Partial Draft of Brief Section III

    III. THE DISTRICT COURTIS FINDTNG THAT VOTING TS RACIALLY
POLARIZED IS CORRECT

In determining whether a method of election violates Section

2, a trial court must evaluate "the extent to which voting in
the elections of the state or political subdivision is racially
polarized." S. Rep. at 29. The Court below evaluated the lay

and expert testimony on this question and found "that within all
the challenged districts racially polarized voting exists in a

persistent and severe degree." J.S. App. 38a. Appellants

contend that this finding is erroneous as a matter of law.

The District Court articulated the relevant guestion as

whether a substantial enough number of white citizens are

unwilling to vote for black candidates that the polarization

operates, under the election method in guestion, to diminish the

opportunity of black citizens to elect candidates of their
choice. J.S. App. I4a-15a, 4la. This inguiry is plainly

consistent with the statutory language of S2.

The District Court, in accordance with the experts for
appellants and appellees, defined racially polarized voting as

the extent to which black and white voters vote differently from

each other in relation to the race of the candidates. J.S. App.

39a, n.29. T. 50, 1404. See also Citv of Rome v. United

States, 446 U.S. I56, IB3-187 (1980), affirming 472 F. Supp.

22L,226 (D.D.C. L97) ("Racial bloc voting is a situation where,

when candidates of different races are running for the same

office, the voters will by and large vote for the candidate of

their own race." This is the IegalIy correct inguiry, and the

trial judge's finding of the extent of racially polarized voting



is not clearly erroneous.

A. Summarv of the District Court:s Findinqs. The District
Court examined seven factors in determining that votingr was

severally racially polarized.

1. The testimony of numerous 1ay witnesses was that
black candidates could expect little support form the white

community in the form of endorsements, campaign contributions,
or votes. Icites including Malone f think] ftris testimonyr

credited by the local judges, J.S. 38a, was born out by the

statistical analyses of the two expert witnesses for appellees,

as follows.

2. Expert witnesses for appellants and appellees

agreed that the correlation coefficient is the standard measure

of whether black and white voters vote differently from each

other. T. 60, 1445.1 Correlations above an absolute value of

.5 are relatively rare. The correlations in this case had

absolute values between .7 and .98, with most above .9. J.S.

App. 39a, n.30.

3. The statistical significance of all correlation
coefficients in this case was less than.00001, meaning that the

probability of their appearing by chance was less than I in
100,000. J.S. App. 38a and n.30. Appellants' expert agreed

with this determination. T. L446.

lcorrelation coefficients rangle
stronger the correlation between the
race of the candidate voted for, the
of the correlation coefficient.

between 0 and + I.0. The
race of the v6ter and the
higher the absolute value

2-



4. Accepting the combined ecological regression and

extreme case estimates of appelleesr expert, Dr. Grofmanr2 the

court examined the percentage of white and black voters who

voted for the black candidates in each of the 32 election

contests ( 53 primaries plus general elections ) which had had a

black candidate during the three election years prior to the

triaI. J.S. App. 4Ia-46a.

The Court found thatr on the average, 81.7t of

white voters did not vote for any black candidate in the primary

elections, and "approximately two thirds of white voters did not

vote for black candidates in general elections even after the

candidate had won the Democratic primary and the only choice was

to vote for a Republican or no one." J.S. App. 40a. Given the

small percentage of black voters, the failure of this number of

whites to vote for black candidates presents a substantial

2AppeIlants' conceded that the method used by appellees to
assess the extent of racially polarized voting is standard in
the literature and that the statistical analysis performed by
appellantsr expert was done accurately, T. 77, 1445.

3-



barrier. 3

To make matters worse, because of the multiseat

electoral system, a candidate may need votes from more than 408

of the voters to win.4

For example, in the Wake County primary in 1980,

there \^/ere 9 candidates f or 6 seats. If the votes were spread

evenly and all voters voted a full slate, each candidate would

get votes from 6/9 or 672 of the voters. In such circumstances

it would take more than 672 of the votes to win.

5. Although, in primaries, black voters ranked black

candidates first or first and second, white voters almost always

ranked them last or next to the last. PX 11, App. 3. In

3These numbers must be examined in the context of the
particular election district in guestion. The lower the black
population of the district, the more white voters it takes
voting for the black candidate to make it possible for him to
win. For example, if 45t of the registered voters regularly
vote for black candidates, that polarization of voting may not
impede the ability of black citizens to elect candidates of
their choice.

