REVISED APRIL 1, 1999
UNITED STATES COURT OF APPEALS
FOR THE FIFTH CIRCUIT
_______________________
No. 98-50227
_______________________
ROBERT VALDESPINO; BRENDA ROLON,
Plaintiffs-Appellants,
v.
ALAMO HEIGHTS INDEPENDENT SCHOOL DISTRICT; ETHYL WAYNE, In her
official capacity as a member of the Board of Trustees of the
Alamo Heights Independent School District, San Antonio, Texas;
HARRY OREM, In his official capacity as a member of the Board of
Trustees of the Alamo Heights Independent School District, San
Antonio, Texas; STEPHEN P. ALLISON, In his official capacity as a
member of the Board of Trustees of the Alamo Heights Independent
School District, San Antonio, Texas; ANNE BALLANTYNE, In her
official capacity as a member of the Board of Trustees of the
Alamo Heights Independent School District, San Antonio, Texas;
THOMAS A. KINGMAN, In his official capacity as a member of the
Board of Trustees of the Alamo Heights Independent School
District, San Antonio, Texas; TERRI MUSSELMAN, In her official
capacity as a member of the Board of Trustees of the Alamo
Heights Independent School District, San Antonio, Texas; VICKI
SUMMERS, In her official capacity as a member of the Board of
Trustees of the Alamo Heights Independent School District, San
Antonio, Texas,
Defendants-Appellees.
Appeal from the United States District Court
for the Western District of Texas
March 11, 1999
Before JONES, SMITH, and EMILIO M. GARZA, Circuit Judges.
EDITH H. JONES, Circuit Judge:
The panel hereby withdraws its previous opinion and
substitutes the following:
Plaintiffs-Appellants claim that the at-large, by-place,
majority-vote elections for positions on the AHISD board of
trustees dilute their votes as Hispanics in violation of Section 2
of the Voting Rights Act of 1965. See 42 U.S.C. § 1973 (1994) (as
amended). The district court found that the Plaintiffs failed to
make out a vote dilution claim because they could not prove, under
the first Gingles threshold factor, that Hispanics are a
“sufficiently large and geographically compact [group] to
constitute a majority in a single-member district.” Thornburg v.
Gingles, 478 U.S. 30, 50, 106 S. Ct. 2752, 2766 (1986). All the
issues on appeal involve proof of the first Gingles factor. In
particular, we reject the appellants’ contention that a “majority”
may be less than 50% of the citizen voting-age population. As
appellants’ other contentions fare no better, the judgment is
affirmed.
I.
The School District conceded at trial that the
Plaintiffs’ demonstration district1 did comprise a majority of
Hispanic voting-age citizens according to 1990 census data. The
1
The “demonstration district” is the hypothetical single-
member district used by voting rights plaintiffs to demonstrate
that they can satisfy the first Gingles factor (i.e., that their
group could constitute a majority in a single-member district).
Because the AHISD Board of Trustees has seven members, the
Plaintiffs must propose a demonstration district that would be
appropriate if the at-large district were divided into seven
single-member districts.
2
School District, however, presented evidence that demographic
changes between the 1990 census and the 1997 trial had eliminated
that majority. AHISD is a small district in which a few strategic
land-use changes could and did significantly alter the district’s
population and neighborhood ethnic mix.
The School District’s evidence was presented in expert
testimony by Dr. Bill Rives, a demographer. Using the 1990 census
data as a baseline, Rives investigated post-1990 changes in the
school district’s housing stock to determine how the population had
changed in the Plaintiffs’ demonstration district and in the school
district at large. He testified that this methodology is “by far
the most popular demographic estimation technique” and is
especially appropriate for small areas.
Rives testified that two main trends combined to leave
the Plaintiffs’ demonstration single-member district
“underpopulated” in 1997. Since 1990, the demonstration district
had lost population (and the proportion of Hispanics in the
demonstration district declined) because a large apartment complex
had closed, been renovated, and reopened with a smaller number of
residents. Simultaneously, the population of the school district
at large had increased because of substantial new residential
development in the Lincoln Heights area (formerly a quarry and
cement plant), outside the demonstration district. As a result of
these changes, the Plaintiffs’ demonstration district no longer
approached one-seventh of the school district’s population, and
thus could not be a proper single-member district.
