*2 THORNBERRY, Before RUBIN and JONES, Judges. Circuit PER CURIAM: We are asked black and Mexican- plaintiffs-appellants American to review City at-large, whether the of Austin’s ma jority place system for election of seq. members violates Section 2 et Voting Rights Act of 42 U.S.C. (1982) (as amended). seq. et (a) voting qualification prerequi- or “searching and No performed district court standard, practice, or site to or present ‘past evaluation practical imposed applied by procedure shall be Cong., 2d 97th reality’,” S.Rep. No. in a any State or subdivision Cong. & Sess., in 1982 U.S.Code reprinted *3 results in a denial or manner which 177, 208, both before Thorn Admin.News abridgement right any of the citizen of 80, 478 U.S. burg Gingles, v. to vote on account of and, the United States (1986), addi 2752, L.Ed.2d 25 92 color, in or or contravention of the race evidence, on re Thornburg after tional in guarantees set forth section 150. The 798 F.2d mand from this Court. title, 1973b(f)(2) provided of this We af found no district court violation.. (b) of this section. subsection firm. (b) (a) ... A violation of subsection if, totality on the established based I. INTRODUCTION circumstances, politi- it is shown that the a members and six council Austin elects processes leading to nomination or cal using at-large an majority vote mayor by political election in the State or subdivi- city capital Although the Texas system. equally open participation sion are not and a Mexican-Ameri- had both a black has by protect- of class of citizens members a years, this nearly 15 can on its council (a) by ed ... in that its mem- subsection From 1924 to always the case. was not opportunity than have less other bers at-large 1951, elected council was participate members of the electorate places by plurality vote. designated without political process repre- and to elect for the unsuccessfully ran Black candidates of their choice. The extent to sentatives 1969, 1949, 1951, and a protected which members of a class have was defeated candidate Mexican-American to office in the State or been elected Building skele- upon this in a 1969 run-off. political subdivision is one circumstance plaintiffs, black history, electoral tal Provided, may considered: which be individuals, and Mexican-American a nothing That this section establishes MALDEF,1 prove seek to two NAACP protected right members of a to have city charter contentions: the 1953 equal to their in numbers class elected amendments, sys- place a which introduced population. in the proportion requirement, were majority vote tem and (1988). Congress provid- 1973 42 U.S.C. § discriminatorily and thus war- motivated totality of the circumstances ed that the relief; Voting Rights Act and that rant whether, specific because of determines notwithstanding significant successes district- practice like multimember years, minority candidates in recent blacks of a State or ing, political processes suffered ac- Mexican-Americans have open partic- equally are not subdivision relief, dilution.2 As racial vote tionable Simultaneously, Con- races. ipation all pos- court two they urged upon the district requirement propor- rejected any gress single-member council dis- forms of sible S.Rep. representation race. See tional member, mayor one a six council tricts: 417, supra. No. nearly majority scheme with in- significant Supreme Court’s first district; or an voting age citizens in one 2 is found amended Section terpretation of member, mayor plan in one eight council Gingles, U.S. and Mexican-Americans which blacks (1986). The Su- 92 L.Ed.2d S.Ct. districts. comprise majority in two would report of the Sеn- preme Court viewed accompanying the Judiciary ate Committee analysis starts with amended Our 2 as instructive Section Rights that amended Voting Act: bill 2 of the Section intervened in Task Force The Black Citizens stand for National NAACPand MALDEF 1. The City participated of Colored on the side for the Advancement case and Association this Legal People Defense and Mexican-American of Austin. Fund, respectively. voting; plain- might proba- of racial bloc factors that be the existence of the various prove it. Voting Rights Act claim: tiffs must tive of a Report specifies factors The Senate 44-46, at 106 S.Ct. at Thornburg, 478 U.S. relevant to a typically which 2763-64. voting-related history of dis- claim: the Discussing the issue of multimember dis- political or crimination in the State subdi- tricts, again noted that Court vision; voting in the the extent to which per violative of voters’ se political or elections of the State subdivi- rights. Thornburg, 478 U.S. racially polarized; the extent to sion is held, “Minority S.Ct. at 2765. It subdivision State that the multimember form of who contend *4 voting practices procedures used or has districting prove violates must that the § opportunity tend to enhance the use of a electoral structure multimеmber against minority the discrimination operates to minimize or cancel out their group, unusually large such as election ability preferred to elect their candidates. districts, majority requirements, vote and See, e.g., S.Rep. 16.” Id. voting; prohibitions against bullet the that, The Court then noted exclusion of of the members many While or all of the factors listed in group slating processes; from candidate Report may the be relevant to a Senate group the extent to which mem- through claim of vote dilution submer- past the effects of discrimina- bers bear districts, gence in multimember unless education, employ- tion areas such as following conjunction there is a of the ment, health, and which hinder their abil- circumstances, the use of multimember ity participate effectively politi- in the generally impede districts will not the process; cal use