Wards Cove Packing Company, Inc. v. Antonio Petitions and Briefs

    The Supreme Court of the United States

Wards Cove Packing Co., Inc.
versus (87-1387)
Antonio

Petitions and Briefs JUN L

naac? legal oknse fund
Libra ry
™W050N  STREET 

YORK, N. Y. 10013

Labor Law Series

Volume 22, No. 10 
1988/89 Term of Court

Law Reprints



TABLE OF CONTENTS

Wards Cove Packing Co., Inc.
versus (87-1387)
Frank Antonio, et al.

PAGE
Docket Sheet. . . . . . . . . . . . . . . . . . . . . . . ii
Petition for Writ of Certiorari. . . . . . . . . . . . 1
Opposition. . . . . . . . . . . . . . . . . . . . . . . . 35
BRIEFS ON THE MERITS
Petitioner. . . . . . . . . . . . . . . . . . . . . . . . 51
Respondent. . . . . . . . . . . . . . . . . . . . . . . . 113Reply Brief of Petitioners. . . . . . . . . . . . . . 183
AMICI CURIAE BRIEFS ON THE MERITS
American Civil Liberties Union. . . . . . . . . . . . 215American Society for Personnel Administration. .. 291
Center for Civil Rights. . . . . . . . . . . . . . . . 327Chamber of Commerce of the United States. . . . . . 353Equal Employment Advisory Council. . . . . . . . . . 389Lawyers' Committee for Civil Rights Under Law... 425NAACP Legal Defense & Educational Fund, Inc. . . . . 461
National Association for the Advancement of
Colored People. . . . . . . . . . . . . . . . . . . . 533

United States. . . . . . . . . . . . . . . . . . . . . . 555

- i -



DOCKET SHEET

Ho. 87-1387-CFX 
Status: GRANTED

Docketed: 
February 9, 1988
Vide:

Entry Date

1 reb 9 1988
2 Mar 11 1988
3 Mar 16 1988
5 Jun 24 1988
6 Jun 30 1988

8 Jul 12 1988
9 Aug 29 1988

10 Sep 9 1988
1 1 Sep 9 1988
12 Sep 9 1988
1 3 Sep 9 1988
1 4 Sep 9 1988
15 Sep 10 1988
16 Sep 10 1988
17 Sep 13 1988
19 Sep 13 1988
20 Oct 24 1988
21 Nov 4 1988
22 Nov 4 1988
23 Nov 4 1988
24 Nov 4 1988
25 Nov 5 1988
26 Nov 22 1988
27 Dec 2 1988
28 Dec 5 1988
29 Jan 18 1989

Title: Hards Cove Packing Company, Inc., et al.,
Petitioners
v.
Frank Atonio, et al.

Court: United States Court of Appeals
for the Ninth Circuit

Counsel for petitioner: Fryer,Douglas
Counsel for respondent: Arditl.Abraham A.

Note Proceedings and Orders

G petition for writ of certiorari filed.
Brief of respondent Frank Antonio, et al. In opposition 

filed.
DISTRIBUTED. April 1, 1988 
REDISTRIBUTED. June 29, 1988
Petition GRANTED, limited to Questions 1, 2, and 3 
presented by the petition.
Order extending time to file brief of petitioner on the 
merits until September 10, 1988.
Record filed.

* Certified copy of original record and proceedings,
volumes 1-54, 6 boxes and 1 envelope, recleved.

Brief amicus curiae of Center for Civil Rights filed.
Brief amicus curiae of American Society for Personnel 

Administration filed.
Brief amicus curiae of United States filed.
Brief amicus curiae of Egual Employment Advisory Council 

filed.
Brief amicus curiae of Chamber of Commerce for the United 

States filed.
Joint appendix filed.

• Brief of petitioner Wards Cove Packing Co., Inc. filed. 
Lodging recived. (12 binders - box).
Order extending time to file brief of respondent on the 
merits until November 5, 1988.
SET FOR ARGUMENT. Wednesday, January 18, 1989. (4th 
case) (1 hr.)

Brief amici curiae of ACLU, et al. filed.
Brief amici curiae of HAACP Legal Defense, et al. filed. 
Brief amicus curiae of Lawyers' Committee for Civil Rights 

Under Law filed.
Brief amicus curiae of NAACP filed.
Brief of respondents Frank Atonio, et al. filed.