3
To correct for the underpopulation, Rives added territory
to the demonstration district. He added a contiguous area to the
north that had been included in some of the appellants’ prior
proposed demonstration districts. That northern area ran clear to
the edge of the school district and had just about the right number
of people to make a proper district. Furthermore, if it were not
added to the Plaintiffs’ district, the northern area would have to
be attached to a different district via a mile-long, narrow strip
of unpopulated land. After the northern area was added to the
demonstration district population, Hispanics made up only 47.9% of
the voting-age citizen population of the revised demonstration
district. Even if the demonstration district were then partially
depopulated (by 8.1% of the ideal population), this number would be
48.3%.
The general thrust of Rives’s testimony had been clear
for some time before trial. For example, in a November 1996
affidavit attached to the Defendants’ Response to Plaintiffs’
Motion for Partial Summary Judgment, Rives had explained that no
additions to the Plaintiffs’ proposed demonstration district could
sufficiently increase the Hispanic population because “[t]here are
no Hispanic-majority blocks that are adjacent to the proposed
district.” As the trial grew nearer, Rives updated his analysis,
redoing calculations based on more recent data from the apartment
complexes about the ethnicity of their residents. Ultimately, the
district court relied on Rives’s April 1997 Report.
4
Much of the controversy in the case comes from Rives’s
August 1997 Report. In July 1997, updated rental rolls became
available from The Reserve, the large apartment complex within the
Plaintiffs’ demonstration district that had been renovated and
reopened since the 1990 census. Rives then recalculated the
results in his April 1997 report after learning that The Reserve
had a slightly higher proportion of Hispanic residents than he had
earlier believed. But he also corrected a calculation error in his
April Report that had understated population growth in the school
district at large. Nevertheless, he concluded that the revised
data still did not yield a majority of Hispanics among voting-age
citizens within the Plaintiffs’ demonstration district. The August
Report was given to the Plaintiffs in early August 1997, not long
before the original trial setting. On August 22, however, the
district court reset the trial for Monday, September 15.
The Plaintiffs’ proffered expert witness, George Korbel,
claims that he was surprised by the conclusions in Rives’s August
Report. In response, Korbel scrambled the week before the
September trial date to conduct a door-to-door survey of the
residents in a small area to the south of the demonstration
district. He thought he could find there a high proportion of
Hispanic residents that could increase their demonstration
district’s population without diluting its Hispanic majority. At
4:21 P.M. on Friday, September 12, the Plaintiffs faxed to the
School District’s counsel a letter disclosing the existence of this
5
new survey. At 4:13 P.M. on Saturday, September 13, the Plaintiffs
faxed the data from the survey.
On the Monday morning set for trial, September 15, the
School District filed a motion to strike the survey on grounds of
unfair surprise. The Plaintiffs’ lawyer told the district court
that their case in chief would rest entirely on 1990 census data,
but that if Rives testified for the School District that more
current data changed the Hispanic majority, then the Plaintiffs
might use the recent survey as rebuttal testimony. The district
court postponed until rebuttal any ruling on the motion to strike
and granted a motion in limine to prevent mention of the survey
during the case in chief or cross-examination. During the
Plaintiffs’ rebuttal, the School District renewed its objections to
the survey evidence, and the district court granted the motion to
strike. The Plaintiffs filed an offer of proof as to what their
expert witness would have testified about the survey.
In its findings of fact and conclusions of law, the
district court reiterated that Korbel’s survey constituted unfair
surprise and was excluded under Local Rule CV-16(e). To
accommodate the Plaintiffs’ objections to the lateness of Rives’s
August Report, the district court decided to rely solely upon the
April Report, which it found to be “thoroughly documented, [with]
a high degree of accuracy,” and “clear, cogent, and convincing
enough to override the presumptive correctness of the prior
decennial census.” Relying on Rives’s report, the district court
found that the Plaintiffs had not proved a demonstration district
6
with less than 10% population deviation that included more than 50%
Hispanics among its voting-age citizens.
