Eastland v. Tennessee Valley Authority Brief Amicus Curiae

    IN THE
UNITED STATES COURT OF APPEALS 

FOR THE ELEVENTH CIRCUIT

No. 82-7008

FRANK L. EASTLAND, et al. ,
Plaintiffs-Appellants, 

v.
TENNESSEE VALLEY AUTHORITY, et al.,

Defendants-Appellees.

On Appeal From the United States District Court 
For the Northern District of Alabama

BRIEF AMICUS CURIAE OF THE NAACP LEGAL DEFENSE 
AND EDUCATIONAL FUND, INC. IN SUPPORT OF 
PETITION FOR REHEARING AND SUGGESTION 

FOR REHEARING EN BANC

JACK GREENBERG 
CHARLES STEPHEN RALSTON 
0. PETER SHERWOOD 
ERIC SCHNAPPER 

Suite 2030 
10 Columbus Circle 
New York, New York 10019 
(212-586-8397)

Counsel for Amicus



IN THE
UNITED STATES COURT OF APPEALS 

FOR THE ELEVENTH CIRCUIT

No. 82-7008

FRANK L. EASTLAND, et al.,
Plaintiffs-Appellants, 

v.
TENNESSEE VALLEY AUTHORITY, et al.,

Defendants-Appellees.

On Appeal From the United States District Court 
For the Northern District of Alabama

BRIEF AMICUS CURIAE OF THE NAACP LEGAL DEFENSE 
AND EDUCATIONAL FUND, INC. IN SUPPORT OF 
PETITION FOR REHEARING AND SUGGESTION 

FOR REHEARING EN BANC

The amicus NAACP Legal Defense and Educational Fund,
Inc., submits this brief because of our ongoing respon­
sibilities as counsel in a substantial number of Title VII 
and other discrimination cases now pending in this circuit 
and this court. We urge that rehearing be granted in this 
case, not because the panel opinion is necessarily incorrect, 
although we believe that is the case, but because the reasoning 
of that opinion is as baffling as it is far reaching. The 
opinion's often contradictory statements of legal principles, 
at times compounded by the manner in which those principles



are applied, will necessarily produce enormous confusion 
among litigants and the lower courts regarding how discrim­
ination is to be proved, who is entitled to challenge discrim­
inatory practices, and under what conditions trial court 
findings may be affirmed on appeal. We set out below the 
novel questions raised, but not answered, by the panel opinion.
We believe that the efficient administration of justice 
would be best served if the May 2 opinion were modified to 
make clear here and now what it has decided, rather than 
leaving that task for "clarification" over a period of years 
by panels with other membership and perhaps different views.

(1) what are the appropriate standards for 
determining the scope of a class?

The district judge in this case drastically narrowed 
the scope of the proposed class. Progressively eliminated 
from the group of blacks that the named plaintiffs could 
represent were employees outside the OACD department, employees 
represented by the Tennessee Valley Trades and Labor Council, 
managerial employees, unsuccessful applicants, and administrative 
employees. (P. 2868). After insisting on "careful attention 
to the requirements of Rule 23," the May 2 opinion concludes 
that the trial court "did not abuse its discretion in narrow­
ing the class" (P. 2869). What those requirements are, and 
what standards are to be applied to any exercise of trial 
court discretion, are critical to the correctness of the 
panel decision, and to its significance for future trial and 
appellate litigation.

-2-



The decision's analysis of these requirements and 
standards, however, which we would have thought central to 
the appeal, is limited to an assertion that the trial judge 
determined that "the representative parties lacked sufficient 
nexus with the putative class members to adequately protect 
their interests." (Id..) The panel explains neither what 
it believes, or what the trial court believed, this "nexus" 
test to require. The result is worse than merely obscure, 
for although this analysis is opaque by itself, the facts 
disclosed by other aspects of the opinion raise even more 
questions. There is, for example, considerable controversy 
about when in a Title VII action successful applicants can 
represent a class that includes unsuccessful applicants; in 
this case, however, one of the named plaintiffs, Eastland, 
was an unsuccessful applicant. If applicants cannot represent 
applicants, who can? The exclusion of TVA employees outside 
the OACD department would appear to have some basis if there 
were no questions of law or fact common to both OACD and 
non-OACD employees. But the panel opinion itself holds, 
correctly in our view, that proof of discrimination against 
non-OACD employees is "corroborative evidence" relevant to 
the claims of black OACD workers. (P. 2872). If the exis­
tence of such common issues does not provide the requisite 
"nexus," what more was required?

