Ms. Velasquez; in the 2010 Guzman contest, 32.3% of Precinct 2’s Latino voters
101
supported Ms. Guzman, whereas 90.9% of Precinct 2’s Anglo voters supported Guzman; and, finally, in
the 2010 Sanchez contests, only 32.8% of Latinos voters supported Mr. Sanchez, whereas 92.9% of
Anglo voters supported his candidacy. The statistical evidence thus demonstrates that these three
candidates were not able to capture a majority of the Latino vote; for each of these candidates Latinos
coalesced around the Democratic Anglo candidate, rather than the Latino Republican candidate.
Therefore, the success of Latino candidates who are not the Latino-preferred candidates does not prevent
Plaintiffs from establishing threshold evidence of racial bloc voting, for this evidence still confirms that
Anglos and Latinos are voting for different candidates.
Dr. Barreto testified that the statistical evidence ultimately demonstrates that the Anglo majority
votes as a bloc to defeat the Latino-preferred candidate, limiting the opportunity for Latino influence.
(Tr. 1:114). The degree of polarization was significant: in some examples, the voting was almost entirely
polarized, with 95% of Latino voters supporting a candidate, but only 10% of Anglo voters supporting
that same candidate. Dr. Barreto further testified that in the 2010 elections, the racial polarization had
reached a fever pitch, with above average turnout by Anglos and low turnout by Latinos enabling Anglo
voters to dictate the outcome of the election in districts like Precinct 2, (Tr. 1:117), engendering defeat
for Latino-preferred candidates like then-Commissioner Garcia. (Tr. 1:118).
On review, the Court finds that Plaintiffs have presented regression analysis that indicates the
presence of racial bloc voting. In the twenty-six elections examined, Anglos and Latinos supported
different candidates. The Latino-preferred candidate was defeated in every single election except one,
Commissioner Garcia’s 2002 inaugural run for county commissioner; in the other twenty-five elections,
the Anglo-preferred candidate prevailed. See (Tr. 1: 143). Accordingly, the regression evidence
presented shows: (i) a clear and consistent relationship between the race of the voters and their candidate
preference; and (ii) that Anglos are usually able to defeat the Latino-preferred candidate.
102
ii.
Contest
2002 County
Commissioner
Precinct 2
2010 County
Commissioner
Precinct 2
2002
Judge
County
Criminal
Court No. 2
2002
Judge
County
Criminal
Court No. 5
2002 County
Treasurer
2006 County
Treasurer
2006 Judge,
County
Criminal
Court No. 2
2006 District
Judge,
183
Judicial Dist.
2006
Lieutenant
Governor
2008 District
Judge,
55th
Judicial
District
2008 Family
District Judge,
312th Judicial
District
Ecological Inference Analysis
RACIAL BLOC VOTING: ECOLOGICAL INFERENCE ANALYSIS
Cand.
Range Level of
Most Likely
Range Level
Most Likely
Support by
Estimate of
of Support by Estimate of
Latinos
Support by
non-Latinos
Support by
Latinos
non-Latinos
ENDOGENOUS ELECTIONS
Sylvia Garcia
39.9%-97.0%
87.6%
24.4% -64.3%
31.0%
Sylvia Garcia
Silvia
Pubchara
46.7% - 89.2%
80.6%
EXOGENOUS ELECTIONS
2002 Elections
25.7%-96.8%
83.6%
17.0%-60.2%
25.7%
31.3%-56.3%
35.9%
Blanca E.
Lopez
26.0-96.8%
83.9%
31.4%-56.3%
35.9%
Richard
Garcia
24.8%- 96.2%
82.5%
30.0%-55.2%
34.9%
2006 Elections
4.6% - 74.4%
19.7%
36.1% – 65.1%
58.8%
26.6%- 95.8%
81.7%
33.9% - 62.6%
39.8%
Vanessa
Velasquez
4.7% - 74.7%
19.6%
36.0% - 65.1%
58.9%
Maria L.
Alvarado
20.5%-92.4%
73.9%
27.2%-57.1%
34.9%
2008 Elections
32.7% - 94.3%
81.1%
33.4% - 62.5%
39.6%
33.0% - 94.5%
34.1% - 63.1%
40.3%
Orlando
Sanchez
Silvia
Pubchara
Dion Ramos
Robert
Hinojosa
81.3%
103
2008
State
Senator
2010 County
Treasurer
2010 Judge,
County
Criminal
Court No. 11
2010 Judge,
County
Criminal
Court No. 764
2010 Judge,65
County
Criminal
Court No. 5
2010 Judge,
County
Criminal
Court No. 4
2010 Judge,
County
Criminal
Court No. 2
2010 Family
District Judge,
314th Judicial
Dist
2010 Family
District Judge,
312th Judicial
District
2010 District
Judge,
183
Judicial Dist.
Richard J.
(Rick)
Noriega
31.6% – 94.3%
79.9%
33.6% - 63.2%
40.4%
2010 Elections
9.9% - 73.4%
27.8%
43.2% - 72.9%
64.5%
Orlando
Sanchez
Mark Diaz
29.2% – 91.7%
76.4%
27.3% - 56.5%
34.5%
Shelia Acosta
29.5% - 91.9%
76.7%
27.6% - 56.8%
34.7%
Alfred “Bud”
Valdez
28.6% - 91.2%
75.2%
26.7%-55.9%
34.1%
Alfred G.
“Al” Leal
29.5% -92.1%
76.8%
29.1%- 58.3%
36.2%
Mary
Connealy
Acosta
29.2- 91.9%
76.3%
28.6 – 57.9%
35.9%
David
Longoria
30.1%- 92.6%
77.7%
29.7% -59.0%
36.7%
Robert
Hinojosa
29.8%-92.3%
76.9%
29.3%-58.5%
36.5%
Vanessa
Velasquez
9.2%-72.1%
25.8%
43.6%-73.1%
65.3%
64
The data submitted in connection with this contest is inconsistent: coefficients from the regression (i.e. the
percentage Latino and the constant (i.e. the percentage white)) the aggregated vote estimates (which simply sum
the coefficients and the constant to estimate what fraction of a designated racial group voted for the Latinosurnamed candidate). See (Plaintiffs’ Ex. 39, at 90). Where there is such a conflict, the Court relies on the
coefficients derived from the regression, rather than the aggregated vote estimates.
65
The data submitted in connection with this contest is inconsistent: coefficients from the regression (i.e. the
percentage Latino and the constant (i.e. the percentage white)) the aggregated vote estimates (which simply sum
the coefficients and the constant to estimate what fraction of a designated racial group voted for the Latinosurnamed candidate). See (Plaintiffs’ Ex. 39, at 87). Where there is such a conflict, the Court relies on the
coefficients derived from the regression, rather than the aggregated vote estimates.
104
2010 Family
District Judge,
312th Judicial
Dist.
2010 Justice,
1st Court of
Appeals,
Place 4
2010 Justice,
Supreme
Court, Place 9
2010
Commissioner
of the General
Land Office
2010
Lieutenant
Governor
Dion Ramos
29.4%-92.0%
76.4%
28.8%-58.0%
36.0%
Michael
Gomez
29.1% -91.7%
76.1%
27.5%- 56.7%
34.8%
Eva Guzman
9.0% -72.7%
25.8%
41.9% -71.7%
63.9%
Hector Uribe
27.8% - 90.6%
73.8%
26.3% -55.7%
34.2%
Linda
ChavezThompson
26.5% - 90.0%
72.0%
26.2% -55.9%
34.6%
The ecological inference findings confirm the results of the ecological regression. For the
endogenous elections, the ecological inference results indicated that Ms. Garcia most likely received
87.6% of Precinct 2’s Latino vote in her 2002 bid for County Commissioner and most likely received
only 31.0% of the non-Latino vote in that 2002 election. In 2010, however, the ecological inference
results indicate that Ms. Garcia most likely received 80.6% of Precinct 2’s Latino vote but only received
25.7% of the non-Latino vote.
It is important to note here that Plaintiffs’ ecological inference has only examined the trends of
Latinos and non-Latinos rather than Latinos and Anglos. The non-Latino variable includes the AfricanAmerican, Anglo, and Asian-American voters of Precinct 2, but is primarily composed of AfricanAmericans and Anglos. According to Dr. Barreto, the ecological inference results for the non-Latino
metric approximate the voting behavior of Anglos. In fact, Dr. Barreto testified that, generally, the nonLatino results overestimate the Anglo support for the Latino-preferred candidate because the non-Latino
designation includes African-Americans, who traditionally support the Latino-preferred candidate,
thereby increasing the share of the non-Latino electorate that supports the Latino-preferred candidate.
105
For each of the twenty-six elections, Dr. Barreto constructed ecological inference density plots that
graph the density of non-Latino voter support for the Latino preferred candidate; the density plot graphs
the percentage of non-Latino voters who voted for the Latino-surnamed candidate against the density of
non-Latinos across the precinct.
For the twenty-one elections where the Latino-surnamed candidate was the Latino preferred
candidate, the density plots of the non-Latino vote feature a small peak occurring where the percentage
of non-Latino voters who supported the Latino-surnamed candidate approaches 100%. This small peak
indicates that one faction of the non-Latino amalgam supports the Latino-preferred candidate at near
100% levels of support. There is a second large peak where non-Latino support for the Latino-surnamed
candidate is below 50%, which indicates that another faction of the non-Latino composite does not
support the Spanish-surnamed candidate. According to Dr. Barreto, the small peak corresponds to the
small African-American faction of the non-Latino voters, who offer near-universal support of the
Spanish-surnamed candidate, and the second larger peak corresponds to the larger Anglo faction of the
non-Latino voter group, who tend not to support the Latino-preferred candidate.
In the five elections where the Anglo majority supported the Spanish-surnamed candidate, the
peaks are inverted. The small peak always occurs below the 50% mark—often occurring near the 0%
mark—and the large peak occurs closer to the marker indicating 100% support. For these five elections,
the small and large peaks still correspond to African-American and Anglo voters, respectively, but this
time, the African-American voters are universally voting against the Spanish-surnamed candidate and
the Anglo voters are supporting the Spanish-surnamed candidate.
The Court finds the ecological inference data imprecise. Plaintiffs’ failure to isolate the voting
trends of Anglos from African-Americans deprives the data of optimal utility, but the data is
nevertheless probative on the question of racial bloc voting. The testimonial evidence and the density
graphs conclusively establish that the inclusion of African-American voters in the non-Latino voter
group generally, save for the five elections discussed above, increased the non-Latino support for the
106
Latino-preferred candidate. And even with these inflated numbers, the evidence consistently showed that
non-Latinos in Precinct 2 (Anglos and African-Americans combined, but not independently) do not
support the Latino-preferred candidate to the same degree as Precinct 2 Latinos. Thus, even if the
ecological inference data is not entirely precise, it is nevertheless probative. The Court finds that the
ecological inference corroborates the ecological regression data’s findings that Anglos do not support
the Latino-preferred candidate.
b.
Anecdotal Evidence of Racial Bloc Voting
Though statistical analysis is the starting point for an assessment of racial bloc voting, anecdotal
evidence may also be relevant to the existence of racial bloc voting. Euclid, 580 F. Supp. 2d at 600.
In this case, several of the fact and expert witnesses offered anecdotal evidence of racial bloc
voting. For instance, Senator Garcia, relying on her own experience as a political candidate, testified that
voting in Harris County is racially polarized, such that Anglos often vote in a bloc to defeat the Latinopreferred candidate. (Tr. 3:32). Senator Garcia explained that being Latina gave her an advantage with
Latino voters—whom she described as having a natural tendency to vote for a candidate they know is
Latino or who possesses a Spanish surname—but also made her an unpalatable candidate to some Anglo
voters. (Tr. 3:33). Thus, for instance, when she was running for City Controller, she had to expend an
extraordinary effort reaching out to certain sectors of the Anglo community to try to persuade them to
consider her as a candidate. (Tr. 3:32). Although these efforts were effective with some sectors of the
Anglo community, there were other pockets of the Anglo community that were impenetrable. Senator
Garcia felt that this latter group was simply resistant to the idea of a Latino candidate and would not,
under any circumstances, vote for her because she was Latina. (Tr. 3:32-33).
According to Garcia, other Latino candidates in Harris County have also experienced this
resistance from certain Anglo voters. (Tr. 3:33). For instance, in the recent case of Judge Medina, a
Latino Republican, his challenger, Judge Devine, publicly admitted that he decided to contest incumbent
Judge Medina in the Republican primary because Devine believed that Judge Medina would have a
107
difficult time emerging from the primary given his Spanish surname. (Tr. 3:33). Ultimately, Justice
Devine defeated Medina in the primary. (Tr. 3:33). The electoral travails of Orlando Sanchez offers
another example. When Mr. Sanchez first ran for County Treasurer he was unable to overcome some of
the racial barriers of the electorate, resulting in his defeat. (Tr. 3:34). It was only after the Republican
leadership made sure that he had a clear field in the primary that he was able to, in his second campaign
for County Treasurer, eventually secure the Republican nomination. (Tr. 3:34).
This sense of resistance was echoed by Councilman Gonzalez, who indicated that while
campaigning on behalf of another Latino candidate, a citizen told him, “I’m not going to vote for that
wetback . . . and you tell them to go back to Mexico.” (Tr. 2:123). Councilman Gonzalez indicated that
during the course of his career he has heard similar things said about him and of other Latino candidates.
(Tr. 2:123).
David Walden,66 Commissioner Morman’s chief of staff, explained that in a general election,
Republican candidates in Harris County are generally not the preferred candidates of most Latino
citizens but are instead the preferred candidate of most of the County’s Anglo voters. (Tr. 3:131).
According to Mr. Walden, this general proposition also holds true for the commissioner’s court races;
Mr. Walden explained that in the commissioner’s court race for Precinct 2, the majority of Latinos
supported the Democratic nominee, whereas a majority of the Anglos voted for the Republican nominee.
(Tr. 3:132). In fact, Mr. Walden testified that, in 2010, when Mr. Morman was elected to the
Commissioner’s Court, he was not the Latino-preferred candidate for the seat. (Tr. 3:130-31).
Mr. Walden testified that bringing more Anglos into Precinct 2 will, as a general matter, increase
the support for the non-Latino preferred candidate. (Tr. 3:132). And in a closely held precinct, like
66
Mr. David Walden has been involved in Harris County politics for over thirty-five years. In the past, he has
served as a liaison for Judge Jon Lindsay and chief of staff to former Houston Mayor Bob Lanier, among others.