overt or subtle appeals political ability minority repre- voters to elect campaigns; racial the extent to which of the mi- sentatives of their choice. members nority group public have elected to 47-50, Id. 478 at at U.S. 2765-66. jurisdiction Report office in the ... The The circumstances cited the Court are demonstrating notes also that evidence minority group: that the unresponsive that elected officials are (1) sufficiently demonstrate that it particularized needs of the members large geographically compact to con- minority group and that the policy majority single-member stitute a in a dis- underlying the State’s or the trict; prac- subdivision’s use of the contested (2) cohesive; politically show that it is may tice or structure is tenuous probative value. (3) majority demonstrate that the white specifically recognized Court that: sufficiently votes as a bloc enable Although Report espouses the Senate a special it—in the absence of circumstanc- flexible, fact-intensive test for 2 viola- § es, such as the candidate run- tions, it limits the circumstances under ning unopposed usually to defeat — may proved violations § minority’s preferred candidate. First, ways. devices, three electoral 50-52, id. 478 U.S. at See 106 S.Ct. at elections, at-large such as not be 2766-67. per considered se violative 2. Plain- opinion The remainder of the Court’s that, tiffs must demonstrate under the upon elaborates the nature of the evidence circumstances, totality of the the devices relevant to these three factors. unequal result in access to the electoral process, Second, Neither the nor the dissenters [citation omitted] conjunction allegedly Congress of an in Thornburg dilutive elec- itself— —nor adopted toral mechanism propor- and the lack of talismanic test for a Section representation tional upon alone does violation. not es- Reliance either the Court’s violation, tablish voter dilution analysis, threshold the “total- [citation omitted] Third, ity circumstances,” results test does not upon assume alleged below, found therefore discussed that the discriminatory acts intentional change fact-bound, intensely in- 1953 charter did not result from a local becomes intent, discriminatory racially district has not been highly dependent upon the quiry such, clearly purpose, maintained for that and has had conclusions. As court’s no such result. While the court acknowl- applies to district court’s test erroneous eighth edged that a black candidate ranked findings. Thornburg, among 14 2781; Baytown, in the number of votes received Campos City S.Ct. at Cir.1988). candidates in the court found (5th 840 F.2d “pure folly major it that a and radical of re- this deferential standard Despite change system in the entire electoral was a view, eyes to the fact may not close our we candidacy reaction to the of a black man analysis, fluidity that the of the Section for office.” who ran may encourage glossed by Thornburg, facts aggregate statistics or litigants Between 1951 and a 15 member hope that a indiscriminately in the propose changes rather committee was created to simply count facts rather court will district in the charter. The Austin amend- carefully. The district analyze than them product year-long ments of a were *5 short-coming prey to that committee, court did not fall study by this which included Indeed, the case was well-tried representatives here. communities. copious, compre- and resulted propositions repre- sides on in both The voted findings trial court. We by hensive the complete of the structure sented overhaul findings some detail must review those city government. There were more than appellants’ as- perspective lend to the separate propositions to Is- on the ballot. signments of error.3 power city manag- the sues included candidates,
er, eligibility requirements for II. OF DISTRICT SUMMARY limited-purpose annexation. and the use of COURT OPINIONS light wide-ranging changes pro- In updated ex population, Austin’s to to conclude posed, the court was unable 362,000, of whom approximately ceeds 12% racial discrimination motivated two of of whom are Mexican- are black 18% changes. those The court found insuffi- pertinent findings American. The court’s any racial ten- cient credible evidence (1) grouped topics: under six histo it sion existed in Austin in 1953or that was (2) slating; districting; ry of multimember change the time that the to “well known at (4) (3) elections; re appeals past racial requirements di- place would sponsiveness city to council minority voting power.” Evidence lute needs; (5) justification tenuousness of the absent, stated, support- was (6) districting; the Thorn for multimember me- of the charter amendments or the ers burg factors. emphasized potential the vote dilution dia change. The court noted aspects of the History multimember district- adopt staggered city did not also that the ing. vote-diluting ar- potentially or other terms Proposition effected at-large rangements. implemented Austin an election elections, brought change place also calling for a member to system five including setting the changes, large plurality other council elected at on a basis. about Saturday in elections on the first mayoral position by then filled an date for was requirement lowering the sitting April, residence among election the five council councilmen, eliminating the re- city adopted a ma- members. petition of 25 voters to quirement for a jority requirement vote and chose to elect filing. The court by designation. Contrary accompany a candidate’s place councilmen contentions, of Austin did appellants’ which will be concluded that noted, court, reasoning previously the facts and on which the district on re- We recite 3. As Court, ultimately concluding district court relied in mand from this took additional evidence