CIRCULATED.
Lodging received. (10 copies).

X Reply brief of petitioner Wards Cove Packing Co., Inc. 
filed.
ARGUED.

- 11 -



N o. 87-1387

IN THE

Supreme Court of the United States

October Term, 1987

W a rd s  Cove P acking Company. Inc.. 
Castle  & Cooke. Inc..

Petitioners,

v.

F rank  A tonio. et al.,

Respondents.

PETITION FOR WRIT OF CERTIORARI 
TO THE UNITED STATES COURT OF APPEALS 

FOR THE NINTH CIRCUIT

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ikkelborg . B roz. 
W ells & F ryer 
Suite 3300
1001 Fourth Avenue Plaza 
Seattle, Washington 98154 
(206) 623-5890
A tto rneys for Petitioners

* Counsel of Record 

February 9, 1988

1



1

QUESTIONS PRESENTED

1. Does statistical evidence that shows only a concen­
tration of minorities in jobs not at issue fail as a matter of law 
to establish disparate impact of hiring practices where the 
employer hires for at-issue jobs from outside his own work 
force, does not promote-from-within or provide training for such 
jobs, and where minorities are not underrepresented in the 
at-issue jobs?

2. In applying the disparate impact analysis, did the 
Ninth Circuit improperly alter the burdens of proof and en­
gage in impermissible fact-finding in disregard of established 
precedent of this Court?

3. Did the Ninth Circuit commit error in allowing plain­
tiffs to challenge the cumulative effect of a wide range of 
alleged employment practices under the disparate impact 
model?

4. Was it error for the Ninth Circuit to expand the reach 
of the disparate impact theory to employment practices such 
as word of mouth recruiting, subjective application of hiring 
criteria, and other practices that do not operate as “automatic 
disqualifies?”

2



11

Petitioners are Wards Cove Packing Co., Inc., and Castle 
& Cooke, Inc., who were defendants in the trial court pro­
ceeding. (Claims against a third defendant, Columbia Wards 
Fisheries, were dismissed. This was affirmed on appeal. 
See fn. 2 infra.)

Respondents are Frank Atonio, Eugene Baclig, Randy del 
Fierro, Clarke Kido, Lester Kuramoto, Alan Lew, Curtis Lew, 
Joaquin Arruiza, and Barbara Viernes (as administratix of 
the Estate of Gene Allen Viernes), who were individual plain­
tiffs and representatives of a class of all nonwhite employees 
in the trial court proceeding.

Rule 28.1 disclosure:
Wards Cove Packing Company,
Inc. is a privately-held 
domestic corporation.
Castle & Cooke, Inc. is a 
publicly-held and traded 
domestic corporation.

LIST OF PARTIES

3



Ill

TABLE OF CONTENTS

Questions Presented...................................   i

List of P arties .............................................. ................  . ii

Thble of C ontents.................. .. ................... .. iii

List of Authorities........ ....................................................... vi

Opinions Below . . ....................................................   1

Jurisd iction .........................................  . . .  .....................2

Pertinent S ta tu te ..................    2

Statement of The C ase ....................   3

A. Nature of the Case........ ................ .................... .. 3

B. Material F a c ts ...........................     4

C. Court of Appeals Rulings................ ................ .. . . . 6

Reasons for Granting the Petition................................. .. . . 8

Page

I. The Simplistic Notion That Racial Imbalance Can 
Establish Disparate Impact in the Face of Findings 
That Minorities Are Not Underrepresented in the 
Jobs at Issue is Not Supported By the Decisions of 
This Court and is Rejected by Several Other Circuits; 
is a Fundamental Misconception of the Role of Statis­
tics in Proving Discrimination; Has Far-Reaching, 
Ominous Implications for Employers; and Is Out of 
Step With the Congressional Policy of Title VII of the 
Civil Rights Act of 1964....................................................8

4



II. The Ninth Circuit’s Application of the Disparate 
Impact Theory Represents a Radical Departure 
from Established Precedent of This Court, and 
Threatens to Revolutionize the Allocation of Proof
in Discrimination Suits................................................11

A. In Reaching for a Basis to Vacate the Dis­
trict Court’s Judgment, the Ninth Circuit Has 
Ignored Prior Precedent of This Court and the 
TVial Court’s Findings..............................................11