On appeal, the Plaintiffs present three arguments: that
they were not required to meet a “bright line” test of 50% Hispanic
voting-age citizens in their demonstration district; that the
School District’s evidence did not adequately overcome the presumed
accuracy of the 1990 census data; and that the district court
abused its discretion in excluding Korbel’s proposed rebuttal
testimony about the last-minute, door-to-door survey.
II.
This court reviews district court “findings on the
Gingles threshold requirements for clear error.” League of United
Latin Am. Citizens v. Roscoe Indep. Sch. Dist., 123 F.3d 843, 847
(5th Cir. 1997). See also Gingles, 478 U.S. at 77-79, 106 S. Ct.
at 2780-81. A district court’s refusal to allow an expert to
testify as a rebuttal witness may be overturned only for abuse of
discretion. See Tramonte v. Fibreboard Corp., 947 F.2d 762, 764
(5th Cir. 1991); Bradley v. United States, 866 F.2d 120, 124 (5th
Cir. 1989).
III.
The Supreme Court has established a three-part threshold
inquiry when a racial or ethnic minority group asserts that its
distinctive votes have been submerged by the racial majority in a
multimember legislative district. The minority group must be able
to (1) “demonstrate that it is sufficiently large and
geographically compact to constitute a majority in a single-member
7
district,” (2) “show that it is politically cohesive,” and (3)
“demonstrate that the white majority votes sufficiently as a bloc
to enable it -- in the absence of special circumstances ... --
usually to defeat the minority’s preferred candidate.” Gingles,
478 U.S. at 50-51, 106 S.Ct. at 2766-67.
The Plaintiffs here argue that the district court erred
in applying the first Gingles factor as a “bright line”
requirement. The Plaintiffs advert to the general purpose of the
Gingles factors, which is to provide a framework for showing that
there could be “a single-member district in which they could elect
candidates of their choice.” This is intended to support the
proposition that the Plaintiffs need only show generally their
electoral potential. The Plaintiffs further argue that the Supreme
Court has disavowed “mechanical[]” application of the Gingles
factors.2 And they complain that the district court did not
evaluate evidence of vote dilution under the totality of the
circumstances test.
All of these complaints are baseless. In reality, this
court has interpreted the Gingles factors as a bright line test.
Each factor must be proved before it is necessary to proceed to the
2
In Voinovich v. Quilter, the Supreme Court did say, “the
Gingles factors cannot be applied mechanically and without regard
to the nature of the claim.” 507 U.S. 146, 158, 113 S. Ct. 1149,
1157 (1993). The Court did so, however, because it was adapting
the Gingles test, which was designed for challenges to multimember
districts, so that it could be used for challenges to the packing
of minority voters into existing single-member districts. Because
that changed context is not relevant to this case, which challenges
a multimember district, plaintiffs have no need to invoke non-
mechanical application of the Gingles factors.
8
totality of the circumstances test. We have repeatedly disposed of
vote dilution cases on the principle that “[f]ailure to establish
any one of these threshold requirements is fatal.” Campos v. City
of Houston, 113 F.3d 544, 547 (5th Cir. 1997); accord Rangel v.
Morales, 8 F.3d 242, 249 (5th Cir. 1993); Overton v. City of
Austin, 871 F.2d 529, 538 (5th Cir. 1989). See also Growe v.
Emison, 507 U.S. 25, 40-41, 113 S. Ct. 1075, 1084 (1993) (“Unless
[the three Gingles factors] are established, there neither has been
a wrong nor can be a remedy.”).
Furthermore, contrary to the Plaintiffs’ suggestion, this
court has required vote dilution claimants to prove that their
minority group exceeds 50% of the relevant population in the
demonstration district. In Gingles, the Supreme Court required
plaintiffs to demonstrate “a majority.” 478 U.S. at 50, 106 S. Ct.
at 2766. Both of the Fifth Circuit cases cited by the Plaintiffs
assumed that 50% was the threshold for “majority” and simply
addressed what evidence could be used to prove that the 50%
threshold was met. In Brewer v. Ham, the court acknowledged that
a super-majority of black residents could be used to prove that
blacks constituted a majority of voting-age residents. See 876
F.2d 448, 452 (5th Cir. 1989) (citing cases with raw super-
majorities of 65.9%, 71.5%, and higher). In Westwego II, this
court repeated Brewer’s holding and expanded on it in a footnote.