findings light Thornburg. that there was no 2 violation. and amended its groups recognized The court that various adopted it intend to discriminate when energy the elec- system money, in 1953.4 donate time and to majority, at-large electoral candidates, single it no tion of but found 2.Slating. strength slating group to be of sufficient by dеny the intent mean- or motivated slating as the The district court defined dilute their ingful access to minorities or to candi- package or slate of creation of short, voting strength. the court found office, orga- dates, filing an before by fac- politics that Austin are influenced strength make sufficient nization with constantly shifting coali- tions who form stamp approval of merely a the election other than racial tions motivated issues group. Slating candidate pre-ordained exclusivity. control effective ac- operate could thus The trial cess of minorities to ballot. appeals. 3. Racial evidence that minorities ex- court found no taking perience any such roadblocks to memory campaign No Austin in recent any place their on the ballot for ethnicity. race has included issues of Further, the court found that no seat. particularly The court dismissed group political power or financial has appeals contention that subliminal racial resources to control the election of a slate accompanied rejection the voters’ in 1985 of of candidates. proposing single-member an amendment districts. accepted arguendo appel- The court Progres-
lants’ identification of the Austin Responsiveness (APC) needs. “powerful sive Coalition as a endors- *6 Nevertheless, ing group.” appellants the city The court considered the Austin failed to establish that members of minori- especially council to be attentive to the ty groups participate not allowed to in were suggested needs of minorities. It that an minority the APC or candidates were at-large system electoral forces each coun- contrary, denied its endorsement. On the respond minority cil member to com- at least one black and one Mexican-Ameri- munity, support to whom he must look for APC and by can have endorsed the polls. at the The court considered testimo- by every city elected the voters in council ny by minority council members that election since 1975. At least one black fully represent minority felt the unable candidate was endorsed and elected in the community, testimony but it found this 1971 and 1973 elections. A second Mexi- vague. The court observed numerous can-American, DeLeon, was endorsed but officials, city including city manager, the not elected 1981. city manager, senior assistant and seven The court minority department found that candi- out of 17 heads are black or dates, also approval Hispanic. Among who received the city’s the various boards the minority community, by commissions, are endorsed and a total of 16% the APC. It also appointments found that in 1984-85 were black and groups participate in the APC endorsement were Mexican-Americans. These 11.5% process. Austin, observed, In the figures higher court a were than those of 1976-77. exist, endorsing groups number of city passed each of The antidiscrimination laws cov- pursues agenda. which its ering public accommodations, own No evi- employment, suggested dence that the value or effect of housing Sixty and per- 1976 and 1977. an APC endorsement is significant city’s $273,000,000.00 more cent of the in commu- than that of the NAACP or the nity development Mexican- grant bloc funds between Democrats, American or that APC endorse- 1979 and 1984 housing. was related to The highly ment is more coveted city than that of corporation also funds a to make mort- any group. other gage loans to persons, provides low-income city enlarged again 4. Thе council provide was in 1967 separate an amended to for the elec- city providing amendment to the mayor. Apparently, charter for six tion of a neither of these council members. steps challenged by the charter appellants. was is op- Thornburg poor, the medical services essential factors. and con- transients a shelter
erates Focusing the dilution threshold vote Forty-two center. job training structed Thornburg, enumerated in the factors recreation city’s parks and percent of city various council election court reviewed city percent of all forty-five through and and ana- facilities results from 1965 three of Aus- lyzed precincts in the statistics from 52 Austin are located facilities concerning involving elections minorities having high- eight fiscal districts tin’s city ran for council.5 The court found who populations. est city council elections in which a minori- composition ethnic during the 1960’s ty candidate ran were approxi- force Austin’s work was city of racially polarized. (compared to mately black 12.2% 19% (com- Hispanic population); total 20.6% compactness. geographical A. and Size pоpulation), total pared to 18.7% analysis requires first approximately Anglo (compared to 59.1% court to determine whether the minori- Minority workers population). in total 68% geo- ty population sufficiently large extensively in administrative employed compact majori- graphically to constitute city, numbers and their positions with single-member Appellants ty in a district. categories have in such percentages embodying either a six- proposals furnished addition, city 1975. In increased since mayoral plan or an one council and election promoted business effectively has eight-one plan expanded which through contract awards. enterprises its agreed The court two members. Furthermore, in- County, Travis proposal, proposed that in the six-one Austin, enjoys and Mexican- black ap- cludes contain an largely district would black legis- representation population, the state but a proximately American 50.7% population voting age black Independent School 45.8%. The Austin lature. population would The Mexican-American one at least one black and District has had 13.9%, of which were estimated be 12% Hispanic member since voting age. eight-one plan, In the *7 voting age population of and a black 60.6% 5. Tenuousness. 