As counsel for plaintiffs in a number of class actions 
now pending in this circuit, we are unable to discern from

-3-



the May 2 opinion what type of evidence is now necessary or 
even relevant to determining the scope of a class. Since 
the actual circumstances of this case appear to provide 
precisely the elements hitherto thought sufficient for a 
certification including at least some of the excluded groups, 
no trial judge in this circuit can reasonably be expected to 
understand the new rules for defining the scope of the class. 
Any defendant could find in the panel opinion some basis for 
insisting that a trial court enjoys unlimited and standardless 
discretion to define a class as narrowly as it pleases. The 
panel may not have intended to establish such a rule, or to 
revolutionalize or even alter the standards for certifying 
or defining a class, but the language and details of its 
opinion invite the contrary conclusion.

(2) Where a plaintiff has shown that black employees 
earn significantly less than whites, how many. 
possible reasons for that difference must it 
eliminate in order to establish a prima facie 
case of discrimination?

Here, as in many other Title VII cases, the plaintiffs 
sought to prove the existence of racial discrimination by 
showing that black employees earned less than whites. The 
plaintiffs, of course, did not limit their case to this 
disparity alone; they also affirmatively established that 
the wage differential was not the result of differences in 
seniority or education, since the disparity remained even 
when these circumstances were taken into account. The issue 
on appeal was what more, if anything, plaintiffs were re­
quired to prove.

-4-



The difficulty with this aspect of the May 2 decision 
is not that it fails to articulate any relevant standard, 
but that it appears to announce four different rules. TVA 
objected to plaintiffs' statistics because they did not take 
into consideration possible differences in the employees' 
pre-TVA work experience. The panel rejected this contention, 
insisting that if a defendant believed such an additional 
variable would explain away the wage disparity, the defendant 
could not merely speculate that that might be the case, but 
had the burden of adducing credible evidence to support its 
claim:

In light of the fact that TVA's own regressions 
did hot account for this variable ... the omission 
should not have substantially impaired the validity 
of Eastland's model. "A defendant's claim that the 
plaintiff's model is inadequate because a variable 

. has been omitted will ordinarily ride on evidence 
showing (a) that the qualification represented by 
the variable was in fact considered, and (b) that 
the inclusion of the variable ... changes the results 
of the regression so that it no longer supports the 
plaintiff." (P. 2874, n. 14) (emphasis added).

Elsewhere, in discussing promotion discrimination, the opinion 
indicates there may be some important exculpatory explanations 
that a plaintiff must disprove even if the defendant itself 
never offers evidence to support them; plaintiff's statistics 
regarding promotion rates, this passage suggests, must be 
limited to individuals with "the minimum objective qualifica­
tions necessary for one to be eligible for promotion." (P.
2877) (Emphasis deleted). Third, the court states that the 
value of a statistical analysis "depends in part upon the 
inclusion of all major variables likely to have a large 
effect" (P. 2875, see also id. n. 15); on this view a plaintiff 
is not required to disprove any particular non-racial explanation,

-5-



but the number and significance of the factors the plaintiff
does prove irrelevant determine the weight to be given to
the plaintiff's statistical evidence. Finally, a fourth
passage appears to hold that a plaintiff must consider each
and every conceivable non-discriminatory cause for the salary
disparities, and affirmatively demonstrate that none of them
was responsible for these differences. Thus the panel held
that proof of average wage disparities between blacks and
whites of equal length of service, job schedule, and education,
did "not demonstrate discrimination because the other factors

%

which influence salary are not considered." (P. 2872)
The differences among these standards is of critical 

importance to the resolution of many if not most Title VII 
cases. If a plaintiff is required, in order to establish a 
prima facie case of discrimination, to disprove in advance 
every possible non-racial explanation for differences in the 
treatment of blacks and whites, proof of discrimination will 
be virtually impossible. If there are certain "important" 
defense explanations which a plaintiff must anticipate and 
disprove, the opinion provides no guidance as to what they 
are. In the face of this opinion, plaintiffs' counsel cannot 
know what they are required to prove to establish a prima 
facie case, defense counsel cannot know when they are 
obligated to put on a case of their own, and the trial courts 
are told only that there are at least four possible standards 
for assessing evidence of this kind.

-6-



(3) May present employees challenge discrimination in their 
initial assignments?

Prior to the decision in this case there was, so far as
we are aware, no dispute that if an employee was assigned to
a job on the basis of race, he or she could bring suit under
Title VII. Such discrimination traditionally occurs during
the initial assignment of new workers. Many of the Title
VII cases in this circuit, and the old Fifth Circuit, involved
attempts by blacks to win the right to move into jobs which
had traditionally been reserved i*n this way for whites.