Here, however, while there are a large number of black
citizens, because they are submerged into such large multimember
districts, they are a sma1l percentage of the total electorate.
For example, in House District #36 (Mecklenburg County), there
are 107,006 black residents, PX 4(b), more than enough for two
whole House Districts Icite to Stip. with ideal House District
sizel, but because they are submerged into an eight member
district, they are only 26.58 of the population. Because the
percentage of the registered voters in each of the districts
which is black is relatively Iow, ranging from 158 to 292, it
takes littIe polarization to impede materially the ability of
the black community to elect candidates of its choice.

4u.c.c. s. SI63-111 (a) ( 2 ) provides , "rf more candidates
obtain a majority then there are positions to be filled, those
having the highest vote (egua1 to the number of positions to be
filled) shalI be declared the nominees."

-4



general elections, white voters almost always ranked black

candidates either last or next to last in the multi-candidate

field except in heavily Democratic areas; in those latter,

"white voters consistently ranked black candidates Iast among

Democrats if not last or next to last among all candidates."

J.S. App. 40a.

The District Court determined how often the

candidates of choice of white voters and of black voters were

different. If white voters as a group are selecting different
candidates than black voters as a group, assuming black voters

are in a minority, the polarization diminishes the chances that

the black voter's candidate will be elected. T. 162-L63.

In describing this analysis the Court used the

term "substantively significant". J.S. 39a-40. The Court found

that in 30 of 32 elections analyzed, the black candidates who

were the choice of black voters were ranked last or near last
such that they lost among white voters. J.S. 40a, n.31.5

7. The trial court determined that in order to have

any chance of overcoming the barrier created by the

sAppellants mischaracterize this analysis asserting that the
District Courtrs finding of severe and persistent racially
polarized voting is in error because it was found "whenever less
than 40t of the white voters cast a ballot for black
candidates." Br. 36. This argrument misapprehends the multiseat
election system in North CaroIina, under which aIl candidates
run againts each other, and depending on the number of seats in
the district, that number of top vote-getters win. N.C.G.S.
51163-111, supra. The number of voters it takes to win depends
on the number of candidates, the number of seats, and how the
voters spread their votes. A voter could get votes from more
than 50t of voters and lose orr in some circumstances, votes
from less than 50t of the voters and win.

6.

5-



unr^rillingness of most white voters to vote for any black

{candidate, black citizens must single shot vote, thus forfeiting
their right to vote for a fuII slate of candidates. J.S. App.

4Ia. This finding was in accordance with the testimony of both

appellees' expert witness, T. 85, and appellantsr expert. T.

L437 .

B. Appellantsr corltention that racial polarization of

votinq has no leqal siqnificance unless it alwavs qausee blacks

to lose, A. Br. 35, 40, is erroneous.6 Appellants' novel

standard, not adopted in any case cited by defendant, is another

way of arguing that any minority electoral success precludes a

finding of a violation of Section 2, an argument which defies

the intent of Congress. See S. Rep. at 29, n.II5 and pp.

supra. 7

6fhe solicitor General does not adopt appellants' proposed
standard but articul-ates the inguiry as whether "the impact of
racial bloc voting in combination with the challenged procedure

here, multimember districts -- deprives black voters of equal
access to the electoral process..." SG Br. 3l-32. Assuming
that the Solicitor General includes with "egual access to the
electoral process ", as the statutory language of Section 2 does,
an egual opportunity to elect candidates of black votersl
choice, the Solicitor General does not disagree with the
District Court's conception of the guestion. The Solicitor
General simply disagrees with District Court's finding of fact
as to its ans\^rer.

Talthough this is an erroneous statement of the limit of
legalIy significant racially polarized voting, it is noteworthy
that racially polarized voting was sufficiently extreme that it
did cause the black candidate to lose in 658 (2L of 32) of the
election contests analyzed in which the black candidate received
substantial black support. Appellantsr statement that "two
thirds of all black candidates have been successful", App. Br.
45, is misleading since it only counts black candidates who made
it to the general elections and ignores the many black
candidates who lost in the primarj-es. [add to this note t of
whites who lost also possible move to extent of electorial
success section]

6-



Appellants support their argument that racially
polarized voting is legally significant only if outcome

determinative by citing Roqers v. Lodqe, 458 u.s. 613 (1992) and

two rower court cases, each decided after Mobire_v. Borden,

supra, finding discriminatory intent under the Fourteenth

Amendment. The level of raciarry polarized voting necessary to
support an inference of discriminatory intent under the

Fourteenth Amendment and the extent necessary to support a

finding of discriminatory result under 52 are not the same.