(Tr. 3:153). Currently, Mr. Walden is the chief of staff for Commissioner Morman. (Tr. 3:127). As chief of staff,
he is responsible for the Commissioner’s scheduling, the Commissioner’s court agenda, and the day-to-day
operation of the Commissioner’s office. Through his experience in Harris County politics he is familiar with the
voting behavior of the residents in Precinct 2. (Tr. 3:130).
108
Precinct 2 was during the 2010 election, taking Latinos out of that district will make it more difficult for
Latinos to elect the candidate of their choice. (Tr. 3:133). Mr. Walden further testified that, by the same
token, bringing Anglos into a closely held precinct, such as Precinct 2 in the 2010 election, will also
make it more difficult for Latinos to elect the candidate of their choice. (Tr. 3:134).
This testimony was echoed by Plaintiffs’ and Defendants’ experts. Plaintiffs’ expert Mr.
Birnberg testified that Latinos are politically cohesive, but Harris County’s Anglo population tends to
support the candidate that is opposing the Latino-preferred candidate, resulting in the defeat of the
Latino-preferred candidate. (Tr. 3:113-14). Mr. Birnberg—in keeping with the testimony of Mr.
Walden—opined that a redistricting plan that brings more Anglos into Precinct 2 makes it more difficult
for Latinos to elect their candidate of choice. In his opinion, the Latino-preferred candidate cannot win
an election under the Revised A-1 map. (Tr. 3:119). Defendants’ expert Dr. Murray noted in his expert
report that “majority Anglo areas do not usually support the candidate preferred by Hispanic voters.”
(Defendants’ Ex. 14, at 4). In fact, Dr. Murray notes that the dual forces of low voter turnout by Latino
voters coupled with the strong voter participation of Anglo voters who “do not usually support” the
Latino-preferred candidate make it very difficult for Latinos in Precinct 2 to elect their candidates of
choice. (Defendants’ Ex. 14, at 4).
c.
Summary
On the whole, Plaintiffs’ statistical and anecdotal evidence demonstrates the existence of racial
bloc voting racial polarization. Plaintiffs’ regression and ecological inference data show a consistent and
sizeable divergence in the voting patterns of Latinos and Anglos. In keeping with these statistical
findings, the fact and expert witnesses also testified that Latinos and Anglos generally prefer different
candidates and this divergence in voting preference usually results in the defeat of the Latino-preferred
candidate. Together, the statistical and anecdotal evidence demonstrates the existence of racial bloc
voting.
109
2.
Defendants’ Rebuttal
Once plaintiffs have offered facially plausible statistical evidence of racial polarization, the
burden then shifts to defendants to offer proof that the divergent voting patterns of the Anglos and
Latinos can be explained by some non-racial factor, such as partisanship. See Teague, 92 F.3d at 290;
see also Armstrong v. Allain, 893 F. Supp. 1320, 1330-31 (S.D. Miss. 1994).
In this case, Defendants contend that Dr. Barreto’s analysis attributes the differences in voting
preference to race when those differences should really be attributed to partisanship. Dr. Alford
explained that most of the elections included in Dr. Barreto’s statistical analysis featured an Anglo
Republican candidate running against a Latino Democratic candidate. Thus, the consistent relationship
between Anglo voters’ rejection of the Latino Democratic candidate appeared to be a function of race.
See (Tr. 4:61-64). However, Dr. Alford contends that when one analyzes the Anglo voting trends in
general election races featuring an Anglo Democrat and an Anglo Republican, it becomes clear that
Anglo voters are rejecting Democrats as a whole, not merely Latino Democrats. In an effort to prove
this, Dr. Alford first reviewed the Anglo voting trends in the contests selected by Dr. Barreto (i.e., those
contests between an Anglo Republican and a Latino Democrat) to determine the percentage of Anglos
supporting the Republican candidate when the Democratic candidate was Latino. He then reviewed the
Anglo voting trends in contests between an Anglo Republican and an Anglo Democrat to determine the
percentage of Anglos supporting the Republican candidate when both candidates were Anglo. Finally,
he compared the percentage of Anglos supporting the Republican candidate when the Republican
candidate was Anglo and when the Republican candidate was Latino to determine whether the race of
the candidate impacted the amount of support that candidate would receive from Anglo voters.
In contests between an Anglo Republican and a Latino Democrat, Dr. Alford found that 87.6%
of Anglo voters supported the Anglo Republican candidate. The results of his analysis are detailed in the
chart below.
110
ESTIMATED ANGLO SUPPORT IN GENERAL ELECTION CONTESTS FROM 2002-2010
WHERE A LATINO DEMOCRAT CANDIDATE RUNS AGAINST AN ANGLO REPUBLICAN
CANDIDATE67
Estimated
% of Anglo
Year
Election Contest
Democratic Latino
Vote for
Candidate
Repub.
Candidate
2002
Judge, County Criminal Court No. 2
Silvia Pubchara
88.7%
2002
Judge, County Criminal Court No. 5
Blanca E. Lopez
88.6%
2002
County Treasurer
Richard Garcia
90.0%
2002
Justice, Supreme Court, Place 1
Linda Yanez
91.6%
2006
2006
Lieutenant Governor
Judge, County Criminal Court No. 2
Maria Luisa Alvarado
Silvia Pubchara
84.7%
79.7%
2008
2008
2008
U.S. Senator
Family District Judge, 312th Judicial District
District Judge, 55th Judicial District
Richard J. Noriega
Robert Hinojosa
Dion Ramos
81.5%
82.3%
83.4%
2010
Lieutenant Governor
Linda Chavez-Thompson
2010
Commissioner of the General Land Office
Hector Uribe
st
2010
Justice, 1 Court of Appeals District, Place 4 Michael Gomez
2010
District Judge, 55th Judicial District
Dion Ramos
rd
2010
District Judge, 183 Judicial District
Michael Gomez
2010
Family District Judge, 312th Judicial District
Robert Hinojosa
2010
Family District Judge, 314th Judicial District
David Longoria
2010
Judge, County Criminal Court No. 2
Mary Acosta
2010
Judge, County Criminal Court No. 5
Alfred Valdez
2010
Judge, County Criminal Court No. 7
Sheila Acosta
2010
Judge, County Criminal Court No. 4
Alfred G. Leal
2010
Judge, County Criminal Court No. 11
Mark Diaz
AVERAGE % OF ANGLO VOTE FOR THE REPUBLICAN CANDIDATE
90.4%
91.4%
90.6%
88.2%
90.5%
87.6%
87.1%
88.6%
87.9%
87.9%
87.9%
91.0%
87.6%
Similarly, in contests between an Anglo Republican and an Anglo Democrat, 87.6% of Anglo
voters supported the Anglo Republican candidate. These results are detailed in the chart below.
67
Source: Defendants’ Ex. 68, at 1. Dr. Alford’s description of the elections contests, included in his report,
conflicts with the descriptions included the County’s cumulative report of election results, included in
Defendants’ Exhibit 51. Where there is a conflict, the Court has adopted the description employed in Defendants’
Ex. 51, the Cumulative Report of Election Results.
111
ESTIMATED ANGLO SUPPORT IN GENERAL ELECTION CONTESTS FROM 2002-2010
WHERE AN ANGLO DEMOCRATIC CANDIDATE RUNS AGAINST AN ANGLO
REPUBLICAN CANDIDATE68
Estimated
% of Anglo
Year
Election Contest
Democratic Anglo
Vote for
Candidate
the Rep.
Candidate
2002
Commissioner of Agriculture
Tom Ramsay
92.6%
2002
Judge, Court of Criminal Appeals, Place 1
Pat Montgomery
91.6%
2002
Chief Justice, Supreme Court
Richard Baker
92.1%
2002
Controller of Public Accounts
Marty Atkins
95.7%
2002
Commissioner of the General Land Office
David Bernsen
89.6%
2002
Lieutenant Governor
John Sharp
82.1%
2002
Member, State Board of Education, District 7
Richard Hargrove
84.4%
2002
Justice, Supreme Court, Place 4
Margaret Mirabal
82.4%
2002 AVERAGE
88.8%
2004
2004
President and Vice President
Justice, Supreme Court, Place 9
John Kerry/ John Edwards
David Van Os
2004 AVERAGE
86.2%
87.6%
86.9%
2006
2006
2006
2006
2006
2006
2006
2006
Attorney General
Commissioner of Agriculture
Controller of Public Accounts
Governor
Commissioner of the General Land Office
Railroad Commissioner
United States Senator
Justice, Supreme Court, Place 2
David Van Os
Hank Gilbert
Fred Head
Chris Bell
VaLinda Hathcox
Dale Henry
Barbara Ann Radnofsky
William E. “Bill” Moody
2006 AVERAGE
86.8%
84.1%
86.5%
81.9%
85.0%
82.5%
85.3%
79.9%
84.0%
2008
2008
Judge, Court of Criminal Appeals, Place 3
Member, State Board of Education District 7
Susan Strawn
Laura Ewing
2008 AVERAGE
83.9%
82.3%
83.1%
2010
2010
2010
2010
2010
2010
Attorney General
Commissioner of Agriculture
Judge, Court of Criminal Appeals, Place 6
Governor
Justice, Supreme Court, Place 3
Justice, Supreme Court, Place 5
Barbara Ann Radnofsky
Hank Gilbert
Keith Hampton
Bill White
Jim Sharp
Bill Moody
2010 AVERAGE
96.0%
95.3%
95.3%
82.0%
94.2%
92.2%
92.5%
AVERAGE FOR ALL ELECTIONS
68
87.6%
Source: Defendants’ Ex. 68, at 2. In Defendants’ Exhibit, Dr. Alford uses a shorthand to identify the election
contests. The Court has cross-referenced this short-hand with the contests listed in Defendants’ Ex. 51 to populate
the election contest with the descriptions included in Defendants’ Ex. 51.
112
Dr. Alford’s analysis thus indicated that 87.6% of Anglo voters supported the Anglo Republican
candidate when the Anglo Republican candidate was facing a Latino Democratic candidate and 87.6%
of Anglo voters supported the Anglo Republican candidate when the Anglo Republican candidate was
facing an Anglo Democratic opponent. (Tr. 4:63). In Dr. Alford’s opinion, this evidence proves that
Anglo voters are rejecting Democrats, regardless of whether that Democrat is an Anglo or a Latino. (Tr.
4:63, 69). In Dr. Alford’s view, the driving force in partisan contests in the general election is
partisanship rather than race or ethnicity. (Tr. 4:67, 68-69).
This suggestion that partisan affiliation dictates electoral outcomes was mentioned in other
witness testimony. First, Dr. Murray testified that, in the general election, members of the same political
party who are of different races or ethnic groups win together and lose together such that there is no
appreciable difference in the victory count of Democratic candidates based on race. (Tr. 4:43). Second,
Dr. Barreto testified that there were instances in which a minority candidate has run as a Republican and
Republican Anglo voters supported the minority candidate. (Tr. 1:149). On cross-examination, Dr.
Barreto could not identify a general election race where a Republican Latino ran but was unable to
capture the support of the Anglo community. (Tr. 1:150). Third, when describing her 1996 defeat in the
race for County Attorney, Senator Garcia explained that 1996 “was one of the years we all experience
here in Harris County where the presidential candidate did not [carry the County, so] there was no coat
tails and many of our Democratic judges and others lost during that time, including myself.” (Tr. 2:8).
Although Senator Garcia was describing the dynamics of the 1996 election, which is somewhat
temporally remote, the sentiment of the statement is very much in keeping with Defendants’ theory that
political affiliation, rather than race, best explains the inability of Latino-preferred candidates to prevail
in the general election.
On review, the Court finds this evidence raises a weak inference that a non-racial factor explains
the divergence in the voting patterns of Anglos and Latinos. “A person’s politics is rarely as readily
discernible—and never as permanently discernible—as a person’s race. Political affiliation is not an
113
immutable characteristic, but may shift from one election to the next; and even within a given election,
not all voters follow the party line.” Vieth, 541 U.S. at 287.69 Because the racial bloc voting inquiry
examines whether Anglo and Latino voting patterns diverge over a period of time, see Gingles, 478 U.S.
at 57; Teague, 92 F.3d at 288-89, where Defendants contend that partisanship rather than race best
explains racially divergent voting patterns the Court needs some indication that the partisanship effect is
an enduring. In this case, Defendants’ evidence suggests, but does not conclusively establish, that over
the course of several elections, Anglos and Latinos have been voting based on partisan affiliation. And
this partisanship explains the divergence in their voting patterns.
Dr. Alford’s analysis demonstrates that Anglo voters support Anglo Republicans when the Anglo
Republican is running against a Latino Democrat and also support Anglo Republicans when the Anglo
Republican is running against an Anglo Democrat. While it is true that the analysis showed that Anglo
support for the Anglo Republican was identical regardless of the race of the Democratic challenger, in
the end Dr. Alford’s analysis merely proves that Anglo voters consistently support Anglo Republicans.
Although Anglo voters’ rejection of the Democratic candidate, regardless of that candidate’s race, could
indicate the primacy of partisanship over race, Anglo voters’ consistent support for the Anglo
Republican, regardless of the race of the Republican’s opponent, does not necessarily indicate the
dominance of partisanship. That said, the fact that Anglo support for the Anglo Republican is the same
no matter the race of the Democratic challenger does seem to favor Defendants’ view. Moreover, the
testimony indicated that in Harris County, Democrats rise and fall together. Together, the analysis and
testimony suggest that the divergence in the voting patterns could be explained by partisanship.
69
See also Meghan Ashford-Grooms, Rick Perry Says He Switched to GOP at a Younger Age Than Reagan Did,
Austin Am. Statesman, Sept. 6, 2011, available at http://www.politifact.com/texas/statements/2011/sep/06/rickperry/rick-perry-says-he-switched-gop-younger-age-reagan/
114
3.
Plaintiff’s Final Burden
If the defendant can offer evidence that some non-racial factor accounts for the racially polarized
voting pattern, the plaintiff must come forth with evidence establishing the primacy of race in explaining
the divergent voting preferences of Anglos and Latinos. See LULAC IV, 999 F.2d at 850-51.
Dr. Barreto testified that in the course of his research on partisanship and ethnicity, he found that
it was nearly impossible to disentangle partisanship from race. Dr. Barreto explained that, in Texas, the
history of partisanship is directly and intimately related to race and racial attitudes, such that it is
impossible to distill voter behavior related to race from voter behavior related to partisanship. According
to Dr. Barreto, race cannot be divorced from partisanship because the greatest predictor of Republican
party affiliation for Anglo voters in the southern United States is the voter’s racial attitudes. After
controlling for a voter’s stated preference on issue positions, an Anglo voter’s political affiliation is
determined by the racial attitudes the voter possesses. (Tr. 1:102). Dr. Barreto testified that Republicans
are perceived as being more racially conservative on civil rights, immigration, and other racial issues.