In this case, however, the district court forbad present
employees from litigating such assignment claims. The panel
opinion summarizes and appears to approve the district court
action in the following language:

The certification order limited the class to past 
and present employees, thereby excluding hiring 
or applicant claims. The district court refused 
to consider the initial assignment claims because 
they were "applicant claims" .... [T]he finding that 
... initial assignment is part of the hiring process 
is not clearly erroneous. (P. 2871, n. 9)

The rule apparently sanctioned by the court of appeals is
that the legality of a practice injuring members of a class
cannot be litigated on behalf of the class if that same
practice also injures individuals who are not members of the
class. OACD's insistence that it hires applicants "for
specific positions" is not, of course, different in kind
from the widespread pre-1965 practice of hiring blacks
for certain positions and whites for others. We do not

-7-



understand what legally relevant meaning could be given to a 
finding that "initial assignment is part of the hiring process." 
If, as part of the hiring process, newly selected employees 
are assigned to jobs on the basis of race, it is those employees 
who are injured by that practice. Except where assignment 
policies affect hiring decisions because only "white" positions 
are vacant, assignment practices would not themselves injure 
unsuccessful black applicants who were rejected on account 
of race; thus in many cases unsuccessful applicants would 
lack standing^to challenge assignment practices. If success­
ful applicants cannot challenge such discriminatory practices 
because the practices are "part of the hiring process," and 
unsuccessful applicants cannot challenge them for want of 
standing, no one can attack them at all. Read literally, 
the panel opinion appears to hold that at times, either as a 
matter of law or in the discretion of the trial court, 
discrimination in initial assignment is simply immune from 
attack. Since the panel cannot have intended to legalize 
this form of discrimination, the meaning of its opinion is 
entirely unclear.

The unusual conclusions to which this approach leads 
are illustrated by the panel's discussion of the promotion 
claims of plaintiff Nash. Nash complained that he had been 
denied promotion into an M schedule job because of his race.
The panel opinion notes:

-8-



In defining the class the district court excluded 
employees on the M schedule and subsequently ruled 
that promotions to the M schedule were not a part 
of the case. We have already determined that the 
court did not abuse its discretion in narrowing the 
class. (P. 2880)

The logic of the discussion of assignment claims is here 
taken to the natural but extreme conclusion that an in­
dividual who is the victim of discrimination cannot 
challenge that discriminatory practice if it also injured 
other individuals who are not part of a certified class.
Here, as with assignments, the panel cannot conceivably have 
intended to announce or approve such a rule, but what it did 
intend the quoted passage to mean is a matter of conjecture.

(4) Can proof of discrimination in promotion or assignments 
be proved by disparities in overall wages

Part 1(c)(1) of the May 2 opinion holds unequivocally, 
and correctly in our view, that proof of disparities in the 
salaries of blacks and whites may be used to establish dis­
crimination in promotions or assignments:

Higher salaries generally coincide with higher 
level jobs. If blacks at TVA earn less than 
whites it is because they are assigned to lower 
ranking, lower paying positions (p. 2872)

The panel noted that Fifth Circuit precedent supports the 
use of wage disparities to prove such discrimination. (Id.)

The panel's analysis of the record in this case, however, 
appears to disavow this standard. Plaintiff offered a statis­
tical analysis of OACD wages which showed that blacks were 
paid less than whites even when education, length of service, 
and "a large number of variables" were taken into account.

-9-



(P. 2876). The panel, however, dismissed this evidence of 
discrimination because it failed to "control for job category." 
(Id.) There are five job schedules and several score job 
categories within the relevant units of OACD. "Controlling 
for job category" meant that allegedly underpaid blacks were 
compared only with whites in the'same job category, not with 
similarly or less qualified whites in other job categories.
Thus the TVA analysis relied on by the panel demonstrated 
decisively that the differences in the wages of comparable 
black and white employees were due in large measure to the 
fact that a disproportionate number of whites were in»the 
better paid categories, while blacks were more frequently in 
the lower paid job categories.

This analysis might weigh in favor of the defendant if 
the plaintiffs were complaining only about promotion dis­
crimination within each category, not between categories, 
and if the categories were so different (e.g. lawyers, 
physicists, foreign language translators) that no employee 
ever moved from one to the other. In fact, however, promo­
tions did occur into the better paying job categories (see, 
e.g., P. 2880), and the existence of discrimination in such 
promotions was expressly alleged by the plaintiffs (Id.).

Thus, although salary disparities may in theory be used 
to prove discrimination in promotions and assignments, that 
evidence can, in light of the May 2 decision, be "rebutted" 
by a showing that the reason blacks make less is "simply 
because they are more highly represented" (P. 2876) in the 
poorly paid jobs due to disparities in assignment or promotion.

-10-



This "rebuttal" consists of precisely the kind of discrim­
ination the plaintiff is seeking to prove. The panel's 
actual resolution of the case seems squarely.to make proof 
of such discrimination literally self-defeating.