57.2%, could occur in population of black city’s reasons The court found that the plan type In neither of would one district. maintaining at-large place re- voting age its command a Mexican-Americans membership eight- are although in an quirements majority, one district single Mexi- proposal the of include a tenuous. Each time issue one would 57.5% voting a population, with placed on the can-American districts has been member The Mexican- population of age 49.8%. ballot, recently in Janu- it has failed—most specific a electoral dis- American claim to defen- The court credited the ary 1985. appellants inseparable, is as trict therefore at-large system the dants’ evidence found, from its acknowledge and the court access, particularly minority be- enhances for Thorn- politically claim to be cohesive community the is substan- cause population. the black burg purposes with throughout city. It tially dispersed the carry appellants did not concluded that neither the black noted thаt The court establishing plaintiffs that an at- had sub- their burden of nor Mexican-American plan in would maintained on a tenuous six-one which large system is mitted a the vot- a of separately constitute pretext for discrimination. as a basis can-American, they contained a 60- recognized judge inherent because the 5. The district pre- relying generated population. from Nine on results limitation 89% Mexican-American votes, proportion of total but population a small the containing such he observed that no black a 50-85% cincts black, other more reliable method the predominantly were identified estimating breakdown of the vote the ethnic Anglo greater sample a or rest in the had 90% Among precincts, provided. the had population. predominantly Mexi- were classified as seven contrast, signifi- the rated as proposed By dis- court ing age population in the concerning the cant the statistics upon various com- tricts. The court relied frequency with which blacks and Mexican- Thornburg pointed ments which supported the same candidates. Americans voting age majority of requirement of a aptness comparison The of such a was proposed single-mem- minority voters in a Thornburg’s drawn from comment that “a ap- The court discounted ber district. showing significant mi- that a number of fig- pellants’ request that the 1980 census nority group usually members vote for the updated ures to include all members way proving same candidates is one minority population who were eleven necessary cohesiveness to a vote eligible in their among older claim_” dilution 478 U.S. at concluded, court also how- district. The at 2769. Between S.Ct. ever, permits it to order an preferred Mexican-American candidate eight expansion city council to mem- supported by average was an guarantee 69% remedy that would bers—a voters, pre- Austin’s and the black black majority voting age population dis- supported by ferred candidate was an aver- permit adoption of a district trict and would age of of Mexican-American voters. Mexican-Americans, 73% although not in which however, significantly, More in those same majority, aggregated could be elections, Anglos voted with one or the possi- This blacks to achieve such a result. minority groups other frequently more bility led the court to consider whether Thus, groups together. than the two voted politically Hispanics blacks and cohe- Anglos against in elections where mi- ran according Thornburg’s sive second supported norities and a different candi- prong. date did groups, than minori- together ties voted five times and voted B. Political cohesiveness of the differently result, four times. This groups. found, hardly court evidence of cohesive- The court found that blacks and Mexi- ness.6 are, groups, can-Americans as individual compared court also the results of a politically cohesive. black, 1987 race for Place 1 in agreed The court that the black and Mex- Davis, Terry Mexican-American, and a Gil- ican-American communities share common Martinez, bert were candidates. concerns and have common histories of dis- race, Davis received vote 12.1% not, however, crimination. It did rate as precincts, Mexican-American to Martinez’s conclusive the affidavits and precincts In the black Martinez 46.7%. *8 depositions politically from active Austin vote, garnered compared of the 10.8% to perceptions citizens that reflect their of the Davis’s The сourt held 63.5%. that such cohesiveness minority between the two support evidence likewise failed to a find- First, groups. the court noted that such ing of cohesiveness between blacks and perceptions up by survey were not backed Mexican-Americans. The court considered or other demonstrative evidence to rein- likely it large that a district with a Second, force these conclusions. the court minorities, percentage of both black and testimony found such counter-balanced oppose Mexican-American candidates will city office, affidavits submitted on behalf of the so, of each other for and if did the any systematic politi- Austin which denied Anglo residual vote would determine the cal cohesiveness. outcome of the election. minority