The footnote, much cited by the Plaintiffs, explained that those
plaintiffs “unable to produce hard data” on voting-age population
because of the way census data are collected and reported would be
9
able to submit “other probative evidence” to prove voting-age
population. See Westwego Citizens for Better Gov’t v. City of
Westwego, 906 F.2d 1042, 1045 n.3 (5th Cir. 1990). In context,
Westwego II’s statements did not alter what must be proved, only
what can be used to prove it. The Plaintiffs still must meet their
burden of proving that Hispanics constitute more than 50% of the
relevant population in their demonstration district.
Finally, this court has already determined what factors
limit the relevant population in the district: voting-age and
citizenship. This was made clear in Campos v. City of Houston, 113
F.3d 544, 548 (5th Cir. 1997) (courts “must consider the citizen
voting-age population” in evaluating the first Gingles factor).
See also Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368, 372
(5th Cir. 1999). Given that the Supreme Court has avoided the
issue of what population to use for the first Gingles factor,3 and
that other circuits have used the same approach as Campos,4 the
district court used the correct legal test here.
3
See Johnson v. De Grandy, 512 U.S. 997, 1008-09, 114 S. Ct.
2647, 2656-57 (1994); Growe v. Emison, 507 U.S. 25, 38 n.4, 113 S.
Ct. 1075, 1083 n.4 (1993).
4
See Negron v. City of Miami Beach, 113 F.3d 1563 (11th Cir.
1997) (using citizen voting-age population for first Gingles
factor); Romero v. City of Pomona, 883 F.2d 1418 (9th Cir. 1989)
(same), overruled in part on other grounds by Townshend v. Holman
Consulting Corp., 914 F.2d 1136, 1141 (9th Cir. 1990). Cf. Barnett
v. City of Chicago, 141 F.3d 699, 704 (7th Cir. 1998) (Posner,
C.J.) (using citizen voting-age population to determine
proportionality for Section 2 challenge to gerrymandering of
single-member districts), cert. denied sub nom. Bialczak v.
Barnett, 118 S. Ct. 2372 (1998).
10
IV.
In this case, evaluating the district court’s application
of that test involves two questions: whether the School Board’s
evidence was adequate to counter the Plaintiffs’ census data, and
whether the district court abused its discretion in excluding the
Plaintiffs’ proposed rebuttal evidence.
A.
Except for a cavil, the parties and the district court
essentially agree about what standard should be required to
overcome census data.5 As the district court summarized it:
[C]ensus figures are presumed accurate until proven
otherwise. Proof of changed figures must be thoroughly
documented, have a high degree of accuracy, and be clear,
5
The cavil is that Plaintiffs attempt to articulate a two-
step test: “The decennial census is controlling unless there exists
‘clear, cogent and convincing evidence’ that the decennial figures
are no longer valid and that other figures are valid.” For this
proposition, however, they cite only Garza v. County of Los
Angeles, 756 F. Supp. 1298, 1345 (C.D. Cal. 1990). In fact, the
Garza court specifically rejected the notion of a two-step test:
17. In order to overcome the presumption in favor of
the 1980 census data, plaintiffs need not demonstrate
that the census was inaccurate.
18. It is sufficient to conclude that there has been
significant demographic changes [sic] since the decennial
census and that there exist[] post-decennial population
data that more accurately reflect[] evidence of the
current demographic conditions.
Id.
11
cogent and convincing to override the presumptive
correctness of the prior decennial census.