(Tr. 1:105). Democrats, on the other hand, have historically been viewed as being more tolerant of civil
rights, immigration reform, and other issues of racial equality. (Tr. 1:104-05). According to Dr. Barreto,
the association of racially conservative positions with the Republican Party and racially tolerant
positions with the Democratic Party dates back to McGovern’s Southern Strategy. Prior to President
Lyndon B. Johnson’s Great Society and the racially and socially progressive pieces of legislation
stemming from that era, the southern states were Democratic strongholds. (Tr. 1:106-07). After
President Johnson’s passage of civil rights legislation such as the Voting Rights Act, the Democratic
Party, though once viewed as hospitable to the needs of Southern white voters, was perceived as
advancing minority interests. This resulted in a dramatic realignment of Anglo southern voters, whereby
Southern white voters abandoned the Democratic Party and shifted their allegiance to the Republican
Party. Although the origin of this political realignment harkens back to the late 1960s and early 1970s,
115
this dynamic has become a fixture of the modern political landscape so that today much of the South,
including Texas, remains a Republican stronghold.
With this context, Dr. Barreto contends the partisan affiliation of voters is deeply tied to their
racial attitudes, such that one cannot distinguish partisan affiliation from race. According to Dr. Barreto,
these attitudes are reflected in the results of his electoral analysis, which indicates that Spanishsurnamed candidates were often defeated.70 Dr. Barreto further explained that the five elections where a
Republican Spanish-surnamed candidate prevailed in the general election do not undermine the
significance of the relationship between race and partisan affiliation. According to Dr. Barreto, minority
Republican candidates who actively speak out against topics such as immigration, racial injustice, and
affirmative action, in step with the Republican platform, are able to neutralize race in gaining white
support. (Tr. 1:103, 105).71 On the one hand, this, of course, undermines the very proposition the
professor seeks to establish. If Anglo Republicans are willing to vote for minority candidates who adopt
the tenets of Republican ideology and take strides to remind voters of their steadfast and earnest
commitment to those precepts, then it would appear that there is some point at which Anglo voters’
political inclinations outweigh their racial attitudes. Dr. Barreto, however, suggested that this
interpretation misses the mark. Although Anglo voters’ willingness to cast a vote for a minority
Republican candidate appears to indicate the triumph of political affiliation over racial attitudes, Dr.
Barreto explained that in truth Anglo Republican voters view these minority candidates through a racial
lens. (Tr. 1:105). First, the minority candidates who espouse these views are not perceived as being
members of the racial group with whom they share phenotypic similarities. Second, even if the Anglo
Republican voters are voting for a person of Latino descent, they are still not voting for the Latino-
70
Compare (Plaintiffs’ Ex. 39), detailing the voting trends of the electorate in the twenty-six analyzed races, with
(Defendants’ Ex. 51), which details the electoral results of the twenty-six races.
71
Interestingly, Senator Cruz recently gave on interview on July 4, 2013 supporting the very proposition that the
evidence established at trial. See, e.g., David Welina, Ted Cruz and His Texas Electorate at Odds on Immigration,
Nat’l Pub. Radio, July 4, 2013, available at http://www.npr.org/2013/07/04/198781839/ted-cruz-and-his-texaselectorate-at-odds-on-immigration.
116
preferred candidate and therefore they are still voting adversely to the interests of the Latino community.
See (Tr. 1:162). In the five elections where a Latino candidate ran as a Republican (i.e., the two elections
involving Valasquez, the two elections involving Sanchez, and the election with Eva Guzman), the
Latino candidate was not the Latino-preferred candidate. (Tr. 1:162).72 Dr. Barreto thus concluded that
race and partisanship were permanently intertwined such that one cannot distinguish the effects of one
from the other and the divergent voting patterns of Anglos and Latinos were a function of race and
politics.
Mr. Birnberg also rejected the idea that the Anglo population’s tendency to support the candidate
opposing the Latino-preferred candidate was merely a function of politics free from any racial
overtones. See (Tr. 3:114). However, unlike Dr. Barreto, Mr. Birnberg testified that race could be
isolated from partisanship by examining the voting behavior of Anglos in primary elections, which
control for the political party of the candidates. According to Mr. Birnberg, “if a Latino can never get
elected in a Republican primary and Anglos always get elected in the Republican primary in Harris
County,” that is a strong indication that partisan choice does not explain the inability of white voters to
support the Latino-preferred candidate, but is more consistent with racial block voting. See (Tr. 3:116).
Mr. Birnberg testified that since 2004, only one Latino, Senator Rafael Edward (“Ted”) Cruz,73
has prevailed in a contested Republican primary against an Anglo candidate in Harris County. (Tr.
3:114, 115). According to Mr. Birnberg, Anglo voters will vote for an Anglo Republican before they
72
The mere fact that Anglo Republican voters have supported a Latino Republican candidate in a general election
does not, without more, counsel against a finding of racial polarization. Again, the Voting Rights Act is not
concerned with the ethnic origins of the elected representatives who serve the citizenry. Although it is often the
case that the citizenry prefers the candidate who shares their ethnic origin, (Tr. 1:163), ethnic consonance between
candidate and voter is not the test by which courts determine if voters have the ability to choose that candidate as
their preferred candidate. To find that Anglo Republican voter support of the Latino Republican counsels against
a finding of polarization changes the focus of the racial bloc voting inquiry from whether Anglos defeat the
Latino-preferred candidate to whether Anglo voters defeat the candidate who happens to be Latino. But again, the
Latino-preferred candidate and the Latino candidate are not synonymous. Although the Latino-preferred candidate
will often be a Latino candidate, conflating the two designations is improper and is anathema to our voting rights
jurisprudence because the case law commands that political cohesion be proven. To assume that the Latino
candidate is the Latino-preferred candidate violates the most basic tenets of Gingles.
73
This occurred in the 2012 Republican primary for U.S. Senate.
117
vote for a Latino Republican, even if the Latino Republican is already an elected official, and in some
cases, the preference for an Anglo Republican candidate will trump the incumbency advantage of the
Latino candidate. (Tr. 3:114). In the 2010 race for Harris County Tax Assessor-Collector, for example,
Leo Valasquez lost the Republican primary even though at the time of the primary race, Mr. Valasquez
was the incumbent Harris County Tax Assessor-Collector. (Tr. 3:114). Similarly, former Justice David
Medina was targeted in the Republican primary for a seat on the Texas Supreme Court because he was a
Latino candidate; the challenger thought that Justice Medina was vulnerable in the Republican primary,
even though Justice Medina enjoyed incumbency status, because Justice Medina was Latino. (Tr.
3:114). Mr. Birnberg stated that these examples show that when Harris County Republicans are given a
choice between an incumbent office-holding Latino Republican and an Anglo Republican candidate, the
electorate chooses the Anglo candidate. (Tr. 3:115).
The one recent exception to this general trend is the election of Senator Cruz. In the May 2012
Republican primary for U.S. Senate, Senator Cruz failed to capture support from the Anglo Republican
electorate, with his challenger Lieutenant Governor David Dewhurst capturing the lion share of the
support from Harris County Republicans. (Tr. 3:115). Nevertheless, Lieutenant Governor David
Dewhurst only won a plurality of the ballots cast; he did not win the majority of votes, which, of course,
resulted in a run-off election between Mr. Dewhurst and the second-place finisher, Ted Cruz. In the July
2012 runoff, Ted Cruz enjoyed a sizeable victory over his opponent. According to Mr. Birnberg, this
electoral victory does not blunt his observation that Anglo Republicans prefer Anglo candidates over
Latino candidates because, in his view, Mr. Cruz’ election was an anomaly; the run-off election occurred
in late July and his electoral victory was engineered by a small number of highly motivated voters, who,
though numerically small, still made up a significant share of the small number of the electorate that
participated in the July run-off. (Tr. 3:115).
Plaintiffs have produced countervailing evidence that race is a consideration for voters in
deciding the candidate for whom they will cast a ballot. According to Mr. Birnberg, the Latino
118
community does not merely vote straight party ticket, instead they are bloc voting Latino surnamed
candidates. See (Tr. 3:118). Moreover, when Anglos have the opportunity to choose between a Latino
Republican candidate or an Anglo Republican candidate, they tend to choose the Anglo Republican
candidate. Thus, the evidence shows that race is playing a factor in the decisions of both Anglos and
Latinos in their selection of candidates. Accordingly, the Court finds that Plaintiffs have established the
third Gingles factor.
D.
TOTALITY OF THE CIRCUMSTANCES
Even if plaintiffs establish the three Gingles preconditions, they cannot succeed on their Section
2 claims unless they establish that under the totality of the circumstances Latinos do not have the same
“opportunity [as] other members of the electorate to participate in the political process and elect
representatives of their choice.” 42 U.S.C. 1973(b). In conducting the totality of circumstances analysis,
courts must consider: (1) the history of voting-related discrimination in the state or political subdivision;
(2) the extent to which voting in the state or subdivision is racially polarized; (3) the extent to which the
state or subdivision has used voting practices or procedures that tend to enhance opportunities for
discrimination against the minority group; (4) whether minority candidates have been denied access to
any candidate-slating process; (5) the extent to which minorities have borne the effects of past
discrimination in relation to education, employment, and health; (6) whether local political campaigns
have used overt or subtle racial appeals; (7) the extent to which minority group members have been
elected to public office in the jurisdiction; (8) whether there is a significant lack of responsiveness on the
part of elected officials to the particularized needs of members of the minority group; and (9) whether
the policy underlying the use of voting qualifications is tenuous. See Gingles 478 U.S. at 36-37. The
court must also consider whether the number of districts in which the minority group forms an effective
majority is roughly proportional to the minority group’s share of the population in the relevant area. See
Perry, 548 U.S. at 425; Fairley, 584 F.3d at 673.
119
The plaintiffs do not need to prove a particular number of the above-listed factors or prove that a
majority of them point one way or the other. Rather, “the final determination of whether the voting
strength of minority voters is canceled out demands the court’s overall judgment, based on the totality of
the circumstances and guided by those relevant factors in the particular case.” Whitfield v. Democratic
Party, 890 F.2d 1423, 1432 (8th Cir. 1989) (internal quotations and modifications omitted).
A defendant may try to rebut the plaintiffs’ claim of vote dilution via evidence of “objective,
nonracial factors under the totality of the circumstances standard.” Teague, 92 F.3d at 283. Ultimately,
“the final determination of whether the voting strength of minority voters is canceled out presents very
complex political and legal issues whose resolution requires a comprehensive canvassing of the relevant
facts,” De Grandy, 512 U.S. at 1011, and “an intensely local appraisal of the design and impact of the
contested electoral mechanisms.” Gingles, 478 U.S. at 79; LULAC, 548 U.S. at 426. A totality-of-thecircumstances approach means that there is no entitlement to a particular outcome if one proves specific
facts. There are only shadings and nuance, which different people estimate differently. Litigants are
entitled to an honest application of the legal framework, not to a particular result. Teague , 92 F.2d at
292.
1.
History of Official Discrimination
The first factor listed in the Senate Report is “the extent of any history of official discrimination
in the state or political subdivision that touched the right of the members of the minority group to
register, to vote, or otherwise to participate in the democratic process.” S. Rep., supra, at 28, 1982 U.S.
Code Cong. & Admin. News, at 206 (emphasis added). Thus, Congress was concerned not only with
present discrimination, but with the vestiges of discrimination which may interact with present political
structures to perpetuate a historical lack of access to the political system. Gingles, 478 U.S. at 69.
Evidence that is relevant under this factor includes the existence of racially segregated schools and
public facilities. LULAC III, 986 F.2d at 747 (internal citations omitted).
120
Mr. Birnberg and Senator Garcia explained that, historically, Latinos have been discriminated
against in education. See (Tr. 3:85) [Birnberg]; (Tr. 3:9) [Garcia]. When Mr. Birnberg first came to
Harris County in the 1970s, Houston Independent School District (“HISD”) was largely segregated on
the basis of race and ethnicity. (Tr. 3:86). Segregation was so entrenched in the tradition of HISD that
the school district attempted to circumvent the spirit of Brown’s instruction to desegregate by
reclassifying Latinos as Anglos such that schools were “desegregated” by integrating Blacks with
Latinos. (Tr. 3:86).74 During this period, there was significant disparity between the educational
experiences of Anglo students and Latino students; Latinos were receiving a decidedly inferior
educational experience characterized by inadequate facilities, outdated curricula, and limited educational
enrichment opportunities. (Tr. 3:87) [Birnberg]; (Tr. 3:11) [Garcia].75 The effects of these educational
disparities can still be observed today; at present, there are over a hundred thousand limited English
proficiency Latino adults, many of whom are the product of this discriminatory educational system. (Tr.
3:87).
The Court finds that Plaintiffs have offered evidence that Harris County has a history of
discriminating against Latino citizens.
2.
Racial Polarization
The second factor under the totality of the circumstances analysis considers “the extent to which
voting in the elections of the state or political subdivision is racially polarized.” S. Rep. at 29. This
factor is the “linchpin of a section 2 vote dilution claim.” Gretna, 834 F.2d at 499; see also Westwego
III, 946 F.2d at 1120. When the pattern of racially polarized voting is severe—that is, when minority and
white voters rarely engage in cross-over voting—an election scheme is more likely to dilute minority
voting strength in violation of Section 2. LULAC III, 986 F.2d at 748-49.
74
Senator Garcia testified that Latinos objected to this practice and boycotted the County schools, leading to the
development of huelga schools. (Tr. 3:10).
75
In fact, HISD remained under a court-ordered desegregation plan until 1983. See generally Ross v. Houston
Independent School Dist., 699 F. 2d 218 (5th Cir. 1983).
121
The statistical evidence presented at trial indicated a high degree of racial polarization. Dr.
Barreto’s election analysis demonstrated that Anglos and Latinos consistently supported different
candidates. In the general election, Latinos supported Latino Democratic candidates at very high rates.
In fact, Latino support for the Democratic candidate ranged from a low of 69.7%, in the 2010 Linda
Chavez-Thompson contest for Lieutenant Governor to a high of 95.9% for Blanca E. Lopez in the 2002
race for county criminal court judge. Anglo support for Latino-preferred candidates paled in
comparison: Anglo support for the Latino-preferred candidate ranged from a low of 2.7% in the 2010
Hector Uribe contest to a high of 19.7% in the 2006 Silvia Pubchara contest for county criminal court. In
every election analyzed, the Latino-preferred candidate was incapable of attracting even twenty percent
of the Anglo vote. This divergence in voting usually resulted in the defeat of the Latino candidate. In
fact, of the elections analyzed by Dr. Barreto, the Latino-preferred candidate only prevailed once: in the
2002 Precinct 2 County Commissioner race. In that race, Sylvia Garcia garnered an estimated 94.4% of
the Latino support but only 15.4% of the Anglo vote, which again suggests a high incidence of racial
polarization.