6. Between supported 1983 a was a can- Mexican-Americans a different can- Anglos supported didate in Anglos. Significantly, 19 races. the minor- didate than did the court Thus, ity eight races, of those 19 races. in those found that in four of these nine the black- races, eight city cohesively. supported majority the entire voted candidate did not receive a races, minority two of the 19 the candidate did of Mexican-American votes and the Mexican- plurality support not American-supported receive of either communi- candidate did not receive a races, ty. remaining majority In the nine blacks and of black votes. support voting. bloc mitted to a Section 2 claim Evidence of white when C. racially polar- there is otherwise a lack of its discus- The district court commenced Austin, voting. noted, ized as the court prong third sion polarized voting found no racial since the not observing Supreme the Court did that “gentle- late 1960’s. The court found no type on what of evidence reach a consensus agreement” might man’s that minorities bloc-voting majority proves that a white prevail only in Places and 6 on the minority’s preferred the usually will defeat council, allowing polarization the racial opinion Brennan’s candidate. Justice remaining support the council seats. To flatly the stated four members of Court conclusion, in this the court or intent are irrelevant observed that that causation 1987, comparing voting the behavior of whites 1985 and Martinez Gilbert obtained a the minority, and a and that the race of majority Anglo precincts of votes in certain disa- candidate is irrelevant. Justice White running even while for Place 1. opinion. greed portion with this the Jus- Second, persuasive no found concurrence tice O’Connor’s four-member voting. evidence of white bloc Between although evidence of the rea- stated that winning 1975 and the candidate in voting patterns could not be sons behind Austin council pre- races was the black showing plaintiff’s to rebut a of racial used ferred candidate of the time and was 85.7% cohesiveness, such evidence is rele- preferred the Mexican-American candidate vant anglo time. Since 1975 80% voters answering question the whether bloc given from of their 47% vote 84% voting by consistently white voters will prevailed candidates who have minority candidates. evi- defeat Such city council elections. suggest dence would that another candi-
date, explained why The court in detail equally preferred by it did the group, might greater accept probative testimony able to attract the be support in future elections. appellants’ expert white Dr. Miller whose bivari- ecological regression analysis áte was a at at appellants’ polarization cornerstone regard- The district court concluded that assuming proof. Even that Dr. Miller’s less of which view of white bloc is results are reliable and indicate that some adopted, evidence failed to polarized, recent elections have been demonstrate it. found that those results do court not neces- pointed The district court first out that sarily that show white will in Thornburg all members of the Court usually vote as a to defeat the minori- bloc consistent, agree sustained ty preferred Additionally, candidates.7 presumptively electoral success is “incon- analysis Dr. court found Miller’s statistical allegation sistent with ... [an] seriously The court said: flawed. ability repre- to elect ... analysis highly of Dr. Miller sus- equal sentatives of their choice is not pect for several reasons. Dr. Miller used enjoyed by majority.” white different measures to determine Black at U.S. S.Ct. 2780. Since voting strength. and Mexican-American consistently repre- there has been a black *9 He failed to take into account the differ- council, city sentative on the and since 1975 voting pre- population ence in sizes of rеpre- has been a Mexican-American there analysis impossi- cincts. His resulted on the council. The court sentative instances]_ results some ble suggests Thornburg [in held that while addition, completely he failed to establish minority candi- the occasional election confidence for the results of his owing special level dates to circumstances such regression analysis.... Without such a opposition incumbency may as lack of measure, if way there is no to determine racially polarized rebut the existence of statistically significant. voting, per- such circumstances not be the results are See, e.g., supra, Dr. Miller’s data. 7. the court’s conclusions at n. 6 drawn from 62, short, accept racially polarized voting. 478 U.S. at court, felt it could not analysis at face value. at 2772-73. Justice White disa expert’s 106 S.Ct. portion opinion and greed with this of the
III. DISCUSSION
particularly took issue with its statement
virtually
preferred by
ev-
issue with
that the race of the candidate
Appellants take
legal
of the dis- minority
and
conclusion
is irrelevant to racial bloc
ery factual
83,
voting analysis.
trict court.
preferred candidates. We
servatively
uncertainty, by focusing
to this
appellants’
contention that Aus-
dress
on statistical evidence of election contests
plan
tin’s multimember district
was enacted
involving minority
by rely
candidates and
perpetuated
with an intent to discrimi-
results,
ing largely on election
rather than
minority groups.
against
nate
factors, in drawing
extrinsic
its conclusions
Racially
Voting.
polarization.9
on racial vote
A. Evidence of
Polarized
the absence
Supreme
guidance,
of further
Court
we do
perplexing aspect
A
most
methodology, although
not criticize this
it
is its discussion
what constitutes evi
only permissible wаy
approach
is not the
legally significant, racially polar
dence of
2 claims. Cf.
Haley,
Houston v.
voting.
prong
plurali
The third
ized
(5th Cir.1988),
F.2d 341
vacated on other
ty opinion requires proof that
the white
(5th Cir.1989).
grounds,
at
S.Ct. at 2767.8 Justice Brennan’s
(or
yielded proportional
greater)
has
repre
opinion for four members of the Court
sentation on the Austin
holds that statistical evidence that different
council for over 10
differently, by
years,
races
racially
vote
itself—without
refuted the contention of
proof
polarized voting.10
of causation or intent —establishes
Earlier,
opinion explained
significance
consistently
8.