This standard appears to be an elaboration on one used by the
Seventh Circuit. See McNeil v. Springfield Park Dist., 851 F.2d
937, 946 (7th Cir. 1988). Two Fifth Circuit cases are relevant.6
The first is Westwego II, which, as mentioned above, opened the
door to the use of non-census data when census data are not
sufficiently probative of the voting-age proportion of a
population. See Westwego, 906 F.2d at 1045 n. 3. The second is
Perez, in which this court affirmed a district court’s decision
that the plaintiffs’ population projections were too unreliable to
overcome 1990 census data. See Perez, 165 F.3d at 373. Based on
Westwego and Perez, the district court properly acknowledged the
persuasiveness of census data while admitting evidence that
demonstrated its inaccuracy in this case. Because the district
court found that the School Board’s 1997 population data overcame
the 1990 census figures, the question is whether that finding was
clearly erroneous.
6
The only Supreme Court authority on this matter is indirect.
See Karcher v. Daggett, 462 U.S. 725, 732 n.4, 103 S. Ct. 2653,
2659 n.4 (1983) (in reapportionment, a state cannot “correct”
census figures “in a haphazard, inconsistent, or conjectural
manner”); Kirkpatrick v. Preisler, 394 U.S. 526, 535, 89 S. Ct.
1225, 1231 (1969) (a state can consider post-census population
shifts in redistricting if its findings are “thoroughly documented
and applied throughout the state in a systematic, not an ad hoc,
manner”). The Ninth Circuit refused to apply the Seventh Circuit’s
“high standard” of “clear and convincing” evidence “in a case where
intentional discrimination has been proved, and the data is merely
to be used in fashioning a remedy.” Garza v. County of Los
Angeles, 918 F.2d 763, 773 n. 3 (9th Cir. 1990).
12
The Plaintiffs present a laundry list of purported
problems concerning the methodology of Rives, the School Board’s
demographics expert. The School Board’s responses as well as
Rives’s cross-examination at trial suffice to show that the
Plaintiffs’ challenges are generally misdirected, exaggerations of
hypothetical problems, based upon criticisms of assumptions that
played no role in Rives’s methodology, or based on the analysis in
Rives’s superseded 1995 reports. The Plaintiffs’ most emphatic
argument -- that Rives himself admitted his April Report was
“wrong” -- is overstated, because Rives did so only in the context
of explaining how the August Report was based on more current data
and corrected a calculation error. Rives’s admissions did not
affect the underlying finding of both the April and August Reports:
the Plaintiffs’ demonstration district did not contain a majority
of Hispanic voting-age citizens.
The general description of Rives’s methodology given
above reveals that the Gingles I issues in this case do not involve
any complicated statistical formulae or tests of significance that
might bedazzle or bamboozle an unwary district court. Cf. Overton
v. City of Austin, 871 F.2d 529, 544-45 (5th Cir. 1989) (Jones, J.,
concurring) (discussing some district courts’ ill-founded
assumptions about the levels at which correlation coefficients
become statistically significant). The data here were relatively
simple; their manipulation involved only rudimentary arithmetic.
Under these circumstances, the district court did not
clearly err in deciding that Rives’s report demonstrated sufficient
13
post-census demographic changes to erode the Hispanic majority in
the Plaintiffs’ demonstration district. In doing so, we take
special note of the School Board’s responses to the Plaintiffs’
three weightiest methodological criticisms, each of which the
district court could have credited without committing clear error:
(1) the housing stock methodology can be appropriate for
calculating population changes in small areas, (2) it was
appropriate to account for some apartment complexes by projecting
their imminent populations at the end of ongoing lease-up periods,
and (3) despite some lapses, the municipal power company’s records
of new electrical hook-ups were an accurate gauge of newly
developed housing in the entire school district. Further, the
School Board’s methodology was much more sophisticated than the
crude straight-line population projection that was rejected in
Perez. See Perez v. Pasadena Indep. Sch. Dist., 958 F. Supp. 1196,
1212-13 (S.D. Tex. 1997), aff’d, 165 F.3d at 373 (5th Cir. 1999).
B.
Even if the district court properly credited the School
Board’s post-census demographic evidence, the Plaintiffs argue that
it abused its discretion by excluding their proposed rebuttal
evidence about post-census populations.