This degree of racial polarization is even observable in the few contests where Anglos do support
Spanish-surnamed candidates. In these contests, Anglos support Latino candidates that have been
rejected by Latino voters. In the elections analyzed by Dr. Barreto, there were five elections in which the
Spanish-surnamed candidate was not the Latino-preferred candidate but was instead the Anglo-preferred
candidate.76 In these elections, Anglo support for the Latino candidate ranged from a low of 75.6% in
the 2006 Vanessa Velasquez race for district judge to a high of 95.6% in the 2010 Vanessa Velasquez
race for district judge.77 This remarkable support from Anglo voters dwarfed the meager support these
candidates received from Latino voters. Latino support for these candidates ranged from a low of 24.6%
76
The five elections were the 2006 election of Vanessa Velasquez, the 2006 election of Orlando Sanchez, the
2010 election of Eva Guzman, the 2010 election of Vanessa Velasquez, and the 2010 election of Orlando
Sanchez. See (Plaintiffs’ Ex. 39).
77
In contrast, only 26.5% of Latino voters supported Ms. Velasquez in her 2006 election bid and only 28.8% of
Latinos supported her in her 2010 bid.
122
in the 2006 Orlando Sanchez race for County Treasurer to 32.8% in the 2010 Orlando Sanchez race for
County Treasurer.78 In each of the five elections, the difference between the Anglo support for the
Latino candidate and the Latino support for that candidate exceeded forty-five percentage points, and in
the case of Orlando Sanchez’s 2010 election, the differential exceeded sixty percentage points.
These statistical results are in keeping with the documentary evidence from Dr. Murray, which
indicated that although there are pockets of Anglo communities in Harris County that have sizeable
concentrations of cross-over voters willing to politically align themselves with Latinos, such as Houston
Heights and the area bordering Rice University, by and large Anglo communities in Harris County
oppose the Latino-preferred candidate. On the whole, the Court finds that Plaintiffs have produced
significant evidence that voting in Harris County is very racially polarized.
3.
Existence of Any Electoral Mechanisms that Enhance the Opportunity for
Minority Vote Dilution
The third factor asks courts to consider other voting practices that may interact with the
challenged election scheme to dilute minority voting strength. LULAC III, 986 F.2d at 749.
Plaintiffs’ witnesses testified that Harris County, the state of Texas, and private citizen groups
have adopted many policies that adversely impact Harris County’s Latino community. See (Tr. 3:112).
a.
County Practices
As for the County, Senator Garcia, Mr. Birnberg and Dr. Barreto testified that Harris County has
erected barriers to the voter registration that have adversely impacted the County’s Latino community.
Senator Garcia testified that sometime during the first half of her first term as commissioner, the
County was sending mass mailings to voters in Precinct 2 threatening to purge them from the voter rolls.
(Tr. 3:28). In fact, during her second year as commissioner, she received a letter threatening to purge her
from the voter rolls because she did not live at her residence. (Tr. 3:28). The next day, she confronted
78
In contrast, 77.4% of Anglo voters supported Mr. Sanchez in his 2006 election bid for office and only 92.9% of
Anglos voters supported him in his 2010 bid for treasurer.
123
the tax assessor, Paul Bettencourt, about the practice of sending these letters. (Tr. 3:29). She asked him
for a list of how many people were getting these letters. The list, once compiled, was detailed by
precinct; the list indicated that Precinct 2 received the largest proportion of purging letters compared to
all the other precincts. (Tr. 3:29). This purging campaign has a chilling effect on voter participation in
that most people who receive these letters will assume that they are ineligible to vote in the upcoming
election. (Tr. 3:29). Senator Garcia, of course, was not dissuaded by the purging letter, but as she aptly
noted, she is not the average voter; she was aware of her voting and registration status and had direct
access to the County tax assessor. See (Tr. 3:30). She estimates that Precinct 2 lost a lot of voters in
connection with this purging campaign. (Tr. 3:30).
Mr. Birnberg testified that in 2008, a number of Latino residents reported that the Harris County
Tax Assessor’s Office, which is responsible for voter registration, was not timely processing the voter
registration applications of Latino citizens. Dr. Murray’s expert report indicated that “[t]he Harris
County Tax Assessor-Collector and Registrar of Voters, Paul Bettencourt, used his position in 2008 to
slow the dramatic rise in voter registrations that year among younger, mostly minority, applicants.”
(Plaintiffs’ Ex. 59, at 17). He explained that Bettencourt’s office had a “practice of rejecting thousands
of applications because of minor errors that were not disqualifying in other large counties.” (Plaintiffs’
Ex. 59, at 17). According to Dr. Murray’s report,
Bettencourt’s office, claiming to be over-whelmed by the surge in applications in the fall,
also created a massive backlog of unprocessed voter registration cards that prevented
voters who had legally registered from casting regular ballots in 2008. That failure
resulted in thousands of provisional ballots being cast in 2008 by persons who had filled
out registration cards before the early October deadline, but whose names were not on the
rolls when they showed up to vote several weeks later. These actions resulted in the
Texas Democratic Party suing Mr. Bettencourt in federal court, alleging, among other
charges, that he had violated the Voting Rights Act. With depositions scheduled before a
January court date, Mr. Bettencourt abruptly resigned in early December 2008, despite
having just been reelected to a four year term a month earlier.
(Plaintiffs’ Ex. 59, at 17); see (Tr. 3:99). As a result of the suit, the County agreed to modify its voter
registration process to correct these issues. (Tr. 3:99); see also (Plaintiffs’ Ex. 16).
124
However, Dr. Barreto’s expert report suggests that voter registration problems persisted into
2010. Dr. Barreto contends that, in 2010, Harris County continued to wrongfully deny thousands of
Latino voter registration applications. In 2010, 175,729 people submitted voter registration applications
in Harris County. Of this total, 31,190 (or 17.7%) were submitted by people with Spanish surnames,
22,047 were submitted by people living in majority-black zip codes, and 122,492 applications were
submitted by people who do not possess Spanish-surnames and who do not live in majority black zip
codes.79 Of the 31,190 applications submitted by people with a Spanish-surname, 4,199 applications
were not accepted;80 thus, Latino residents of Harris County had a 13.5% rejection rate. See (Plaintiffs’
Ex. 39, at 8). Of the 122,492 submitted by Anglo voters, only 12,435 applications were not accepted;
thus, Anglos had a 10.2% rejection rate. See (Plaintiffs’ Ex. 39, at 8). Dr. Murray contends that the
difference in the rejection rates is practically insignificant. (Defendants’ Ex. 14, at 2). According to Dr.
Murray, if the rejection rate for Latinos and non-Latinos were the same, that would have only resulted in
an additional 924 applications being accepted. This increase “would have made virtually no difference
in the overall registration patterns in the county.” (Defendants’ Ex. 14, at 2).81 In response, Dr. Barreto
contends that while the three percent difference between the rejection rates of Latinos and Anglos, may
not be significant in any one year, if the 3% difference occurs every year over the course of the decade,
then the cumulative effect of the difference is substantial.
79
The first group, comprised of Spanish-surname applicants, is intended to approximate the number of Latino
applicants. The second group, comprised of people living in zip codes where at least 70% of the residents identify
as African-American, is designed to serve as a proxy for black voters. Finally, the third group, comprised of nonSpanish surnamed individuals who do not live in majority-black zip codes, is supposed to approximate the
number of Anglo voter applications.
80
Dr. Barreto further explained that many of the rejected applications were submitted by Latinos residing in
Precinct 2. See (Tr. 1:123); see (Plaintiffs’ Ex. 39, at 9).
81
Dr. Alford also suggested that Dr. Barreto is creating a problem when none exists. According to Dr. Alford,
Harris County’s voter registration rates, even among Latino citizens, is comparable to and in some cases exceeds
the registration rates of Latinos in many other counties in Texas. Dr. Alford thus contends that Harris County does
not have a Latino voter registration problem. See (Defendants’ Ex. 49, at 5). The Court is not persuaded; the mere
fact that Harris County’s Latino voter registration and rejection rates are commensurate with or exceed the
successful registration rates of other Texas counties does not negate the possibility that there is a problem in
Harris County; taking Defendants’ expert’s evidence at face value, Defendants’ evidence only suggests that the
problem in Harris County, to the extent that there is a problem, is not as bad as it is in other counties.
125
In addition to the processing issues, Councilman Rodriguez, Mr. Birnberg and Senator Garcia
testified that Latinos’ ability to exercise their right to vote is hindered by the County’s failure to provide
multilingual voting materials, as required by law. (Tr. 2:123, 3:20, 3:100). She testified that although the
County has improved its record of providing multi-lingual materials over the last fifteen to twenty years,
the County could do more to ensure its residents have access to multi-lingual voter information. (Tr.
3:20-21). Ms. Garcia further explained that access to multilingual voting materials substantially
increases voter participation among Harris County’s Latino population. (Tr. 3:20).
In addition to barriers to processing and provision of materials, Mr. Birnberg explained that the
County is also seeking to change the mechanics of the deputy registrar process such that going forward
there will be fewer avenues by which to register new voters. The deputy registrar program is a means by
which citizens can register to vote. Historically, the deputy registrars would travel into communities and
encourage citizens to complete voter registration forms. Within five days of receiving a completed
registration, the registrars would then submit the completed registration to one of fifteen locations
designated to receive these registrations. (Tr. 3:101). Recently, the Tax Assessor’s Office has reduced
the number of locations at which completed registration forms may be submitted from fifteen to two.
(Tr. 3:101). Moreover, the Tax Assessor’s Office has changed the operation of the registrar program
such that today, a person seeking to serve as a deputy registrar must first complete a County-sponsored
training. (Tr. 3:101). According to Mr. Birnberg, the County made access to the training difficult for
people who lived in districts represented by African-American or Latino state representatives but made
the training conveniently located for people who lived in districts represented by Anglo representatives.
He explained that the County hosted trainings in each state representative district represented by an
Anglo state representative but did not host any training sessions in districts represented by AfricanAmerican or Latino representatives. (Tr. 3:101-102).
Senator Garcia further testified that Latinos encounter resistance from poll workers. (Tr. 3:23).
Throughout her career in politics, she has observed consistent problems with poll workers and access to
126
the polls more generally. (Tr. 3:23). Over the course of her career in politics, voters have told her that
they were told that they: (i) cannot bring someone in to assist them; (ii) are not registered to vote; and
(iii) cannot vote in the election because they do not possess a valid driver’s license. (Tr. 3:24).
Occasionally, she also receives complaints that poll watchers have intimidated Latino voters at the polls,
such as inquiring about the voters’ citizenship status. (Tr. 3:24).
Senator Garcia also testified that Latinos’ ability to exercise their political choice has been
frustrated by their limited access to polling places. (Tr. 3:25). Senator Garcia explained that often
polling places are not located in Latino-populated communities; therefore, Latino voters have to travel a
modest distance to reach their nearest polling station. Second, occasionally, polling places are not
accessible via public transportation, which limits the ability of some voters to get to the polls. Third,
polling places have limited hours that do not accommodate the work schedule of many of the County’s
Latino residents; in her view, extending the hours of operation of the polling locations and/or allowing
early voting would increase the rate of participation amongst Latinos. (Tr. 3: 27).
b.
State Practices
As for the State of Texas, Mr. Birnberg opined that the state’s inability to achieve voting equality
is evidenced by the fact that since 1971 every Congressional, Texas House, and Texas Senate
redistricting that has been adopted by the Texas Legislature has been modified by the courts. (Tr. 3:104).
Mr. Birnberg further testified that more recent examples of the state’s erection of barriers to minority
franchise include: (i) the state’s photo identification law that requires voters to present a state-issued
form of identification in order to vote; (ii) the state’s proposal to limit the categories of person who may
serve as deputy voter registrars to Texas residents; and (iii) the state’s decision prohibiting groups from
paying deputy voter registrars a fee that is based on the number of voter registrations the registrar
facilitates. (Tr. 3:105-112).
Mr. Birnberg testified that Texas has recently passed a photo identification law that requires
voters to present a state-issued form of personal identification, in addition to their voter registration
127
certificate, to the polling station before they will be permitted to cast their ballots. (Tr. 3:107-08).
Ironically, immediately after the Supreme Court’s decision in Shelby County v. Holder, 570 U.S. __ 133
S.Ct. 2612 (2013), striking down section 4(b) of the Voting Rights Act and effectively suspending the
enforcement of Section 5 of the Act, Governor Rick Perry and Texas Attorney General Greg Abbott
indicated that they would begin enforcing the photo identification law. See Michael Cooper, After
Ruling, States Rush to Enact Voting Laws, N.Y. Times, July 6, 2013, at A9; Todd J. Gillman, Texas AG
Greg Abbott: Voter ID law “Will Take Effect Immediately,” Dallas Morning News, June 25, 2013,
available at http://www.dallasnews.com/news/20130625-texas-ag-greg-abbott-voter-id-law-will-takeeffect-immediately.ece.82
In addition to the identification law, the state has also recently modified its deputy registrar
program by limiting the persons eligible to serve as deputy registrars and the persons from whom a
deputy registrar may accept a registration. Under the recent changes, only Texas residents may serve as
deputy registrars. The new changes also prohibit deputy registrars from accepting registrations from
citizens who live outside the County in which the registrar is canvassing. Thus, for example, a deputy
registrar canvassing downtown Houston could not accept a registration from someone who worked
downtown but lived in Sugarland, which is in Fort Bend County rather than Harris County. (Tr. 3: 109110). According to Mr. Birnberg, limiting the deputy registrar position to Texas residents will make it
very difficult to conduct large-scale voter registration drives using deputy registrars. (Tr. 3:109-10).
c.
The Citizenry
In Dr. Murray’s report, he explained that “the rise of the Tea Party in Texas helped bring back
some of the discriminatory electoral strategies from earlier years aimed at minority voters in Texas.
Leading the charge in Houston was the King Street Patriots, a Tea Party affiliate that employed “ballot
82
Section 5 requires certain states and their political subdivisions to obtain federal preclearance before
implementing any changes to their voting laws or practices, and Section 4(b) details the coverage formula that
determines which states and/or political subdivisions of the state are subject to preclearance.
128
security” programs aimed at suppressing black and Hispanic voting in October and November 2010.
(Plaintiffs’ Ex. 59, at 17).