B. The Ninth Circuit Decision is a Revolutionary 
Departure from the Established Rules for the 
Allocation of Proof in a Discrimination Case. . 1 3

1. New Allocation of Proof.................................. 13

2. Hiring C riteria................................................. 15

3. Sources of Employees..................................... 16

III. Allowing Plaintiffs to Challenge an Entire Range
of “Named” Employment Practices Merely Be­
cause the Employers’ Work Force Reflects Uneven 
Racial Balance Is an Improper Application of the 
Disparate Impact Model, Places an Unfair Burden 
on the Employer, and Exacerbates an Existing 
Conflict of Authority in the Circuits........................... 18

iv

TABLE OF CONTENTS, (continued)

Page

5



V

TABLE OF CONTENTS, (continued)

Page

IV. There is a Substantial Conflict in the Circuits as to 
Whether the Disparate Impact Analysis May Be 
Applied to Subjective Decision Making and Other 
Practices That Do Not Act as “Automatic Dis­
qualifies.” ............................... ......................................21

Conclusion................................... ........................  ..........  22

Appendices:

Appendix I: . . 

Appendix II: . . 

Appendix I I I : . 

Appendix IV:. . 

Appendix V:. . . 

Appendix VI: . 

Appendix V II:. 

Appendix VIII: 

Appendix IX: .

ic
. m

*
ii - i

m i*

IV-l*

. VI-l*

VIM*
*

VIII-1 

. IX-1*

Ed. Note: * Denotes material not reprinted herein.

6



Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975) 13

Anderson v. Bessemer City, 470 U.S. 564 (1985) 11,12

Bunch v. Bullard, 795 F.2d 384 (5th Cir. 1986)................ 22

Clark v. Chrysler Corp., 673 F.2d 921 (7th Cir. 1982). . . 9,10,17

Coser v. Moore, 739 F.2d 746 (2d Cir. 1984)....................  10

De Medina u. Reinhardt, 686 F.2d 997 (D.C. Cir. 1982) . . . 9

Dothard v. Rawlinson, 433 U.S. 321 (1977)................  11,15

EEOC v. American N a tl Bank, 652 F.2d 1176 (4th Cir.
1981).................................................................................. 11

EEOC v. Federal Reserve Bank of Richmond, 698 F.2d 
633 (4th Cir. 1983), rev'd on other grounds, sub nom, 
Copper v. Federal Reserve Bank of Richmond, 467 
U.S. 867 (1984).........................................................9,10,22

Freeman v. Lewis, 675 F.2d 398 (D.C. Cir. 1982).............. 13

Furnco Constr. Corp. v. Waters, 438 U.S. 567 (1978) . . . .  18

Griffin v. Board of Regents, 795 F.2d 1281 (7th Cir. 1986)... 22

vi

TABLE OF AUTHORITIES

Table of Cases Page

Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1985). . . 13,15,19,21

Griggs v. Duke Power Co., 401 U.S. 424 (1971)..............  15

Hammon v. Barry, 826 F.2d 73 (D.C. Cir. 1987)............  9,20

7



Statutes Page

Civil Rights Act of 1866, 42 U.S.C. § 1981 . . . . . . . . .  3

Title VII of Civil Rights Act of 1964, as amended,
42 U.S.C. § 2000e-2(a).........    2-3,8,20

28 U.S.C. § 1254(1)..............................  . . . . . . . . . . . .  2

28 U.S.C. § 1331................ 2

28 U.S.C. § 2101(c)............................................... 2

Other Authorities

ix

TABLE OF AUTHORITIES, (continued)

B. Schlei & P. Grossman, Employment Discrimination
Law (2d ed. 1983)..............  11

Miscellaneous

Webster's Third New International Dictionary of the
English Language Unabridged 12

10



IN THE
Supreme Court of the United States

October Term, 1987

W a r d s  Cove P acking  Com pany . Inc..
Ca st le  & Co o ke . Inc..

Petitioners,

v.

F r an k  A tonio . et al.
Respondents.