The reason that polarized voting is relevant to a

finding that an election system is being maintained for a

discriminatory purpose is that if polarized voting is extreme

enough that brack citizens never erect candidates of their
choice, then white elected officials can ignore the needs of the

black community with impunity, and that raises an inference that
they are maintaining the at rarge system for a discriminatory
purpose. See Nevett v. Sides, 571 F.2d 209, 2I7, 223, n.l6 (5th

Cir. L97B) (Nevett II) (holding discriminatory intent to be a

necessary element of a claim of unconstitutional vote dilution).
Thus, in Rogqrs v. -Lodge, the Court concluded that bloc

voting together with lack of erectorar success "bear heaviry on

the issue of purposeful discrimination. " Id. at

7-



( emphasis added ) . 8

WhiIe proof of absolute exclusion may be necessary to

raise an inference of discriminatory intent, it is not necessary

in order to show that black citizens have "less opportunity"

than do whites to elect candidates of their choice. The

statutory language of Section 2 is not limited to situations in
which black candidates have absolutely no chance of being

elected. 42 U.S.C. S1973(b). RacialIy polarized voting can

give rise to this unegual opportunity, even if it does not cause

black candidates to lose every single election.

Appellees know of no court which has adopted

appellants' proposed standard in a Section 2 case. Instead,

other courts have found polarized voting sufficient to support a

violation of Section 2, despite a finding of some electoral

success. In McMillan v. Escambia CounW, 748 F.2d 1037, 1043,

1045 (f lth Cir. I984 ) (t'Icr'IiIlan II ) , the Court f ound racially
polarized voting and a violation of Section 2 despite some black

electoral success, based on a finding that "a consistent

majority of the whites who vote will consistently vote for the

black's opponent. "

BAppellees do not agree that Rogers v. Lodge, supra, stands
for the proposition that racially polarization is legalIy
significant to a finding of discriminatory purpose only if it
uniformly causes electoral defeat. For example, in NAACP v.
Gadsden County_ggLreef_Board, supra, cited by appellants, the

i-=---:__finding of unconstitutional vote dilution was upheld despite the
election of one black candidate to the school board (electoral
success similar to that present here in House District #36 and
House District #2L). But this Court need not consider, in the
context of this case, whether appellants' bold assertion is
correct.

8-



As the Court noted in Major v. Treen,574 F.Supp. 325,

339 (s.o. La. I983) (three judge court):
Nor does the fact that several blacks have gained
elective office in Orleans Parish detract from
plaintiffsr showing of an overall pattern of
polarization... Racial bloc voting, in the context
of an electoral structure wherein the number of
votes needed for election exceeds the number of
black voters, substantially diminishes the
opportunity for black voters to elect the
candidate of their choice.

C. Appellees were not required to prove that white voters'
failure to votg for black candidates was raciallv motivated.

Iinsert 105-115 - edited - original attached]

9-



D. The District Courtrs findinq of the extent of raciallv
polarized votinq is not clearlv erroneous. Based on the

evidence summarized in Part III A, supra, the trial judges found

"that in each of the challenged districts racial polarization in
voting exists to a substantial or severe degree, and that in
each district it presently operates to minimize the voting

strength of black voters." J.S. App. 46a.

Appellants and the Solicitor General argue that the

District Court's finding is hrrong because it was based solely on

the determination that white and black voters chose different
candidates. IciteJ fhe Solicitor General argues that the

District Court used a definition under which any election in
which only 5It of whites declined to vote for a black candidate

would be held to be "Feverally racially polarized" U.S. Br. 29

(emphasis in original.) This argrument rests on an indepensible

misrepresentation of the language of the opinion below. The

guoted reference to differences in the preferences of black and

white voters appears at J.S. 39a, where the District Court

correctly notes the presence of the differences between white

and black voter choice. The term "severe" does not appear in
that punage at all, but is used on the next page in a separate

paragraph to describe elections in which 81.7t of white voters

declined to vote for any black candidate. (.f .S. 40a)

After the sentence guoted by the Solicitor
General, the opinion clearly states, "additional
facts revealed by this data support the ultimate
finding that severe (substantively significant)
racial polarization existed in the multimember
district elections considered as a who1e." J.S.
App.40a.

10



The opinion of the District Court clearly distinguishes

the presence of any difference between black and white voters

from a case in which white overwhelming opposed the candidate

preferred by black votes, and egually clearly characterizes only

the latter as "severe".