(5) Does Pullman-Standard Co. v. Swint apply to affirmances 
of erroneous trial court decisions?

In Pullman-Standard Co. v. Swint, 72 L.Ed. 2d 66 (1982), 
the Supreme Court delineated the respective roles under the 
Federal Rules of Civil Procedure of the federal district 
courts and courts of appeals. Where the trial court's decision 
is tainted by some error, the court of appeals is not ordinarily 
authorized to decide on its own the factual questions at 
issue, or to speculate about how the district court would 
have decided the case had that error not occurred or about 
how the lower court might act on remand. Puliman-Standard 
appears to sanction the affirmance of a trial court's conclu­
sion on the merits only under two circumstances: (1) where
those conclusions, and the premises on which they are based, 
have been found to be correct as a matter of law and not 
clearly erroneous as a matter of fact, and (2) where, although 
there were lower court errors, the record would compel in a 
second appeal the ultimate conclusion initially reached by 
the trial court even if, on remand, that court were to arrive 
at the opposite result. In all other situations a case must 
be remanded for a fresh decision by the district court.

The affirmance in this case, however, falls into neither 
of those categories. It introduces and turns on two apparently 
novel doctrines, —  first, that certain allegedly erroneous 
findings relied on by the trial court need not be reviewed

-11-



if they are "surplusage," and second, that a trial court's 
refusal to consider relevant evidence may at times be "not 
error." Both doctrines seem outside the language, and incon­
sistent with the reasoning, of Puliman-Standard.

The plaintiffs in this court attacked the trial court's 
subsidiary findings regarding promotion rates and labor 
market comparisons; the panel, rather than undertaking to 
decide whether the findings were correct, expressly "disregard[ed] 
these findings as surplusage." (P. 2877 n. 18). If evidence 
regarding promotion rates was irrelevant, the trial court 
errred by relying on it; if such evidence was indeed relevant, 
the correctness of the trial court's ultimate conclusion 
necessarily turned on whether it had properly resolved the 
disputes regarding that evidence. Either way it is difficult 
to understand how that conclusion of non-discrimination 
could be upheld without an examination of the allegedly 
faulty premises on which it was based.

A similar problem is raised by the district court's 
dismissal as "irrelevant" of proof of discrimination against 
blacks outside of the OAC department. (P. 2872, Data Set 
B). The panel found that this evidence "could have been 
considered as corroborative evidence," i.e. that it was 
relevant, but also held that its rejection "was not error."
(Id.) How a refusal to consider relevant evidence can be 
"not error" is not explained. The trial court in any event, 
did not think it could consider such evidence; whether, in 
light of appellate direction to do so, it might have reached 
the same ultimate conclusion on the merits is a matter of 
speculation.

-12-



Although the "surplus finding" and "not error" rule 
appear critical to the outcome of the panel decision, neither 
their rationale nor their scope is explained. Both open new 
questions, to which they provide no answers, regarding 
whether trial judges, or district court litigants, are 
really obligated to conform their actions to the substantive 
legal standards announced by the appellate courts. Both 
raise novel issues, which they do not purport to resolve, 
about the degree to which erroneous district court decisions 
are now-to be exempt from reversal or even scrutiny.

For the above reasons the petition for rehearing should 
be granted, and the case reheard en banc.

CONCLUSION

Respectfully submitted

JACK GREENBERG 
CHARLES STEPHEN RALSTON 
0. PETER SHERWOOD 
ERIC SCHNAPPER

Suite 2030 
10 Columbus Circle 
New York, New York 10019 
(212-586-8397)

Counsel for Amicus

-13-



CERTIFICATE OF SERVICE

I hereby certify that on this 20th day of May, 1983, I 
served three copies each of the Brief Amicus Curiae of the 
NAACP Legal Defense and Educational Fund, Inc., and a copy 
of the Motion for Leave to File Brief Amicus Curiae, on 
counsel for the parties, by causing them to be deposited in 
the United States mail, first class postage prepaid, addressed 
to:

Paul Saunders Cravath Swaine & Moore 
One Chase Manhattan Plaza 
New York, New York 10005
Richard T. Seymour
Lawyers' Committee for Civil Rights 

Inder Law
733 Fifteenth St., N.W.
Washington, D.C. 20005 •
Herbert S. Sanger, Jr.
General Counsel 
Tennessee Valley Authority 
Knoxville, Tennessee 37902
Bernard E. Bernstein 
Bernstein, Susano, Stair & Cohen 
Sixth Floor
First Tennessee Bank Building 
Knoxville, Tennessee 37902
Melvin Radowitz 
Jacobs & Langford, P.A.
1000 Rhodes-Haverty Building 
134 Peachtree St., N.W.
Atlanta, Georgia 30303