[The King Street Patriots] mobilized several hundred ‘patriots’ to monitor early and inperson voting at Harris County polling places, to make sure that the election was not
being stolen. Of course, all the stealing was taking place in inner city minority
neighborhoods, so that is where the all-white, older patriotic brigades needed to be
deployed. The racist element of these efforts was evident in a photo posted in October on
the King Street Patriots website showing a black man holding a sign with the words ‘I
only got to vote once,’ with a white woman behind him holding a second sign saying
‘I’m with stupid.’ The patriots did not turn up any specific examples of vote fraud, but
they did create a number of confrontations with voters and election officials at minority
polling places across the county.
(Plaintiffs’ Ex. 59, at 17); see also (Tr. 4:7).83 Although the King Street Patriots efforts have gained
notoriety, other smaller groups have implemented similar efforts. For instance, Mr. Birnberg testified
that one group encouraged its member to go to Latino polling places dressed up like immigration
officers. There have also been reports of groups distributing fliers in Latino communities warning
residents that police officers would be patrolling the polling places and would arrest anyone who had an
outstanding parking or traffic ticket. (Tr. 3: 90-91). In the report Dr. Murray explains that
underlying the efforts of elected Republicans such as Paul Bettencourt and activists like
the King Street Patriots is the reality that Texas Republicans wrote the African American
vote off in 1964 with the Goldwater-Johnson election. And with the rise of the extreme
right in the 21st century the state’s majority party seems hell-bent on driving Hispanics,
who have traditionally given about a third of their votes to GOP candidates, in the same
direction. If you are not going to compete for the votes of folks who make up half the
total population of Texas, 45 percent of the voting age population, and a third of the total
registered vote, it makes very good sense to both keep down the size of the minority vote
in general elections, and, and we shall see, to make sure the impact of those minorities
who do vote is limited by how representative districts are drawn.
(Plaintiffs’ Ex. 59, at 17).
The Court finds that Plaintiffs have produced evidence that voting practices exist that enhance
the possibility that the County’s map has a dilutive effect.
83
In 2010, Mr. Birnberg contacted the Harris County Attorney’s Office about the King Street Patriots. The
County, in turn, contacted the Justice Department to assist the County in addressing the activities of the King
Street Patriots. (Tr. 3:120).
129
4.
Existence of a Candidate Slating Process
The fourth factor asks courts to consider whether members of the minority group have been
denied access to any candidate slating process. Slating has been defined as “the creation of a package or
slate of candidates, before filing for office, by an organization with sufficient strength to make the
election merely a stamp of approval of the pre-ordained candidate group.” Overton, 871 F.2d at 534.
Where a slating organization exists, “the ability of minorities to participate in that slating organization
and to receive its endorsement may be of paramount importance.” LULAC III, 986 F.2d at 750 (internal
quotation marks and citations omitted). At trial, there was no evidence offered as to the existence of a
candidate slating process.
5.
Socio-Economic Disparity
Often, “political participation by minorities tends to be depressed where minority group
members suffer effects of prior discrimination such as inferior education, poor employment
opportunities, and low incomes.” Gingles, 478 U.S. at 69. Therefore, evidence of socioeconomic
inequalities and depressed political participation “may be probative of unequal opportunity to participate
in the political process and to elect representatives.” Westwego I, 872 F.2d at 1212. Under the fifth
factor, the Court considers “the extent to which members of the minority group in the state or political
subdivision bear the effects of discrimination in such areas as education, employment and health.” Here,
Plaintiffs must prove both depressed political activity and socioeconomic inequality, but they need not
to prove any causal nexus between their disparate socioeconomic status and the depressed level of
political participation. LULAC III, 986 F.2d at 750.
Mr. Korbel testified that in Harris County there is a strong and consistent correlation between
socio-economic welfare and race, such that Latinos are more likely to be economically disadvantaged
than their Anglo peers. Using data from the 2006-2010 ACS, Mr. Korbel found that Anglos have a mean
per capita income of $45,278, which is almost three times the $14,511 mean per capita income for
Latinos. (Plaintiffs’ Ex. 49, at 6). Moreover, median income for Anglo households is more than twice
130
that of Latino households, with median income of Anglos totaling $75,124, compared to the $38,916
median household income of Latinos. See (Plaintiffs’ Ex. 54, at 3); (Tr. 2:16). The ACS data also
indicated that Latinos have a higher incidence of poverty than do Anglos. According to the ACS, 9.2%
of Anglos earn less than 150% below the poverty level, but 34.5% of Latinos earn less than 150% of the
poverty level. See (Plaintiffs’ Ex. 49, at 5); (Tr. 2:8).84
In addition to income disparities, the evidence showed that there are significant educational
disparities between Latinos and Anglos. For instance, Latinos in Harris County are substantially more
likely to be functionally illiterate than their Anglo peers: the ACS indicated that 28.3% of Latinos over
the age of 25 had completed nine or fewer years of education—which, according to Mr. Korbel, is a
proxy for illiteracy—whereas only 1.8% of Anglos over the age of 25 had completed nine or fewer
years. See (Tr. 2:16); (Plaintiffs’ Ex. 54, at 5). Similarly, 43.4% of Anglos over the age of 25 possess a
college degree whereas only 10.8% of Latinos over the age of 25 possess a college degree. (Plaintiffs’
Ex. 54, at 6).
Senator Garcia testified that Latinos were also disadvantaged in terms of housing and
employment. With respect to housing, Senator Garcia testified that there were disparities between the
housing opportunities available to Latinos and those available to non-Latinos. In her view, housing
subsidies are not routed to organizations that serve Latino communities and therefore Latino populations
do not receive these funds, despite their demonstrated need. (Tr. 3:14). Without these subsidies, the
habitable housing options available to lower-income Latinos are limited. (Tr. 3:14-15). With respect to
employment, Senator Garcia testified that in both the public and private employment sector, Latinos are
under-represented in management or executive level positions. (Tr. 3:12).
84
Mr. Korbel explained that poverty should be measured in reference to persons living below 150% of the poverty
line rather than persons living below the poverty line, because the persons living below the poverty line metric
measures those people who are absolutely destitute, living in abject poverty, but excludes those who are
extremely disadvantaged, but not quite destitute. (Tr. 2:8). Accordingly, Mr. Korbel suggests that the persons
living below the poverty line metric underestimates the number of economically disadvantaged persons in the
County.
131
Senator Garcia explained that these disparities in education, employment, and housing adversely
impact Latinos’ ability to participate in the political process. (Tr. 3:17). First, those who want to
participate lack the tools to participate: they do not have access to public transportation options enabling
them to engage in the political process, nor do they possess an education that has equipped them with
the literary skills necessary to engage in the political process. (Tr. 3:18).
Under the segregated system of HISD, there were thousands of children who were not properly
taught English. Today, those same children are adults with limited English proficient adults. (Tr. 3:87).
These limited English proficient adults do not possess the language skills that would facilitate
participation in the democratic process. (Tr. 3:88). Mr. Birnberg testified that in his political experience,
a person’s educational attainment affects their involvement in campaigns or affects their ability to cast
an effective vote. (Tr. 3:87). According to Mr. Birnberg, voting is addictive behavior; once a citizen
begins to participate in the voting process, they continue to vote in subsequent elections. By the same
token, vote avoidance (or failure to vote) is also addictive behavior in that citizens without a history of
voter participation cannot be brought into the voting process. Accordingly, if a citizen’s educational
experience leaves him less able to become involved in the political process earlier and therefore he did
not start voting when he was young, then he is probably still not voting. (Tr. 3:88).
Second, many economically disadvantaged people lose hope that their condition will improve,
eventually becoming despondent and disaffected, believing that government does not and will not
respond to their needs. (Tr. 3:18). This despondency can result in voter apathy. One indicia of this voter
apathy appears in the statistics measuring voter turnout. According to Dr. Murray’s analysis, “voting
turnout is lowest in heavily Hispanic areas.” (Defendants’ Ex. 15, at 2).
In the end, the evidence shows that Latinos are socio-economically disadvantaged and that
Latinos voters have low rates of voter participation.
132
6.
Racial Appeals
The sixth factor asks courts to consider whether political campaigns have been characterized by
overt or subtle racial appeals. The presence of this factor can provide evidence that racial politics
dominate the political process. LULAC III, 986 F.2d at 750.
Mr. Gonzalez testified that political campaigns occasionally feature racial appeals. In one
campaign,85 his Anglo opponent disseminated a mailer featuring two streets. The first street was a wellkept corridor, presumptively located in Harris County, and the second street featured a dirt-road, with
buildings whose architecture, designs, and language written across the buildings suggested that the
buildings were in Mexico. In Mr. Gonzalez’s view, the mailer suggested that under the Anglo candidate
the neighborhood would continue to thrive, but under the Latino candidate the area would regress. (Tr.
2:125).
This testimony was echoed by Mr. Birnberg. (Tr. 3:112). Mr. Birnberg testified that, in 2008,
when Joel Redmond was running for the state legislature from Pasadena, his opponent circulated a
racially charged political advertisement that featured a set of black crows perched atop a series of
telephone lines with the words “Joel Redmond wants you to meet some of his closest friends” written in
blood alongside pictures of Mario Gallegos, then-Senator Barack Obama, and Representative Sheila
Jackson Lee. The tag line read, “Birds of Feather Flock Together.” (Tr. 3:113). According to Mr.
Birnberg, the advertisement was a patent racial appeal that was intended to remind voters that if they
vote for Joel Redmond, who is Anglo, they are voting for someone who socializes and communes with
Latinos and African-Americans. (Tr. 3:113). Mr. Birnberg also noted that Terry McConn had employed
a similar advertisement a few years earlier in his campaign against M.J. Khan for City Council, except
that in place of Barack Obama, the advertisement depicted Craig Washington. (Tr. 3:113).
The Court finds that Plaintiffs have produced evidence that political campaigns are
characterized by racial appeals.
85
Mr. Gonzalez could not remember the campaign in which this incident occurred.
133
7.
Extent of Latino Electoral Success
The seventh factor examines the extent to which members of the minority group have been
elected to public office in the jurisdiction. This factor, along with the extent of racially polarized voting,
is one of the most important factors that the district court must consider in its analysis of the totality of
the circumstances. See Gingles, 478 U.S. at 48 n.15; LULAC III, 986 F.2d at 750-51; Westwego III, 946
F.2d at 1120.
a.
History of Electability
Mr. Birnberg testified that, historically, political representatives in Harris County were elected
under multi member districts. (Tr. 3:89). Under the multi member districts of yesteryear, Latinos had
great difficulty winning city or county-wide races. (Tr. 3:89). In fact, Lionel Castillo’s 1972 election as
Controller was the first time a Latino was elected to city-wide office. (Tr. 3:89). The next Latinos
elected to citywide office were Gracie Saenz, in 1992, and Sylvia Garcia, in 1998. (Tr. 3:89). As for
county-wide races, Senator Garcia testified that the only Latinos to be elected county-wide were
Republican Orlando Sanchez (County Treasurer), Democrat Adrian Garcia (County Sheriff) and a few
county-wide judicial races. See (Tr. 3:35, 36).
However, Dr. Murray claimed that, in recent elections, Latino candidates have enjoyed a good
measure of success. (Tr. 3:247). In 2012, there were five Latino candidates, who were also Latinopreferred candidates running for county-wide office. Four of the five candidates won. The electability of
Latinos is further evidenced by the fact that Latino sheriff candidate Adrian Garcia captured a higher
share of the electorate than any other candidate—Anglo, Black, or Latino—in both 2008 and 2012. (Tr.
3:247). In light of these trends, Dr. Murray suggests that the electorate is amenable to Latino candidates,
and therefore the electability of Latino candidates is increasingly a function of voter turnout. Despite this
relationship, voter turnout in Latino communities is significantly lower than voter turnout in AfricanAmerican or Anglo communities, inhibiting the electability of Latino candidates. See (Defendants’ Ex.
134
14, at 3). In fact, over the last five election cycles, the turnout in predominately Anglo districts is almost
three times the turnout in predominately Latino districts, as indicated in the table below.86
According to Dr. Murray, when this turnout pattern is coupled “with the reality that majority
Anglo areas do not usually support the candidates preferred by Hispanic voters, the bottom line is . . .
there is not a sufficient concentration of [] registered Spanish surname voters to create [an effective
opportunity district]” that could preserve the ability of Latinos to elect the candidate of their choice.
86
Source: Defendants’ Ex. 14, at 4.
135
(Defendants’ Ex. 14, at 4). According to Dr. Murray, the 2011 redistricting of the City of Houston
demonstrates the degree to which the low Latino voter turnout inhibits Latinos’ ability to elect their
preferred candidates. Under the 2011 redistricting, the City gained two new council seats; one of the
seats, District J, would serve as an additional Latino opportunity district. (Defendants’ Ex. 14, at 6-7).
Although Latinos made up 59.3% of the voting age population of District J, when the November 2011
election was held there, a previously-defeated Anglo council candidate, Mike Laster, easily defeated his
opponents Criselda Romero and Rodrigo Canedo; Laster captured 67.3% of the vote whereas Romero
was only capable of securing 21.7% of the vote and only 11.0% of voters supported Canedo.
(Defendants’ Ex. 14, at 7). According to Dr. Murray, this demonstrates that low voter turnout inhibits
Latinos from electing their candidates of choice.
Dr. Barreto rejects Dr. Murray’s suggestion that the inability of the newly created District J to
effectively perform as a Latino opportunity district proves that Precinct 2 cannot perform as an effective
Latino opportunity district. (Tr. 1:134). First, he contends that District J’s inability to perform as an
effective Latino district is not probative of Precinct 2’s ability to perform as an effective Latino
opportunity district because: (i) city council races occur in odd-number years, not even-number years, as
the county commissioner races do; (ii) the turnout in the off-year city council races is very low, limiting
the value of any deductions and extrapolations drawn from these elections; and (iii) District J
corresponds to a relatively small area of the County and therefore any trends that occur in these races
may be unique to this area and are not necessarily indicative of trends occurring in Precinct 2 more
generally. Second, he contends that Dr. Murray’s conclusion incorrectly presumes that depressed voter
turnout amongst Latinos is an enduring and irreversible trend, when in fact Latino voter turnout would
most likely increase if Precinct 2 were designed to be a performing Latino-opportunity district because
Latinos who live in majority-minority districts vote at a much higher rate than Latinos who do not reside
in such districts. (Tr. 1:131-32).