Miller's results
with Justice O'Con-
electoral results on the vote dilution
nor's admonition that factors other than race
process as follows:
might be relevant in the overall decision wheth-
Consequently,
difficulty
electing
if
usually
er white voters will
defeat
can-
proven, minority
white bloc
are not
didates.
voters have not established that the multi-
ability
member structure interferes with their
connection,
properly
10.In this
court
the district
preferred
to elect their
candidates.
held that consistent
electoral success
49 n.
539 results,” reject force into their we cannot the court’s find- Attempting to breathe ing polarized voting elections have opinion clearly that Austin of a lack of expert’s appellants as- racially polarized, erroneous.12 been analysis precise- used Dr. Miller’s sert that properly recognized The trial court also methodology as that of the
ly the same
the limitations inherent in Dr. Miller’s dif-
level. Bivar-
expert at the trial
Thornburg
fering
composition
of the
measures
ethnic
regression analysis was
ecological
iate
precincts. Explaining
of
that his base data
that does not
by
experts, but
used
both
him,
supplied
were the most reliable
Dr.
reliability of the
to establish the
suffice
extrapolated
Miller
the number of black
expert in this case. As
of the
conclusions
precinct
in a
from census data on
recounted,
the district
have earlier
we
population,
regard
age
without
suspect pri-
court found those conclusions
registration.
levels or voter
He estimated
expert’s
failure
marily because
voters,
the number of Mexican-American
coefficients
meaningful correlation
achieve
hand, from the
of
on the other
number
measures of
use of inconsistent
and his
Spanish
precinct
regis-
surnames on
voter
strength.
minority voter
tration lists. These estimates tend to over-
eco-
criticisms are valid. Bivariate
Both
state
vote results unless
has
fre-
logical regression analysis
high
higher
is as
or
than that
voter turnout
after
employed in Section
cases
quently
precinct.
This
of the rest of
overstate-
See,
Campos
City
v.
Thornburg.
e.g.,
of
magnified by
ment is
Dr. Miller’s calcula-
Texas,
supra; Citizens For
Baytown,
Anglo
tion of the
vote as the remainder
Gretna, La., 834
City
v.
Better Gretna
of
subtracting minority
from total
after
votes
Cir.1987),
(5th
F.2d 496
reh. denied
precinct. Consequently,
in
votes
each
Cir.1988).
(5th
to the
Crucial
F.2d 1471
tend to inflate the level of
these measures
analysis are the val-
validity
regression
minority support for a candidate.
this
“r2”,
for “r” and
which measure
ues
case,
expert’s computa-
a number of the
and linear rela-
strength of the correlation
difficulties, yielding
tions reflected such
examined,
being
in
tionship of the variables
figures
support
of Mexican-American
and the
this case the race
the voter
and,
Ang-
in
over
some instances
100%
supports.
candidate he
Mexican-Americans, support levels
los and
case,
“r”
Miller calculated
In this
Dr.
candidates.
of less than
for various
0%
for the
mainly between 0.5 and 0.7
values
Although
agree
perfection
we
that absolute
polarized,
plaintiffs
claim are
43 raсes
statistical data is not to be
the base
only
6 cases.
“r” values above 0.8
ignore
expected,13a trial court should not
“r”
Consequently, squaring these
values
nor the
imperfections
of the data used
dem-
coefficients of determination
calculate
analysis.
limitations of statistical
explains
race
more than half
onstrates that
trial court’s ref-
Appellants also fault the
six of the
of the variance
results from the 1985 and
erence to
plaintiffs.11
con-
races relied on
involving
Martinez as evi-
races
Gilbert
text,
of the district court’s
view
refuting
alleged
polariza-
racial
dence
methodology con-
finding that “Dr. Miller’s
court,
Anglo
The
how-
vote.
practices that
tion
many questionable
so
tains
ever,
city’s “extreme case”
weight in the
relied on the
place
cannot
much
the Court
analyst
precincts,
assumes that the
expresses
percentage
then
11. The "r2" value
explained
according
precincts
the race
to their
variance in the vote that
voters in most
voted
assumptions
of the voters.
ethnicity.
piling
on of
sophisticated
easily
statistical
prey
fall
to more
significance
upon
the statistical
12. Insistence
Wildgen, Adding
errors.
regression analy-
ecological
results in bivariate
Ecological Fallacy
Parameter
Thicket: The
important
in-
particularly
because of the
sis is
Cases, Urban Law.
Control in Vote Dilution
methodology. Briefly
limitations of that
herent
stated,
(1988).
upon
premises
facts
it
correlations
upon
but
as-
individual voter behavior
about
Edmisten,
F.Supp.