PETITION FOR WRIT OF CERTIORARI 
TO THE UNITED STATES COURT OF APPEALS 

FOR THE NINTH CIRCUIT

OPINIONS BELOW

On October 31, 1983, the United States District Court 
for the Western District of Washington (Quackenbush. J.) 
entered an opinion following a nonjury trial. Appendix I. 
An order correcting the opinion and judgment in favor of 
petitioners was entered December 6, 1983. App. II. The trial 
court's decision was published at 34 E.P.D. ^34,347 (Commerce 
Clearing House, Inc.). The opinion of the Court of Appeals 
affirming the judgment was published at 768 F.2d 1120. App.
III. An order that withdrew the opinion and ordered rehearing 
en banc was published at 787 F.2d 462. App. IV. An opinion 
of the en banc Court of Appeals was published at 810 F.2d 477. 
App. V. A second opinion of the original panel of the Court

11



2

of Appeals on remand from the en banc court was published 
at 827 F.2d 439. App. VI. On November 12, 1987 an order 
clarifying the opinion was entered, App. VIII, and a petition 
for rehearing denied. App. IX.1

JURISDICTION

Federal jurisdiction in the trial court was invoked under 
28 U.S.C. § 1331. The decision of the Court of Appeals sought 
to be reviewed was entered on September 2, 1987. App. VI. 
A timely petition for rehearing was filed on September 16, 
1987, App. VII, and the petition was denied on November 12, 
1987. App. IX. Jurisdiction in this Court is invoked under 
28 U.S.C § 1254(1). This petition is timely under 28 U.S.C 
§ 2101(c).

PERTINENT STATUTE

Plaintiffs’ claims arise under Title VII of Civil Rights Act 
of 1964, as amended, 42 U.S.C. § 2000e-2(a):

(a) It shall be an unlawful employment practice for an 
employer—

(1) to fail or refuse to hire or to discharge any indi­
vidual, or otherwise to discriminate against any individual 
with respect to his compensation, terms, conditions, or 
privileges of employment, because of such individual’s 
race, color, religion, sex, or national origin; or

(2) to limit, segregate, or classify his employees or 
applicants for employment in any way which would de­
prive or tend to deprive any individual of employment

1 In addition, the plaintiffs took two interlocutory appeals: One 
unpublished opinion, affirming a denial of a motion for preliminary 
injunction and another affirming in part and reversing in part a 
dismissal of Title VII claims. 703 F.2d 329.

12



3

opportunities or otherwise adversely affect his status 
as an employee, because of such individual's race, color, 
religion, sex, or national origin.

STATEMENT OF THE CASE

A. Nature of the Case.

The named plaintiffs in this class-action suit are former 
employees at several salmon canneries in Alaska. They brought 
this action against their former employers, petitioners Wards 
Cove Packing Company, Inc., and Castle & Cooke, Inc.2 charg­
ing employment discrimination on the basis of race in violation 
of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e, 
et seq., and the Civil Rights Act of 1866, 42 U.S.C. § 1981. 
The class is defined as all nonwhites who are now, will, be, or 
have been at any time since March 20, 1971, employed at any 
one of five canneries.

Following a lengthy non-jury trial, the trial court found 
that plaintiffs had not established discrimination under § 1981 
or Title VII and judgment was entered for petitioners. The 
Ninth Circuit affirmed this decision, but on rehearing en banc 
resolved a conflict within the circuit by determining that the 
impact analysis could be applied to subjective employment 
practices and remanded to the original panel. The subsequent 
panel decision vacated the judgment and remanded to the 
district court with directions to apply the disparate impact 
analysis in a manner inconsistent with decisions of this Court 
and in conflict with other circuits.

2 Claims against Columbia Wards Fisheries, an additional defen­
dant, were dismissed. Atonio v. Wards Cove Packing Co.. Inc., 
703 F.2d 329 (9th Cir. 1983); also see App. III-13-15 and App. VIII.

13



4

B. Material Facts.

Petitioners operate salmon canneries in remote and widely 
separated areas of Alaska. Of eleven facilities, five were 
certified for this class action. The canneries operate only 
during the summer salmon run. For the remainder of the year 
they are vacant. Petitioners' head office and support facilities 
are located at Seattle, Washington, and Astoria, Oregon.

The employment needs to operate the canneries will vary 
with the size and duration of the salmon runs. Petitioners hire 
employees primarily from the Pacific Northwest and Alaska. 
The bulk of employees are “cannery workers,’’ who work in the 
cannery itself on the fish processing, canning lines. Local 37 
of the I.L.W.U. has jurisdiction and a contract for these jobs. 
The remaining jobs at the cannery are carpenters, machinists, 
tender crews, longshoremen, administrative, and other support 
personnel. It is these non-cannery worker jobs which are at 
issue. The non-cannery worker jobs are covered by several 
union contracts. Some are non-union. The trial court’s opinion 
sets forth the facts in detail. (App. I; see also the background 
discussion in first panel opinion at App. 111-3-12.)