136
The Court agrees. The characteristics of the city council races limit their probative value with
regard to commissioner court races. Moreover, the Court finds that the low turnout does not completely
invalidate the historical evidence that Latino candidates have a more difficult time getting elected to
office. Low Latino “turnout does not,” by itself, “militate against finding a Section 2 violation.” See
Benavidez, 638 F. Supp. 2d at 725. Dr. Barreto aptly noted that Latino voter turnout tends to increase
when an area becomes a Latino opportunity district; therefore, the historical patterns of political
participation in Precinct 2’s Latino community may not adequately estimate voter turnout under a nondiluted electoral system. Benavidez, 638 F. Supp. 2d at 725; see also Euclid, 632 F. Supp. 2d at 764.
b.
Statistical Electability Analysis
i.
2001 Map vs. Revised Plan A-1
Relying on the results of his reconstituted election analysis, Dr. Barreto concluded that the
Revised Plan A-1 diminishes the voting power of Precinct 2’s Latino population, thereby reducing its
ability to elect Latino-preferred candidates. (Tr 1:61). Reconstituted election analysis or electability
analysis is a statistical device “that extracts actual election results from a variety of statewide and local
races that subsume the area being analyzed and determines, precinct-by-precinct within the new district,
the racial composition of the vote and the ‘winner’ within the new district. This method of aggregation
allows a researcher to determine how an individual candidate performed within the boundaries of the
target district even though the actual election covered a different geographical area.” Rodrigez, 385 F.3d
at 861; see also Johnson v. Miller, 864 F. Supp. 1354, 1391 (S.D. Ga. 1994) (per curiam).
In this case, Dr. Barreto wanted to use reconstituted election analysis to determine how the 2001
map and the Revised Plan A-1 map affected the electability of recent candidates. Dr. Barreto analyzed
nine electoral races: (1) 2006 primary race for lieutenant governor; (2) the 2010 primary race for
lieutenant governor; (3) the 2010 primary race for land commissioner; (4) the 2002 general election race
for Supreme Court Justice, Position Number 1; (5) the 2008 general election race for U.S. Senate; (6) the
2002 general election race for governor; (7) the 2002 general election race for county treasurer; (8) the
137
2002 general election race for Supreme Court Justice, Position Number 4; and (9) the 2006 general
election race for Presiding Judge of the Court of Criminal Appeals.87 Under the 2001 map, the Latinopreferred candidate, in races one through five, successfully carried Precinct 2. See (Plaintiffs’ Ex. 39, at
12). However, when these same elections were reconstituted under the Revised A-1 map, the Latinopreferred candidate lost; under the Revised A-1 map, the Latino-preferred candidate captured 1-2% less
of the electorate than they would have captured had the election taken place under the 2001 map,
resulting in their defeat.
In the remaining races, races six through nine, the map under which the election occurred did not
change the outcome of the election: under both the 2001 map and the Revised Plan A-1, the Latinopreferred candidate wins. However, under the Revised Plan A-1, the Latino-preferred candidate wins
with a smaller margin of victory than they would under the 2001 map. (Tr. 1:91).
RECONSTITUTED ELECTION ANALYSIS: COMPARISON OF THE PERFORMANCE OF
LATINO-PREFERRED CANDIDATES UNDER THE 2001 MAP & THE REVISED A-1 MAP88
Primary Election Races Where the Latino Preferred Candidate Prevails Under the 2001 Map but is
Defeated Under the Revised A-1 Map
Electoral Race
#
1
Candidates89
% of Support
from Pr. 2
under 2001
Map
% of Support
from Pr. 2
under Revised
A-1 Map
2006 Primary
Lt. Gov.
Grant
Alvarado
Deleon
Margin of Victory for
Latino-Prf. Cand.
37.9%
38.3%
23.8%
+0.4%
38.2%
36.1%
25.7%
-2.1%
87
Difference
Between the
Support
Obtained under
2001 Map and
Revised A-1
+0.3%
-2.2%
+2.0%
-2.5%
In Plaintiffs’ Exhibit 39, the 2006 general election race for the Court of Criminal Appeals is simply denoted as
2006 judicial race. Comparing the candidates listed in Plaintiffs’ Exhibit 39 with the list of election results the
County has provided in Defendants’ Exhibit 51, the Court has discerned that the 2006 judicial race referred to in
Plaintiffs’ Exhibit 39 is in fact the race for presiding judge for the Court of Criminal Appeals. Cf. (Plaintiffs’ Ex.
39, at 13) with (Defendants’ Ex. 51, at p. 6 of November 2006 Election Results). The Court will thus refer to this
election as the 2006 general election race for Presiding Judge of the Court of Criminal Appeals.
88
Source: Plaintiffs’ Ex. 39, at 12-13.
89
In the column labeled “candidates,” the name of the Latino-preferred candidate is emboldened.
138
2
3
2010 Primary
Lt. Gov.
2010 Primary
Land. Comm.
Earl & Katz
(combined)
Chavez-Thompson
Margin of Victory for
Latino-Prf. Cand.
Burton
Uribe
Margin of Victory for
Latino-Prf. Cand.
49.2%
50.1%
+0.9%
50.8%
+1.6%
49.9%
-0.2%
-0.9%
-1.8%
49.4%
50.6%
+1.2%
51.1%
48.9%
-2.2%
+1.7%
-1.7%
-3.4%
General Election Races Where the Latino Preferred Candidate Prevails Under the 2001 Map but is
Defeated Under the Revised A-1 Map
#
Electoral Race
Candidates
% of Support
from Pr. 2
under 2001
Map
% of Support
from Pr. 2
under Revised
A-1 Map
4
2002 Sup. Ct. #
1
Schneider-GOP
Yanez-DEM
Margin of Victory for
Latino-Prf. Cand.
49.9%
51.1%
+0.2%
50.9%
49.1%
-1.8%
Difference
Between the
Support
Obtained under
2001 Map and
Revised A-1
+1.0%
-1.0%
-2.0%
5
2008
Senate
Cornyn-GOP
Noriega-DEM
Schick-LB
Margin of Victory for
Latino-Prf. Cand.
47.1%
50.9%
2.0%
+3.8%
49.5%
49.5%
2.0%
-0.9%
+2.4%
-2.3%
0.0%
-4.7%
U.S.
General Election Races Where the Latino Preferred Candidate Captures a Smaller Fraction of the
Electorate Under the Revised A-1 Map Relative to their Electoral Success Under the 2001 Map
#
Electoral Race
6
Candidates
% of Support
from Pr. 2
under 2001
Map
% of Support
from Pr. 2
under Revised
A-1 Map
2002 Governor
Perry-GOP
Sanchez-DEM
Others-OTH
Margin of Victory for
Latino-Prf. Cand.
50.3%
47.8%
1.9%
-2.5%
51.3%
46.9%
1.8%
-4.4%
139
Difference
Between the
Support
Obtained under
2001 Map and
Revised A-1
+1.0%
-0.9%
-0.1%
-1.9%
7
2002 County
Treasurer
Cato- GOP
Garcia-DEM
Margin of Victory for
Latino-Prf. Cand.
51.2%
48.8%
-2.4%
52.2%
47.8%
-4.4%
+1.0%
-1.0%
-2.0%
8
2002 Sup. Ct. #
4
Smith-GOP
Mirabal-DEM
Margin of Victory for
Latino-Prf. Cand.
46.7%
53.3%
+6.6%
47.7%
52.3%
+4.6%
+1.0%
-1.0%
-2.0%
9
2006 Ct. Crim.
App.,
Pres.
Judge
Keller-GOP
Molina-DEM
Margin of Victory for
Latino-Prf. Cand.
47.1%
50.9%
+3.8%
49.5%
49.5%
-0.9%
+2.4%
-2.3%
-4.7%
Just before trial, Dr. Barreto also performed reconstituted election analysis on the November
2012 race for county sheriff between Democrat Adrian Garcia and Republican Louis Guthrie and the
November 2012 race for 14th Court of Appeals, Place 8, between Republican John Donovan and
Democrat Julia Maldonado. According to Dr. Barreto, the reconstituted election analysis indicated that
Garcia and Maldonado are elected under the 2001 map but are defeated under the Revised A-1 map. (Tr.
1:63).
Based on the results of the reconstituted elections analysis, Dr. Barreto concluded that the
Latino-preferred candidate in Precinct 2 receives one to two percent less support under the Revised A-1
map than he or she would under the 2001 map. (Tr. 1:61, 64). According to Dr. Barreto, this one to two
percent difference usually results in the defeat of the Latino-preferred candidate, which of course
diminishes the ability of Latinos to elect their candidate of choice. (Tr. 1:64).
Dr. Barreto further testified that he performed reconstituted election analysis under the interim
map drawn by the Court. According to Dr. Barreto, Latino-preferred candidates in Precinct 2 capture a
higher share of the electorate under the Court-drawn map than under either the Revised A-1 map or the
2001 map. In fact, Dr. Barreto testified that, under the Court-drawn map, Latino-preferred candidates in
Precinct 2 are able to capture an additional four percent of the electorate than they otherwise would
under the Revised A-1 map. By the same token, Latino-preferred candidates were able to capture an
140
additional one percent of the electorate than they otherwise would under the 2001 map. According to Dr.
Barreto’s reconstituted election analysis, the Latino-preferred candidates are able to secure absolute
majorities of the electorate under the Court-drawn interim map, thereby enabling Latinos to elect their
candidate of choice. (Tr. 1:64).
ii.
Measuring the Effect of Voting Districts Added to/ Removed
from Precinct 2
Dr. Barreto also used the reconstituted maps to determine how the composition of the Precinct 2
under the Revised A-1 Plan affects the opportunities for political influence. In making that assessment,
Dr. Barreto first examined the population changes in the district to determine whether the Revised Plan
A-1 reduced the number of Latinos residing in Precinct 2. He then evaluated the individual voting
precincts that were removed from or added to Precinct 2 under the Revised A-1 map to determine
whether the change disrupted the political balance of Precinct 2, thereby diluting the opportunity for the
Latino citizens of Precinct 2 to give voice to their choice.
In the first stage of his analysis, Dr. Barreto compared the composition of the population living
in Precinct 2 under the 2001 map and the Revised A-1 map to determine whether the Revised A-1 map
reduced the number of Latinos residing in Precinct 2. Under the 2001 map, 60.1% of the population is
Latino; 55.2% of the voting age population is Latino; 34.9% of the citizen voting age population is
Latino; and only 29.7% of Precinct 2’s registered voters are Latino. (Plaintiffs’ Ex. 39, at 4).90 Under the
Revised A-1 map, 58.2 % of the population is Latino; 53.4 % of the voting age population is Latino;
33.8 % of the citizen voting age population is Latino; and 28.8 % of the registered voters in the precinct
are Latino. (Plaintiffs’ Ex. 39, at 4).91 Comparing the two schemes, Dr. Barreto determined that under
90
The total population and voting age statistics are derived using data presented by Harris County on July 25,
2011. The citizen voting age statistics were derived from the 2005-2009 ACS citizenship data. See (Plaintiffs’ Ex.
39, at 4).
91
The total population and voting age statistics are derived using data presented by Harris County on August 5,
2011. The citizen voting age statistics were derived from the 2005-2009 ACS citizenship data. See (Plaintiffs’ Ex.
39, at 4).
141
the Revised Plan A-1, Latinos constitute a smaller share of Precinct 2’s total population, voting age
population, citizen voting age population, and registered voters. (Tr. 1:75).
In the second stage of his analysis, Dr. Barreto compared the voting precincts included in
Precinct 2, in the Revised A-1 map, with the voting precincts that included in Precinct 2, in the 2001
map, to determine which voting precincts were removed or added under the County’s Plan and how
those changes affected Latino political influence in Precinct 2. (Tr. 1:76). In his expert report, Dr.
Barreto indicated that the Revised A-1 map moves twelve voting precincts that were formerly in
Precinct 4 into Precinct 2. According to the report, these twelve voting precincts are very conservative
and have consistently voted against Latino-preferred candidates. See (Plaintiffs’ Ex. 39, at 4). Dr.
Barreto explained that the precincts added to Precinct 2 have very high rates of voter registration and
participation; generally, these voting precincts are largely populated with Anglo-American voters who
consistently vote against the Latino-preferred candidate.92
92
Using ecological regression and ecological inference, Dr. Barreto concluded that Senator Barrack Obama was
the Latino-preferred candidate in the 2008 Presidential Election and that Maria Luisa Alvarado was the Latinopreferred candidate in the 2006 Lieutenant Governor’s race.
142
ELECTORAL RESULTS OF VOTING PRECINCTS MOVED FROM PRECINCT 4 TO
PRECINCT 293
2008 Presidential Election
VTD % Sen. Mc
% Sen.
Precinct
Precinct
Total
Votes for Votes for
Cain
Obama
Location
Location
Votes
Sen. Mc
Sen.
under 2001
under Revised
Cain
Obama
Map
A-1 Map
97
70.2%
27.4%
4
2
2879
2021
790
98
79.6%
18.5%
4
2
1415
1126
262
351
76.9%
21.6%
4
2
3416
2626
739
388
46.6%
52.1%
4
2
2928
1364
1526
502
77.7%
20.5%
4
2
1434
1114
294
604
79.7%
18.7%
4
2
1953
1556
365
636
83.1%
15.0%
4
2
2402
1996
361
658
62.9%
35.4%
4
2
1882
1183
666
659
70.4%
28.6%
4
2
3172
2232
906
674
71.9%
26.4%
4
2
2862
2059
756
700
77.2%
19.8%
4
2
890
687
176
764
60.2%
38.2%
4
2
5054
3041
1932
30287
21005
8773
(69.4%)
(29.0%)
TOTAL
2006 Lt. Governor Election
VTD
%
Dewhurst
%
Alvarado
97
98
351
388
502
604
636
658
659
674
700
764
69.0%
71.2%
78.0%
52.0%
72.9%
70.3%
73.1%
63.1%
71.8%
71.1%
69.8%
63.1%
25.6%
24.2%
18.2%
44.5%
22.3%
23.4%
22.2%
30.9%
24.3%
25.3%
24.3%
32.2%
Precinct
Location
under 2001
Map
4
4
4
4
4
4
4
4
4
4
4
4
Precinct
Location
under Revised
A-1 Map
2
2
2
2
2
2
2
2
2
2
2
2
TOTAL
93
See Plaintiffs’ Ex. 39, at 7.