The district court excluded any evidence from Korbel’s
last-minute survey “because it unfairly surprised the Defendants,”
citing W.D. TEX. R. CV-16(e), under which the district court may,
14
“upon the showing of good cause,” permit a party to supplement the
written summary of an expert’s proposed testimony.7
On appeal, the Plaintiffs offer two reasons why their
evidence was not an unfair surprise: (1) Korbel had testified in
his deposition that Hispanic population was available south of the
demonstration district; and (2) the survey was done in response to
“new methodologies and numbers” in Rives’s August Report and was
made available as soon as it was completed.
The Plaintiffs’ first reason fails to account for how
modern discovery handles expert witnesses. The Local Rule required
a “written summary of [Korbel’s] proposed testimony.” It further
required that summary to include “the basis of the opinions which
purport to be the testimony of the witness” and “specific
references to any exhibits that will be used by the witness in
support of any opinions.” W.D. TEX. R. CV-16(e) & note. It can
scarcely be maintained that Korbel’s cursory reference in a
deposition could serve as adequate notice of his intent to rely on
a door-to-door survey of a specific neighborhood. Nor could that
deposition response have provided sufficient information for the
School Board to prepare to cross-examine Korbel about the survey.
Cf. Sierra Club v. Cedar Point Oil Co., 73 F.3d 546, 571 (5th Cir.
7
Alternatively, in a footnote, the district court noted that
the methodology and execution of Korbel’s survey were too flawed
for the results to overcome the presumptive correctness of the 1990
census. Although it appears quite compelling, the School Board
does not press this line of argument, and we need not pursue it
since we hold that the evidence was properly excluded due to unfair
surprise.
15
1996) (initial expert disclosures not sufficiently “complete and
detailed” to meet discovery order).
The Plaintiffs’ second reason takes no account of the
fact that the district court relied only upon Rives’s April Report,
the admissibility of which the Plaintiffs never contested. If the
survey was made necessary only by the novelty of the August Report,
then apparently it could not have been detrimental to the
Plaintiffs to exclude both.
In sum, the court did not abuse its discretion by
excluding this evidence for unfair surprise when the proffering
party failed to meet its duty to supplement its expert disclosures.
See Alldread v City of Grenada, 988 F.2d 1425, 1436 (5th Cir. 1993)
(no error in excluding expert witness’ testimony when information
crucial to understanding it was not provided until two weeks prior
to trial); Mills v. Beech Aircraft Corp., 886 F.2d 758, 764 (5th
Cir. 1989) (proper use of discretion to exclude results from tests
conducted by plaintiffs’ expert the week before trial).8
8
The survey evidence was also unnecessary once the district
court excluded the District’s August Report. This satisfies the
first factor of a four-factor test that has sometimes been applied
in evaluating a district court’s exercise of discretion: “(1) the
importance of the witness’s testimony; (2) the prejudice to the
opposing party of allowing the witness to testify; (3) the
possibility of curing such prejudice by granting a continuance; and
(4) the explanation, if any, for the party’s failure to identify
the witness.” Bradley v. United States, 866 F.2d 120, 125 (5th
Cir. 1989). See also Sierra Club, 73 F.3d at 572 (using same four
factors in evaluating exclusion of evidence as sanction for
violating discovery order). The Plaintiffs would also appear to
fare quite poorly on the fourth factor, since it was obvious from
the beginning that the School District would present evidence of
1997 population. Neither party addresses the four-factor test on
appeal, though the School District discussed it in its original
16
Because it was not an abuse of discretion to exclude the
survey results on the grounds of unfair surprise, we need not
address whether Korbel’s survey would have constituted proper
rebuttal testimony.
V.
For the foregoing reasons, the district court properly
placed the burden on the Plaintiffs to prove a majority of
Hispanics among voting-age citizens in their demonstration
district; the district court did not clearly err in finding the
School Board presented sufficient evidence to prove demographic
changes since the census; and the district court did not abuse its
discretion in excluding the Plaintiffs’ proposed rebuttal testimony
for unfair surprise.
The judgment of the district court is AFFIRMED.
motion to strike.
17