Petitioners hire all employees except those persons work­
ing on the cannery line (cannery workers) from an external 
labor market which is 10% nonwhite. For the positions at 
issue, nonwhites filled 21% of the at-issue jobs at the class 
canneries and approximately 24% in petitioners’ overall Alaska 
operations. Cannery workers, on the other hand, were hired 
through Local 37 of the I.L.W.U. The composition of Local 37 
is dominated by Filipinoes, as are the crews it dispatches to 
the canneries each summer. In addition, petitioners filled some 
cannery worker jobs for some of the more remote canneries 
from local populations.

In 1974 plaintiffs commenced a class action against peti­
tioners. The suit mounted a broad-scale attack against the

14



5

gamut of petitioners' employment practices. Plaintiffs identi­
fied 16 “practices ’3 which they contended caused an imbalance 
and thus a “concentration'' of nonwhites in the lower-paying 
cannery worker jobs. Plaintiffs used comparative statistics 
to argue that of the total work force, the majority of the non­
whites were concentrated in the lower-paying jobs and that 
there should have been a balance of 50% white/nonwhite 
employees in all job classifications.

After 12 trial days, in which more than 100 witnesses 
testified, over 900 exhibits were admitted, and over 1,000 
statistical tables were submitted, the trial court entered 
extensive findings of fact in a 73-page opinion. App. I. The 
findings determined that plaintiffs’ comparative statistics were 
of little probative value; that the labor supply for petitioners' 
facilities is 90% white; that minorities were not underrepre­
sented in the at-issue jobs; that cannery workers are not the 
appropriate comparison labor pool for at-issue jobs; that 
petitioners hire from an external labor supply and do not either 
promote-from-within or train inexperienced, unskilled workers 
for at-issue jobs; that most jobs at issue require skill and prior 
experience that is not readily acquirable at the canneries; that 
Local 37 provides an oversupply of nonwhite cannery workers 
and that this overrepresentation is an institutional factor in 
the industry.4

In addition, the trial court found that no individual 
instances of discrimination were proven; that petitioners did 
not give job preference to friends and relatives; that plaintiffs’
3 The 16 practices were word-of-mouth recruitment, separate hiring 

channels, nepotism, termination of Alaska natives, rehire prefer­
ence, retaliatory terminations, menial work assignments, frater­
nization restrictions, housing, messing, English language require­
ment, race labeling, subjective hiring criteria, lack of formal 
promotion practices, failure to post openings, and discrimination 
in pay in certain jobs.

4 None of these findings were challenged on appeal.

15



6

“nepotism" statistics were distorted and unreliable; that hiring 
was on the basis of job-related criteria; that hiring of exper­
ienced personnel was a business necessity; that the rehire 
preference clauses in the union contracts operated like a 
seniority system; that housing is not racially segregated, and 
that housing, rehire, and messing policies were all dictated by 
business necessity.

The trial court found that plaintiffs had failed to establish 
intentional discrimination and the disparate impact analysis 
was not appropriate for application to plaintiff’s wide-ranging 
miiltiple practice challenge nor to subjective hiring practices. 
In applying the impact analysis individually to five of peti­
tioners' practices (rehire preference, English language, “nepo­
tism.” housing, and messing), the district court again found 
in favor of petitioners.

C. Court of Appeals Rulings.
On appeal a panel of the Ninth Circuit affirmed the 

judgment, noting, however, that there was a conflict in the 
decisions of several circuits and the Ninth Circuit itself as to 
whether the disparate impact analysis could be applied to 
analyze “subjective practices.” 768 F.2d 1120 (9th Cir. 1985), 
App. III. This opinion was withdrawn after rehearing en banc 
was granted. 787 F.2d 462 (9th Cir. 1985), App. IV. On en banc 
rehearing, the Ninth Circuit held that the disparate impact 
analysis could be applied to such practices. 810 F.2d 477 
(9th Cir. 1987), App. V. The case was then remanded to the 
original panel which sought to apply the impact analysis to 
eight of the 16 “practices” identified by plaintiffs.5 827 F.2d 
439 (9th Cir. 1987), App. VI.
5 The practices selected by the panel were subjective hiring criteria, 

word-of-mouth-recruitment, nepotism, separate hiring channels, 
rehire preferences, housing, messing, and labeling. The Ninth 
Circuit does not explain why these were selected nor what dis­
position was made, if any, of the other eight practices alleged to 
have caused the “imbalance" in hiring.