143
Total
Votes
Votes for
Dewhurst
Votes for
Alvarado
1297
683
1916
688
667
992
1090
1020
1487
1496
477
1997
13810
895
486
1495
358
486
697
797
644
1067
1063
333
1261
9582
(69.4%)
332
165
348
306
149
232
242
315
361
379
116
644
3589
(26.0%)
Dr. Barreto further noted that the Revised A-1 map moves six voting precincts that were
formerly in Precinct 2 into Precinct 1. Dr. Barreto found that the six precincts that were removed from
Precinct 2 under the Revised A-1 map were high-performing Latino voting precincts that consistently
supported the Latino-preferred candidate. For instance, in the 2008 presidential election, these voting
precincts strongly supported the Democratic candidate, then-Senator Barack Obama, with 72.5% of the
voters in these precincts voting for Senator Obama. Similarly, in the 2006 election for lieutenant
governor, these voting precincts again supported the Democratic candidate, Maria Alvarado, with 66.2%
of the voters in these precincts voting for Alvarado over her Republican challenger. See (Plaintiffs’ Ex.
39, at 6).
144
ELECTORAL RESULTS OF VOTING PRECINCTS MOVED FROM PRECINCT 2 TO
PRECINCT 194
2008 Presidential Election
VTD % Sen. Mc
% Sen.
Precinct
Precinct
Total
Votes for Votes for
Cain
Obama
Location
Location
Votes
Sen. Mc
Sen.
under 2001
under Revised
Cain
Obama
Map
A-1 Map
284
10.1%
87.6%
2
1
1502
151
1316
353
18.1%
79.8%
2
1
2215
400
1768
368
28.3%
70.2%
2
1
532
148
367
376
49.1%
47.9%
2
1
1177
578
564
460
33.1%
64.2%
2
1
1540
510
988
605
23.9%
73.8%
2
1
3113
745
2297
10070
2532
7300
(25.1%)
(72.5%)
TOTAL
2006 Lt. Governor Election
VTD
%
Dewhurst
%
Alvarado
284
353
368
376
460
605
14.5%
23.8%
32.2%
53.0%
39.0%
28.9%
82.4%
74.5%
64.5%
42.7%
55.9%
68.2%
Precinct
Location
under 2001
Map
2
2
2
2
2
2
Precinct
Location
under Revised
A-1 Map
1
1
1
1
1
1
TOTAL
Total
Votes
Votes for
Dewhurst
Votes for
Alvarado
697
819
183
627
513
1152
3991
101
195
59
332
200
333
1220
(30.6%)
574
610
118
268
287
786
2643
(66.2%)
Dr. Barreto concluded that the removal of the six Democratic-leaning precincts and the addition
of the twelve Republican-leaning precincts altered the political landscape of Precinct 2 such that, under
the reformulated Precinct 2, there is less opportunity for Latinos to elect the candidate of their choice.
First, the precincts added to the district gave the Anglo-preferred candidate, Senator John McCain, an
approximate 12,000 vote advantage.95 See (Plaintiffs’ Ex. 39, at 7). Second, the precincts removed from
94
See Plaintiffs’ Ex. 39, at 6.
As the table above indicates, the precincts coming into Precinct 2 offered 21,005 votes in support of Senator
John McCain and 8,773 votes for now-President Obama; the difference between the vote counts captured by
Senator McCain and President Obama is approximately 9,000 votes.
95
145
Precinct 2 gave the Latino-preferred candidate, now-President Obama, approximately a 5,000 vote
advantage.96 Given that the Revised A-1 map removes 5,000 net votes for the Latino-preferred candidate
and offsets that subtraction with the addition of 12,000 net votes for the non-preferred candidate, these
modifications to the precinct result in an additional 17,000 net votes for the non-preferred candidate.
According to Dr. Barreto’s reconstituted election analyses, this change of 17,000 net voters results in a
2-3% drop in support for the Latino-preferred candidate, a difference which usually proves fatal to their
electability.
In light of the demographic changes accomplished by the Revised A-1 map, Dr. Barreto
concluded that the Revised A-1 map diminished the opportunity of Precinct 2’s Latino population to
elect the candidate of their choice because the County’s plan substantially altered the political
composition of the Precinct by swapping reliably Democratic voters for reliably Republican voters. See
(Plaintiffs’ Ex. 39, at 4).97
On balance, the Court finds that Latinos have a long history of election failures and the Revised
A-1 map would perpetuate the inability of Latinos to elect the candidate of their choice.
8.
Responsiveness of Elected Officials
An additional factor that in some cases has probative value in the totality of circumstances
inquiry is “whether there is a significant lack of responsiveness on the part of elected officials to the
particularized needs of the members of the minority group.” LULAC III, 986 F.2d at 752; see also
Campos, 840 F.2d at 1250.
96
As the table above indicates, the precincts removed from Precinct 2 offered 7,300 votes in support of nowPresident Obama and 2,532 votes for Senator John McCain; the difference between the vote counts captured by
President Obama and Senator McCain is approximately 5,000 votes.
97
Although the expert report included in Plaintiffs’ Exhibit states that these changes clearly violate Section 5 of
the Voting Rights Act, at trial Dr. Barreto testified that these changes would also diminish the opportunity for
Latino citizens of Precinct 2 to elect their candidate of choice. Although this exhibit often speaks of Section 5, see
e.g. (Plaintiffs’ Ex. 39, at 6), the Section 5 inquiry has already been determined by the Department of Justice and
is beyond the scope of the instant proceedings. Therefore, to the extent that this instrument is referenced by the
Court or relied upon by the parties at trial, it is used only with respect to the Section 2 inquiry that is currently
before the Court.
146
According to Councilman Rodriguez, former Commissioner Garcia was very responsive and proactively communicated with his office. (Tr. 2:133-34). He explained that, during Ms. Garcia’s tenure as
commissioner, he worked with her on infrastructure projects. In that capacity, he found her to be very
responsive. He could pick up a cell phone and call her if he needed her. (Tr. 2:134). Unfortunately,
Councilman Rodriguez has not enjoyed the same level of communication and accessibility with the new
commissioner. According to Mr. Rodriguez, the Precinct 2 commissioner’s office is not as responsive
under Commissioner Morman as it was under Commissioner Garcia. (Tr. 2:134). That said, Mr.
Rodriguez admitted that he has not had many opportunities to interact with Commissioner Morman nor
has he requested a meeting, so the lack of rapport between the two is, in part, a function of the limited
opportunities for interaction between the two. (Tr. 2:134, 136-37).
Senator Garcia explained that when she left office she was working on a series of projects in the
Latino community, which she believes the new commissioner has failed to pursue. (Tr. 3:46). According
to Senator Garcia, the new commissioner has abandoned her efforts on the Freeport Road Project in the
North Channel Area and her efforts to open a clinic in the lower Aldine area. (See Tr. 3:47, 52). She
further claimed that Mr. Morman deserted her efforts to encourage the municipal government of
Pasadena to maintain a Park and Ride program there, by failing to exert influence with the municipal
government to stress the value of the program. Finally, Senator Garcia claimed that Commissioner
Morman abandoned her efforts to preserve Juan Seguin Park as a County asset. (Tr. 3:47-48).
Senator Garcia contends that abandoning these projects imposes substantial costs on the
community. In the case of the road project, the cancellation will impact the mobility of North Channel
residents, limiting their access to bus systems and transportation, which, of course, limits their ability to
access basic services. (Tr. 3:47). As for the clinic, the cancellation will limit the community’s access to
affordable health care services. (Tr. 3:47). The abandonment of the Park and Ride will deprive the
residents of Pasadena of an affordable means of access. Finally, Senator Garcia contends that the
147
County’s dispossession of Juan Seguin Park is significant because the park is named for one of the most
prominent Latino figures in Texas history and therefore holds symbolic value. (Tr. 3:48-49).
Defendants, however, claim that Senator Garcia is misinformed. First, Defendants noted that
Senator Garcia’s information was out-dated and unreliable. In fact, on cross-examination, Senator
Garcia admitted as much when she acknowledged that she did not know the current status of the
Freeport road project, the Pasadena Park and Ride, and Juan Seguin Park, and that the testimony she
offered on direct was based on information that was almost a year old at the time of trial. (Tr. 3:52-53).
As for the clinic in lower Aldine, she admitted that she did not know why the plans changed. (Tr. 3:52).
Second, Mr. Walden, Commissioner Morman’s chief of staff, testified that he has been
responsive to the needs of the Latino community. Mr. Walden testified that the Freeport road project
was not canceled. (Tr. 3:158). He further explained that Commissioner Morman did not abandon the
lower Aldine health clinic. According to Mr. Walden, although there were discussions about creating the
health clinic, the commissioner could not find any measureable source of funding to build or operate the
facility. (Tr. 3:159). Therefore, that project is on hold until funding can be secured for the facility. (Tr.
3:159, 163-64).
As for the Seguin Park facility, Mr. Walden testified that the reports that Commissioner Morman
tried to sell Seguin Park were exaggerated. According to Mr. Walden, the commissioner did not try to
sell the park in the conventional sense but instead had a pre-arranged agreement with the Texas Parks &
Wildlife Department whereby the County would auction the park to the Department and then the county
and state would jointly develop the land. (Tr. 3:159). The transaction had to be arranged in this way
because, according to Mr. Walden, the state preferred to own the land before it would invest monies to
develop it. The transaction fell through after the state concluded that the land was not worth its appraised
value. (Tr. 3:159).
Regarding the Park and Ride, Mr. Walden explained that the Pasadena Park and Ride suffered
from problems associated with low ridership; on its busiest days, only fifteen or so people took
148
advantage of the service. (Tr. 3:160). At some point, the City of Pasadena determined that the benefits of
maintaining a park and ride service did not justify the costs. (Tr. 3:160). Commissioner Morman did not
advocate against the termination of the Park and Ride service because he does not, as a philosophical
matter, interfere with municipal government decisions. (Tr. 3:164-65).
In light of the entire evidence, the Court finds that Plaintiffs have failed to establish that the
newly elected Precinct 2 commissioner is not responsive to the needs of the Latino community.
That said, Senator Garcia testified that Precinct 2 residents feel that the County government, at
large, has been unresponsive to their needs. (Tr. 3:37). According to Senator Garcia, the neglect is not
necessarily the result of animus but instead results from the County’s failure to appreciate and
understand the needs of the Latino community. In her view, the County is unfamiliar with the needs of
the Latino community because there are few people in County government who are both familiar with
the issues facing the Latino community and are able to exert influence to address those issues. (Tr. 3:37).
As a result, the County often does not provide Latino communities with the services these communities
need. (Tr. 3:37). For example, in the lower Aldine area—an impoverished predominantly Latino
neighborhood in Precinct 2 whose residents are still dependent on water wells and septic tanks—the
residents lack park services, endure poor road conditions, and are plagued with health issues related to
their use of water wells. (Tr. 3:37-38). During her tenure as County Commissioner, county personnel
often referred to this area as “no man’s land” because no one wanted to contend with the problems
plaguing this section of the County. (Tr. 3:38). Thus, this area and its residents were largely ignored.
(Tr. 3:38). Senator Garcia testified that other predominately Latino areas, like Cloverleaf and Emos,
were similarly neglected by County officials. (Tr. 3:38-39). In Cloverleaf and Emos, the roads are
poorly maintained, the community lacks sidewalks, and the park and recreational services are lacking.
(Tr. 3:38).
In addition to the County’s failure to provide adequate physical plant services, Senator Garcia
also testified that the County fails to provide “soft” services, like bilingual services or sensitivity to the
149
cultural needs of its Latino residents. (Tr. 3:39). For instance, in the days preceding Hurricane Katrina’s
landfall, Latino residents attempted to call the Harris County Office of Emergency Management
(“HCOEM”) or Houston TranStar, a joint governmental response unit by the City of Houston and
HCOEM, to get information about services and assistance offered by TranStar and/or HCOEM.
However, neither TranStar nor HCOEM had bilingual staffing. Therefore her office served as a de facto
liaison between residents and the local emergency management services. (Tr. 3:39-41).
Senator Garcia explained that the County’s failure to provide “soft” services is also evidenced by
the dearth of county facilities honoring or recognizing the contributions of Latino residents. (Tr. 3:42).
When she first assumed office in 2003, there were few buildings, parks, or streets in Precinct 2 that were
named after Latino citizens. (Tr. 3:42). According to Senator Garcia, the county’s failure to name
facilities in honor of Latinos diminishes the contribution of Latinos and deprives Latino residents of a
sense of pride, inclusiveness, and respect. (Tr. 3:42).
9.
Tenuousness of Harris County’s Reason for Maintaining Electoral Practice
In conducting the totality of the circumstances analysis, the Court also examines whether the
policy underlying the state or political subdivision’s use of the challenged voting practice or procedure
is tenuous. The “tenuousness of the policy underlying the challenged practice, although less important
under the ‘results test’ embodied in Section 2, is nonetheless relevant.” LULAC III, 986 F.2d at 752-53.
Evidence of a tenuous underlying state policy may “indicate that the practice or procedure produces a
discriminatory result . . . [or] that the policy is unfair.” LULAC III, 986 F.2d at 753.
Defendants offered substantial testimony that this round of redistricting posed a unique set of
problems: (1) maintaining Precinct 1’s status as an African-American opportunity district; (2)
redistributing the population of Harris County from the over-populated Precincts 3 and 4 to the underpopulated Precincts 1 and 2; (3) maintaining Latino influence in Precinct 2; and (4) maintaining
incumbent relationships. Although Plaintiffs disagree with the manner in which the County has balanced
150
these competing considerations, there is no evidence that the County’s reasons for adopting and
maintaining the Revised Plan A-1 are arbitrary or without adequate basis.
10.
Proportionality
In looking at the totality of circumstances, the Court must also consider “whether the number of
districts in which the minority group forms an effective majority is roughly proportional to its share of
the population in the relevant area.” Perry, 548 U.S. at 426. To be sure, “nothing in [the Voting Rights
Act] establishes a right to have members of a protected class elected in numbers equal to their
proportion in the population.” 42 U.S.C. § 1973(b). Nevertheless, proportionality provides some
evidence of whether “the political processes leading to nomination or election in the State or political
subdivision are not equally open to participation.” § 1973(b); Perry, 548 U.S. at 437-38. Accordingly,
proportionality has become a key measure of fairness in redistricting. See De Grandy, 512 U.S. at 1018.
For purposes of a vote dilution claim, there is, of course, no “magic parameter” with respect to
proportionality, and “rough proportionality” must allow for some deviations. The degree of probative
value assigned to proportionality may vary with other facts. Perry, 548 U.S. at 426.
Because rough proportionality under the totality of the circumstances test looks to citizen voting
age population at the county-wide level, the Court finds that the ACS figures are reliable for this
purpose. See Perry, 548 U.S. at 426; Comm. for a Fair & Balanced Map v. Ill. State Bd. of Elections,
835 F. Supp. 2d 563, 586 (N. D. Ill. 2011). Dr. Murray testified that, according to the 2010 census data,
25% of the citizen voting age population in Harris County is Latino. (Tr. 4:1). Since there are four
commissioner’s court precincts, if the number of districts in which Latinos formed an effective majority
were proportional to their share of the citizen voting age population, then the Court would expect there
to be one commissioner’s court precinct in which Latinos form an effective majority.