While it is true that a

not vote for any black candidate

even incumbents or those running

appellants assert, that District
on that observation.

majority of white voters did

under any circumstances, not

unopposed, it is not truer ds

Courtrs finding is based solely

The more complex evidentary issue answered by the lower

court was whether the degree of polarization was sufficiently

severe as to materially impair the ability of minority voters to
elect candidates of their choice.9 In concluding that such

impairment had been shown, the District Court noted:

On the average, 81.7t of white voters did not vote
for any black candidate in the primary elections.
In the general elections, white voters almost
always ranked black candidates either Iast or next
to last in the multi-candidate field except in
heavily Democratic areas; in these latter, white
voters consistently ranked black candidates last
or among Democrats if not last or next to last

-gwhfle 
appelfants ao not challenge the method appellees'

expert used to analyse the election returns in general, T. 77,
L445, appellants claim that appelleesr regression analysis is
flawed by what they labeled as the "ecological fallacy. " they
assert that instead of using turnout figures, appellees' expert
used voter registration figures. App. Br. 4I. Not only was
this argument made to the District Court and rejected, J.S. App.
39a, n.29, but also it is not accurate. Dr. Grofman did have
turnout figures for each precinct, and he used a regression
analysis to calculate the turnout figures by race. PX L2 at pp.
3-8. In fact, appellantsr expert, Dr. HoefIIer, admitted that
he did not know what method Dr. Grofman used to calculate
turnout, T. I441-3, and he, therefore, could not express an
opinion about the accuracy of the method.

- 11



among all candidates. (J.S. 40a).

The polarization was most severe in House District #8, where an

average of 92.78 of white voters would not vote for any black

candidate in the primary, J.S. 45a-46a; the District Court

correctly noted that in that district it was mathematically

impossible for a black candidate ever to be elected J.S. 46a.

In the other districts, the degree of polarization was

sufficiently severe enough to be a substantial impediment,

although not an absolute bar, to the election of minority
candidates.

The average portion of white voters willing to support

a black candidate in a primary was 18.3t, not 49t. The

proportion of voters that was white ranged from 70.5t to 84.9t.
(J.S. I9a) fn each of the disputed districts the number of

white voters who in primaries did not support the black

candidate favored by the black community constituted a majority
of the entire electorate. For example, in Senate District *22,

in which 83t of the registered voters are white, an average of

two thirds of the white voters did not vote for the black

candidate in the primaries. That means that a black candidate

starts the campaign with a likely 55t of the electorate who

votes he cannot win. Under those circumstances, the election of

candidates preferred black voters, while not mathematically

impossible, is obviously extremely difficult.
Appellants attack the lower courts' finding of

substantial polarized voting by pointing out those portions of

the record most favorable to their claims. Of the 53 elections

L2



discussed by the trial court, appellants refer only to 8. A.

Br. 36-38. fn most instances, appellants emphasize the election
at which white support for a black candidate was the highest of

any election in that district.l0 the highest proportion of

white support for minority candidates cited by appellants were

in the L982 Durham County general elections and the L982

Mecklenburg County primary. (App. Br. 36-37) But there were no

Republican candidates in the L982 general election in Durham

County, and in the L9B2 Mecklenburg County primary there were

only seven candidates for eight positions in the primary. (.1.S.

_&, __d) rhus the white votes of 472 and 50t,

respectively, represent the number of whites willing to vote for

an unopposed black rather than not vote at aII, rather than the

proportion of whites willing to support in a contested election

a minority candidate favored by the minority community.

A finding of the District Court cannot be found to be

clearly erroneous by cites to only a small portion of the

evidence. In this case, viewing the record as a whole, the

District Court's conclusion that voting is severally racially
polarized is not clearly erroneous.

I0rf,is is true of examples (a ) (b) (h ) (i ) and ( j ) in
Appellants' Brief , pp. __ _ 13 _



HOW THE SYSTEM WORKS

North Carolina's multimember legislative districts have at-

large elections with no full slate reguirement or numbered

seats. In the primaries, there is a majority vote reguirement.

N.C.G.S.5163-111. For multiseat elections, a majority is
calculated as one more than the total number of votes cast,

divided by the number of seats at stake, then divided by two.

Id. If, as is possible depending on the number of candidates in
relation to the number of seats, more candidates obtain a

majority than there are positions to be filled, those having the

highest vote, equal to the number of positions to be filled, are

the winners. 5163-11I(a)(2). Thus, it may take more than a

majority of votes to win. Because the number of votes it takes

to win depends on the number of seats, the number of candidates,

the number of ballots cast by each voter, and the manner in

which voters spread their vote, it is impossible to predict

exactly how many votes it will take to win.

vuw l'ts o N*, ff F-"*#'n hry
3 c+-,'ol^'/e

N,^r 118A

Arrvfl47
f*y llso u'\ F'^'"P