16



7

On remand the Court of Appeals panel affirmed the dis­
trict court on the rehire preference, did not discuss the English 
language requirement, but held that plaintiffs’ “comparative 
statistics,” which showed only a concentration of minorities 
in the cannery worker jobs, were nonetheless adequate to 
require the district court to examine petitioners’ hiring 
practices on grounds of business necessity. In doing so, the 
Court of Appeals did not hold that any practice caused 
disparate impact,® and ignored the district court's findings that 
plaintiffs’ statistics were distorted and unreliable, that peti­
tioners hired more nonwhites than the proportion available in 
the labor supply, and that institutional factors, not the peti­
tioners' practice, caused an overrepresentation of minorities 
in cannery worker jobs.

The court also held, contrary to trial court findings, that 
a preference for relatives (“nepotism”) existed and had an 
adverse impact on nonwhites. Finally, the court questioned 
the district court’s finding of business necessity for petitioners' 
housing and messing practices. The Court of Appeals vacated 
judgment for petitioners and remanded.

9 The Ninth Circuit implied that petitioners ‘‘conceded” causation 
and did not argue no impact was shown. 827 F.2d at 446, 447. 
This is not true. Proof of causation and impact is plaintiffs’ 
burden and petitioners have maintained throughout that plaintiffs 
failed to meet their burden on both.

17



8

REASONS FOR GRANTING THE PETITION

This case raises fundamental questions as to the bound­
aries of the disparate impact theory under Title VII of the 
Civil Rights Act of 1964, 42 U.S.C. 2000e. et seq., and as to the 
role of statistics and the allocation of proof under that theory. 
The Ninth Circuits decision, in direct conflict with several 
other circuits, effectively allows a plaintiff to shift the burden 
of proof to an employer by establishing only that the em­
ployer's work force has an uneven racial balance. Tb reach this 
extraordinary conclusion, the Court of Appeals had to dis­
regard established precedent of this Court and other circuits, 
invent new rules for allocation of proof, and totally ignore the 
trial court’s key findings of fact and the petitioners’ evidence.

This petition should be granted because only this court 
can answer the questions raised, resolve the conflicts created, 
and rectify the wrong that has been done to petitioners.

I. The Simplistic Notion That Racial Imbalance Can Estab­
lish Disparate Impact in the Face of Findings That 
Minorities Are Not Underrepresented in the Jobs at Issue 
is Not Supported By the Decisions of This Court and is 
Rejected by Several Other Circuits; is a Fundamental 
Misconception of the Role of Statistics in Proving Dis­
crimination; Has Far-Reaching, Ominous Implications for 
Employers; and Is Out of Step With the Congressional 
Policy of Title VII of the Civil Rights Act of 1964.

The Ninth Circuit gave plaintiffs’ comparative internal 
work force statistics decisive weight in vacating the trial 
court's judgment for the employers. Id., 827 F.2d at 444-447 
(App. VI, pp. 14-18). However, in doing so the Court of Appeals 
ignored the admonition of this court that the usefulness of 
statistics “depends on all of the surrounding facts and other 
circumstances.” Teamsters u. United States, 431 U.S. 324, 
340 (1977). It also ignored the unchallenged findings of the 
trial court on the labor market.