Gingles
at 368.
13. Cf.
precinct
sumptions usually developed
elec-
from
composition
Estimating the ethnic
tion returns.
*11
analysis in con- B.
Intentional Discrimination.
“homogeneous precinct”
extremely
performed
cluding that Martinez
Appellants alternatively contend that the
Anglo pre-
predominantly
inwell
certain
inaugurated
at-large
City of Austin
its
ma-
suggest that the
did not
cincts. The court
place
jority
system with the intent to dis-
the
precinct percentages reflected
overall
against
criminate
and Mexican-
blacks
analysis
limita-
type
This
vote.
They
Americans.
contend that the court’s
—the
recog-
expressly
tions of which
findings regarding the 1953 charter revi-
Voting
utilized in other
nized—has been
present system
the
sion which introduced
See, e.g., Gingles
Rights
erroneous;
Act cases.
v.
clearly
findings
are
that those
29;
Edmisten,
F.Supp. at 367-68 n.
according
appro-
not
to the
were
evaluated
standard;
Gretna,
legal
priate
Better
834 F.2d
and that where dis-
Citizens For a
criminatory
underlying
intent
a statute’s
496;
Baytown, 840 F.2d
Campos City
v.
original
proven,
enactment is
it is unneces-
1240.
that,
later,
sary
years
to demonstrate
the
Simply put,
the district court’s
discriminatory
statute continues to have a
findings regarding
presence
racially
reject
effect in order to invalidate it. We
polarized voting
clearly
were not
errone
these contentions.
repeatedly
ous. Austin has
elected black
evaluating
When
the historical back
and Mexican-American council members
ground
city’s
revision,
of the
1953 charter
during
past
years.
according
Even
the trial court did
Village
not cite
Ar
voting
analysis,
winning
lington Heights Metropolitan Housing
frequently
candidates
received Development Corporation,
252,
429 U.S.
(50%)
fifty percent
Anglo
well over
555,
(1977).
97 S.Ct.
those
find,
we would have to
267-68,
course,
564-65. Of
contrary
court,
to the district
out,
City points
as the
Arlington Heights
product
slating process
were
of a
suggest
does not
these factors are
essentially
rigged Austin’s
elections
exhaustive.
541 222, 228, possible impact change on wood, “minori- case, (1985). 222 In this L.Ed.2d findings 85 ties.” The court’s are shielded however, the record of sustained clearly erroneous standard of review. city on Austin’s representation simply We are not left with a definite and finding present- a decade belies a over wrong. firm conviction that impact the num- discriminatory from day Appellants’ evidence of intentional dis bered-place system. fruitfully crimination in this case governs Arlington Heights if Even compared with the evidence of discrimina case, findings court’s were this the district in tion Hunter v. Underwood and Carroll conclu inadequate not to reach its ultimate Stallings, ton Branch C.P. v. of N.A.A. the 1953 charter revision did not sion that (11th Cir.1987). F.2d The latter discrimination. result from intentional presented challenge case a Section to a at the findings, Those recited more detail single-member county up in commission set opinion, implicitly treated outset of this representative in 1951. The local Alabama by Arlington each of the factors listed sponsored effecting who had bill Heights. decrease from a commission three-member findings describe the The trial court’s quoted had a related context as background as historical of the revision saying, country “This is a white man’s general the Austin part of a overhaul of keep way.” we must it that 829 F.2d at city charter. The revision to a numbered- Likewise, Hunter, counsel for the place system of council member elections virtually state defendants conceded that committee, proposed by a citizens’ on was challenged statute was enacted with represented, which racial minorities were disenfranchising the intent of blacks. Thirty- year after more than a of effort. 231, 105 U.S. at S.Ct. at 1921. The law had re- propositions two were voted on. Other put passed part of visions to the voters at the same time been the state’s changes in gamut ran the from the council- constitutional convention in which a “zeal manager prohibition to a on city relations supremacy rampant.” for white ran Id. having an interest in council members’ Here, equivocal concerning the record is contracts, changes municipal annexa- adopting majority-vote, the motivation for tion, taxing personnel, policies. While Thus, place charter revision.16 numbered acknowledged proposals that the interpreta district court’s ambit for charter revision followed on the heels tion of the evidence was wide. city council election in which a of a eighth finished in a field of four- candidate competing for six council IV. CONCLUSION
teen candidates
positions, the court found no other evidence
Voting
A
2 of the
claim under Section
hoc,
support
post
propter
discrimination
Rights
questions
Act raises serious
about
The court did not consider that two or
hoc.
integrity
of a
subdivision’s
contemporary
allowing
three
references to
case,
governance.