Under the Revised A-1 map, however, there is not a single precinct in which Latinos form an
effective majority. The disproportionality occasioned by the Revised Plan A-1 lends credence to
Plaintiffs’ claims that the Revised Plan A-1 dilutes their voting power.
151
IV.
LAW & ANALYSIS: CONSTITUTIONAL CLAIMS
A.
VOTE DILUTION
Plaintiffs contend that the adoption of the Revised Plan A-1 also violates the Equal Protection
Clause of the Fourteenth Amendment because the plan was adopted with the intent to dilute the voting
power of Latinos in Precinct 2.
The essence of a vote dilution claim under the Fourteenth Amendment is “that the State has
enacted a particular voting scheme as a purposeful device ‘to minimize or cancel out the voting potential
of racial or ethnic minorities.’” Miller, 515 U.S. at 911 (quoting City of Mobile v. Bolden, 446 U.S. 55,
66 (1980) (plurality opinion)). To obtain relief on a constitutional vote dilution claim such as this, the
plaintiffs must “prove that the purpose and operative effect” of the challenged election scheme “is to
dilute the voting strength of [minority] citizens.” Voter Info. Project, Inc. v. City of Baton Rouge, 612
F.2d 208, 212 (5th Cir. 1980).
To prove discriminatory purpose, the plaintiff does not need to advance direct evidence of
discriminatory intent. Rogers v. Lodge, 458 U.S. 613, 618 (1982). Instead, “discriminatory purpose may
often be inferred from the totality of the relevant facts, including the fact, if it is true, that the law bears
more heavily on one race than another.” Washington v. Davis, 426 U.S. 229, 242 (1976) (internal
quotation marks omitted). Some circumstantial indicia of discriminatory purpose to be considered by the
Court should include: (1) whether bloc voting along racial lines exists; (2) whether minorities are
excluded from the political process; (3) whether minority voter registration is low; (4) whether elected
officials are unresponsive to the needs of minorities; (5) whether the minority group occupies a
depressed socioeconomic status because of inferior education or employment and housing
discrimination; (6) the historical backdrop leading to the passage of the redistricting legislation; (7) “the
specific sequence of events leading up to the challenged decision”; (8) whether the redistricting body
departed from the normal procedural sequence for passing redistricting legislation; (9) whether the
voting strength of a cohesive minority group has decreased or “retrogressed”; and (10) whether district
152
boundaries have been manipulated to adjust the relative size of minority groups, including instances of
“packing.” Backus v. S. Carolina, 857 F. Supp. 2d 553, 558 (D.S.D. 2012).
To prove discriminatory effect, a plaintiff must show that the redistricting scheme impermissibly
dilutes the voting rights of the racial minority. Generally, this requires proof that the racial minority’s
voting potential has been minimized or cancelled out or the political strength of such a group has been
adversely affected. See Mobile, 446 U.S. at 66, 84 (internal quotation marks omitted). Plaintiffs alleging
vote dilution must offer “a reasonable alternative voting practice to serve as the benchmark ‘undiluted’
voting practice.” Reno v. Bossier Parish Sch. Bd. (Reno I), 520 U.S. 471, 480 (1997). That is because
“the very concept of vote dilution implies—and, indeed, necessitates—the existence of an ‘undiluted’
practice against which the fact of dilution may be measured.” Reno I, 520 U.S. at 480. An inquiry “into
dilutive effect must rest on some idea of a reasonable allocation of power between minority and majority
voters; this requires a court to compare a challenged voting practice with a reasonable alternative
practice.” Reno v. Bossier Parish Sch. Bd. (Reno II), 528 U.S. 320, 367-368 (2000).
On the present record, the Court finds that Plaintiffs have failed to prove the County acted with a
discriminatory intent in the adoption or maintenance of the precinct boundaries for the Harris County
Commissioner’s Court. There is no convincing evidence indicating that the Commissioner’s Court drew
the district lines for the purpose of diluting Plaintiffs’ voting strength.
B.
GERRYMANDERING CLAIMS
Plaintiffs contend that the adoption of the Revised Plan A-1 also violates the Equal Protection
Clause of the Fourteenth Amendment because the Plan was a product of racial gerrymandering98 in that
it was drawn with excessive and unjustified use of race and racial data.
98
The term “gerrymander” was coined by combining the last name of Elbridge Gerry with the word “salamander”
in order to describe the “fancied resemblance to a salamander . . . of the irregularly shaped outline of an election
district in northeastern Massachusetts that had been formed for partisan purposes in 1812 during Gerry’s
governorship” of that State. Davis v. Bandemer, 478 U.S. 109, 164 n.3 (1986).
153
Racial gerrymandering of electoral districts involves the “deliberate and arbitrary distortion of
district boundaries . . . for [racial] purposes.” Shaw I, 509 U.S. at 640; Prejean v. Foster, 337 F.3d 504,
509 (5th Cir. 2003). Given the presumption of the legislature’s good faith in redistricting, showing that a
redistricting plan intentionally discriminates is not ordinarily an easy task. Prejean, 337 F.3d at 509. A
plaintiff “must prove that the legislature subordinated traditional race-neutral [re]districting principles . .
. to racial considerations,” such that race was “the predominant factor motivating the legislature’s
[redistricting] decision.” Miller, 515 U.S. at 916; Chen v. City of Houston, 206 F.3d 502, 505 (5th Cir.
2000) (internal citations omitted). The mere fact that race was given some consideration in the
redistricting process does not alone suffice in all circumstances to trigger strict scrutiny. Chen, 206 F.3d
at 506; Theriot v. Parish of Jefferson, 185 F.3d 477, 488 (5th Cir. 1999).
Mr. Walden testified that as chief of staff he is responsible for campaigning for his candidate and
ensuring his candidate’s electability. (Tr. 3:148). Throughout the redistricting process, Mr. Walden
sought detailed information from the County Consultants on Precinct 2’s composition under the 2001
map and its composition under the proposed redistricting. For each voting districts being considered for
removal or addition to Precinct 2, Mr. Walden reviewed detailed information on the straight ticket
voting patterns, racial composition, voter turnout, and the extent to which the support for the Republican
candidate varied based on the race of the Democratic opponent. (Tr. 3:138 -141). Mr. Walden testified
that he used this data to help him evaluate which voting districts Commissioner Morman would
recommend should be included in the re-districted Precinct 2 and which districts should be excluded
from the re-fashioned Precinct 2. (Tr. 3:140).99 Mr. Walden explained that these statistics helped him
evaluate how the political landscape of the map changed with the inclusion or exclusion of a particular
voting precinct, (Tr. 3:155-56), so that he could optimize Mr. Morman’s political advantage. (Tr. 3:169).
99
According to Mr. Walden, the redistricting process is steered by the attorneys hired to manage the process and
the consultants; the individual commissioners cannot dictate how their Precinct will ultimately be composed but
may nevertheless influence the outcome of the redistricting process. (Tr. 3:150-51).
154
Thus, for instance, when the County was considering whether to add the voting precincts east of
Lake Houston to Precinct 2, Mr. Walden was well aware that these voting precincts had a population of
32,353 persons, of which 24,916 (or 77.0%) were Anglo and 4,987 (or 18.4%) were Latino. (Plaintiffs’
Ex. 32, at 4); (Tr. 3:147). Moreover, he understood that bringing these precincts into Precinct 2 would
increase the number of Anglo and Republican voters in the Precinct. (Tr. 3:147); see also (Plaintiffs’ Ex.
32, at 4). Mr. Walden further explained that when he was evaluating whether a voting precinct should be
added to (or remain in) Precinct 2, he carefully examined that voting precinct’s straight ticket voting
trends. See (Tr. 3:142). Thus, he was well aware that there were more straight ticket Republican voters
under the original A-1 map than there were under the 2001 map. See (Tr. 3:148). And he knew these
changes increased Mr. Morman’s political advantage. (Tr. 3:148).
STRAIGHT TICKET ANALYSIS OF PRECINCT 2 UNDER THE 2001 MAP & UNDER THE
A-1 MAP100
Map
2010 Vote
Straight Ticket
Straight Ticket
No Straight
Republican
Democrat
Ticket
2001 Map
134,624
48,303
37,844
48,477
(35.9%)
(28.1%)
(36.0%)
A-1
159,012
62,392
40,114
56,506
(39.2%)
(25.2%)
(35.5%)
In Mr. Walden’s view, there was nothing improper about his consideration of the straight ticket
voting patterns, racial composition, voter turnout, and other precinct-level voting data. As chief of staff,
he was tasked with knowing the political impact of the map and—to the extent possible—making sure
that the map enhanced his candidate’s political viability. See (Tr. 3:155-56). Mr. Walden further
explained that although he was trying to maximize the electoral advantage for Commissioner Morman,
he did so without considering whether the proposed changes would affect the ability of Latino citizens
of Precinct 2 to elect a Latino candidate. (Tr. 3:149). He was merely trying to increase the number of
100
Source: Plaintiffs’ Ex. 25, at 8.
155
Republican voters in the Precinct. In his view, making a precinct more Republican does not necessarily
make it more difficult for Latinos to elect the candidate of their choice. (Tr. 3:150).
Plaintiffs, however, suggest that much of the information sought had a racial dimension. In an
email exchange with another colleague, Ms. Margo White, Mr. Walden requested information about the
Latino political victories in Precinct 2 during the 2010 election. (Plaintiffs’ Ex. 26, at 2); (Tr. 3:143). On
another occasion, Mr. Walden asked the attorneys for the County to determine whether the race of the
candidate had any impact on the electoral decisions of voters in the Aldine area. See (Plaintiffs’ Ex. 27);
(Tr. 3:144).
Mr. Walden claims that he sought this information to confirm his suspicion that the race of the
candidate did not really matter in determining whether the Democratic candidate would prevail in the
Aldine area. See (Tr. 3:145); see also (Plaintiffs’ Ex. 27, at 2). According to Mr. Walden, jurisdictions
like Harris County must always be mindful of the racial composition of their jurisdictions, because that
is a consideration the Justice Department evaluates when determining whether to preclear the County’s
proposed voting change. (Tr. 3:158). Although Mr. Walden did not have any formal responsibility in
connection with the County’s pre-clearance process, his primary goal in obtaining the racial
demographic data was to ensure that the commissioner’s court would produce a map that could get precleared by the Justice Department. (Tr. 3:163). He did not want the commissioner’s court to approve a
plan that could not be pre-cleared by the Justice Department. (Tr. 3:166-68).
On review, the Court finds that Mr. Walden’s candid testimony unsurprisingly demonstrates that
he intended to design a map that would preserve his candidates’ political viability. This, of course, does
not run afoul of the Constitution. And while, to be sure, Mr. Walden did consider the racial composition
of the electorate and the ways in which the voting trends of the electorate change in relation to the race
of the candidates running for office, the evidence presented before the Court does not show that these
racial and ethnic considerations predominated the redistricting process to the subordination of other
traditional redistricting principles. To the contrary, Defendants’ expert, Dr. Murray, offered un-rebutted
156
testimony that the County’s map was drawn in compliance with traditional redistricting principles. Dr.
Murray testified that the redistricting process that Harris County used in 2011 was no different than the
redistricting process used in 2001. (Tr. 3:247). The County drew the map in accordance with its
principles and priorities; there was an open process in Harris County that allowed for and sought out
comment from the public; and then there was reevaluation of the map in light of the public comments.
(Tr. 3:248). In Dr. Murray’s view, there was no racial motivation in the Harris County redistricting
process. (Tr. 3:249).
Ultimately, the evidence simply demonstrates that Mr. Walden was concerned with preserving
his candidate’s incumbency. The evidence shows that Mr. Walden was concerned with increasing the
political clout of Republicans in Precinct 2 while at the same time minimizing the influence of
Democratic factions in the precinct. This conduct speaks more of political jockeying than racial
gerrymandering. Although the Court is mindful of the testimony that Latinos in Harris County tend to
vote Democratic and Anglos tend to vote Republican, these proclivities, without more, cannot transform
partisanship into race discrimination. Second, redistricting, of course, requires the government be
mindful of the racial composition of their jurisdictions. Therefore, mere evidence that the government
considered race information does not, without more, shed light on the question of unlawful
discrimination under the Fourteenth Amendment; instead Plaintiffs must show that racial and ethnic
considerations steered the redistricting process. Plaintiffs have fallen far short of that burden.
V.
CONCLUSION
In conclusion, the Court finds that Plaintiffs have not established that the County’s Plan was
unconstitutional. Plaintiffs did not prove that the County adopted the Plan intending to discriminate
against Latinos nor did they prove that race was the predominant factor in the design of the Plan.
Moreover, because Plaintiffs have not established the first Gingles factor, they cannot prove that, under
the totality of the circumstances, Latinos have been denied an equal opportunity to participate in the
157
political process and to elect candidates of their choice. See Growe, 507 U.S. at 40-41; Fairley, 584 F.3d
at 667; Valdespino, 168 F.3d at 852.
Nevertheless, Plaintiffs “have raised serious and valid concerns that hindrances to equal
opportunity to participate in the political process are present.” Perez, 958 F. Supp. at 1230. Indeed this
Court is troubled by evidence of the range and prevalence of voter suppression tactics employed against
members of the Latino community. While some members of the Supreme Court imagine that barriers to
voting have been eradicated, see Northwest Austin Municipal Util. Dist. No. 1 v. Holder, 557 U. S. 193,
204 (2009) (noting that “things have changed in the South”), the record here is replete with evidence to
the contrary. While, to be sure, the County has taken some strides toward voter equality, it is clear that
more can be done to ensure that all citizens have a full and fair opportunity to participate in the political
process.
Although Plaintiffs have not carried their burden today, the evidence presented leads to the
inescapable conclusion that a Latino opportunity district will be possible in Harris County in the
foreseeable future. The sleeping giant is waking and those standing in its path would be wise to move
out of the way. The writing on the wall is clear: “the times, they are a changing’….”101
The Clerk shall enter this Order and provide a copy to all parties.
1st
SIGNED on this the ____ day of August, 2013, at Houston, Texas.
____________________________________
VANESSA D. GILMORE
UNITED STATES DISTRICT JUDGE
101
Bob Dylan, The Times They Are A-Changin’ (Columbia Records 1964) (a protest song from the sixties).
158
Disclaimer: Justia Dockets & Filings provides public litigation records from the federal appellate and district courts. These filings and docket sheets should not be considered findings of fact or liability, nor do they necessarily reflect the view of Justia.
Why Is My Information Online?