In this
the evi-
mode of
city govern-
“minorities” to take control of
parties
dence offered
both
was volumi-
probative,
such termi-
ment were
because
arguments
presented.
nous and the
well
equivalent
nology was not at that time an
conscientiously
The district court
its turn
Additionally,
City
minorities.
for racial
foregoing
the case. For the
rea-
evaluated
contemporary
arti-
offered numerous
news
sons,
judgment of the district court is
thirty-two charter
cles which describe the
alluding
AFFIRMED.
revisions in detail without ever
already
additionally suggest
repetitious.
testified
Appellants
was
Wendler had
that the dis-
by excluding testimony
their
length,
trict court erred
in evidence
the court had received
Wendler, Sr.,
pro-
era,
witness Ed
who would
regarding
many
articles and exhibits
vided-background on the attitude of "the whites"
testimony. We can-
and it had heard other live
during
on charter revision. The
the discussions
say
its discretion.
the court abused
testimony
grounds
that it
court excluded this
partic-
2 is to facilitate
of amended
JONES,
Judge,
*13
and
Circuit
EDITH H.
pro-
in our
ipation by minorities
concurring:
of their
cesses, by preventing dilution
rea-
and
panel’s
our
result
I concur with
voting age persons can vote.
Only
votes.
also
they go,
I would
soning
but
as far
Pyrrhic victory for a court to
It
a
would be
analysis of
the trial court’s
have addressed
district in
single-member
create
factors:
the size
Thornburg
the first two
absolute,
minority population dominant
mi-
geographical compactness
and
numbers,
age
continued
but
political cohesive-
nority groups, and the
polls. Thornburg
to
defeated at
be
groups.
racial
these two
ness between
fact:
implicitly recognized this
decision that
light
our court’s recent
of
possess
po-
Unless
Rights
applies
Act
Voting
2
Section of
representatives in the
to elect
tential
Edwards,
elections,
judicial
v.
Chisom
challenged structure or
absence
(5th Cir.1988),together with
F.2d 1056
839
they
claim to have
practice,
cannot
round
another Census-related
the advent of
injured by
practice.
redistricting,
would be
legislative
of
we
17,
at 50-51 n.
pellants
the second of these three League
United Latin American Citi
refinements.
Dist.,
Indep.
812 F.2d
zens v. Midland
Sch.
(5th
appellants’ propositions
The
are ill-found-
vacated en banc
Cir.1988) (Brown,
dissenting) (arguing that
expert testimony as to statistically signifi-
the “r” value became analysis of the elec- regression for the
cant correlation coeffi- yielding
tion the lowest that, careful to note
cient. Our court was race, found Campos
“[f]or .52, slope of .33 and
that the “r” value of significance .0456 were the lowest val- MOORE, al., Rachel et but, introduced as testified ues [the Plaintiffs-Appellants, Brishetto, they still plaintiffs’ expert] were significant.” Campos, statistically ought encourage dis- F.2d at 1247. We MISSISSIPPI VALLEY STATE UNIV- litigants pay alike to close trict courts and ERSITY, al., Defendants-Appellees. et “r” signifcance val- attention to No. 87-4830. regression analysis. generated by Sta- ues tool, especially tistics are a mischievous Appeals, Court of United States court, prepared if are not and offered Fifth Circuit. *16 convincing explanation of their with a 1,May and limita- mathematical characteristics customary here of lev- tions. absence rightly ac- reliability
els of statistical was
knowledged by the district court.
CONCLUSION
Previous cases and the above discussion propositions regarding the
establish three Thornburg prongs
first two that should be First, recognized by
explicitly our circuit. single-
voting age majority proposed in a district, population rather than
member
majority, a mi- should determine whether sufficiently
nority group is numerous.
Second, Thornburg analysis should be according existing
conducted number subdivision, in a
of districts
according larger to a number tailored to single-member
accommodate a district ra- Third, majority. although
cial we
now held that two racial minorities jointly politically cohe-
prove that Thornburg, under this factual determi-
sive suggest
nation in one or two cases does not every Mexican- case blacks and satisfy
Americans will their burden
proof. Finally, should insist not courts generat- “r” provision on the values regression analyses, but also on
ed “r”
evidence as to the level at which those significant. become should
values
Courts
2.Even
if total
be considered
notes
"...
minority group
graphically compact,
so
is
eight-member city
yard-
and an
surrounding
popu-
white
violation,
small
relation
appellants’ population
stick for a
2§
in a
lation that it could not constitute
single-member
figures
yield
barely
still
districts that
assure the
district,
these
electing
possibility of
a black council member.
they
would have been able
cannot maintain
population
The
does not
Mexican-American
representatives of their choice in the
to elect
generate
potential
eight-
itself
such
even in an
struc-
absence of the multi-member electoral
observes,
City
plan.
member district
U.S. at 50 n.