18



9

In failing to recognize the significance of the findings, 
particularly as to the labor market, the Ninth Circuit com­
mitted serious error. The decision is in direct conflict with 
Hazelwood School Dist. v. United States. 433 U.S. 299 (1977)
and Johnson v. Transp. Agency, 480 U.S.___ , 94 L. Ed. 2d 615
(1987), which hold that where at issue jobs are filled from 
outside the employer's own work force, it is appropriate to focus 
on the racial composition of the relevant external labor market 
and statistically compare it to the employer’s actual hiring.7 
The post-Hazelwood circuit court opinions agree. E.g., 
Johnson v. Uncle Ben's, Inc., 628 F.2d 419, 425 (5th Cir. 1980); 
Rivera v. City of Wichita Falls, 665 F.2d 531, 544-45 (5th Cir.
1982); De Medina v. Reinhardt, 686 F.2d 997, 1004-1009 (D.C. 
Cir. 1982); EEOC v. Federal Reserve Bank of Richmond, 698 
F.2d 633, 658-62 (4th Cir. 1983), rev'd on other grounds sub 
nom. Cooper v. Federal Reserve Bank of Richmond, 467 U.S. 
867 (1984). See Hilton v. Wyman-Gordon Co., 624 F.2d 379, 
380 (1st Cir. 1980) (plaintiff’s "concentration” evidence 
rebutted by outside labor force statistics): Clark v. Chrysler 
Corp., 673 F.2d 921 (7th Cir. 1982) (any showing of impact 
tefuted by external labor market statistics). Cf. Hammon v. 
Barry, 826 F.2d 73 (D.C. Cir. 1987) (improper to adopt affirma­
tive action plan where minorities not underrepresented in 
comparison to area labor force). The Ninth Circuit did not 
mention, discuss, or heed these decisions.

In effect, what the Ninth Circuit has done is hold that 
a mere internal work force showing of "concentration” of 
minorities, without regard to the factual circumstances, is 
sufficient to establish the disparate impact of the amalgam 
of practices plaintiffs choose to name. This is a direct conflict

7 At trial both parties recognized that establishment of the most 
reasonable proxy for the pool of potential applicants was neces­
sary. Hazelwood v. United States, 433 U.S. 200 (1977). Both 
offered expert and statistical evidence on the labor market and 
the trial court found petitioners’ evidence more probative.

19



10

with at least four other circuits whose post-Hazelwood deci­
sions hold (1) internal work force comparisions are relevant, if 
at all, only in a promotion case or where the employer trains 
its workers for promotion and then, only if plaintiff focuses 
on the qualified segment of the promotion pool, Johnson u. 
Uncle Ben's, Inc., supra, 628 F.2d at 425 (5th Cir.); Ste. Marie 
v. Eastern R. Assoc., 650 F.2d 395, 400 (2d Cir. 1981); EEOC 
v. Federal Reserve Bank of Richmond, supra, 698 F.2d at 
659-60 (4th Cir.); Rivera, supra, 665 F.2d at 541, n.16 (5th Cir.); 
and (2) that a showing of concentration in a hiring case will 
be refuted by external labor market evidence that shows no 
underrepresentation of minorities, Hilton, supra, 624 F.2d at 
380: Coser v. Moore, 739 F.2d 746, 752 (2d Cir. 1984); EEOC 
v. Federal Reserve Bank of Richmond, supra, 698 F.2d at 
658-62 (4th Cir.); Rivera, supra, 665 F.2d at 539, 544-45 (5th 
Cir.). See Clark, supra, 673 F.2d at 929 (external labor market 
data showed no impact in hiring).

The foregoing decisions stand for the proposition that 
plaintiffs cannot establish a disparate impact in hiring for jobs 
at issue with statistical evidence that shows only a concentra­
tion of minorities in jobs not at issue, where the employer has 
hired minorities in their porportion to the labor market and 
hires from an external, not internal, labor pool. The Ninth 
Circuit disagrees, but it stands alone in that disagreement.

The Ninth Circuit is geographically the largest court of 
appeals circuit in America. Tb allow this fundamentally 
erroneous view of the role of statistical proof exposes every 
employer in the West that does not have an “even" racial 
balance in all of its jobs to the threat of litigation and the risk 
of liability, regardless of the particular circumstances of their 
businesses. As discussed below, this is not what Congress 
intended nor do the logical implications of this decision carry 
out the spirit or the letter of Title VII. See III, infra.

20



11

II. The Ninth Circuit’s Application of the Disparate Impact 
Theory Represents a Radical Departure from Established 
Precedent of This Court, and Threatens to Revolutionize 
the Allocation of Proof in Discrimination Suits.

A. In Reaching for a Basis to Vacate the District Court s 
Judgment, the Ninth Circuit Has Ignored Prior Prece­
dent of This Court and the Trial Court's Findings.

First, as pointed out above, the Ninth Circuit did not 
consider the trial court's finding as to the probative value of 
petitioners’ statistical evidence. This was a finding of fact, 
was not clearly erroneous, and should not have been ignored. 
Anderson v. Bessemer City, 470 U.S. 564 (1985). The Ninth 
Circuit could not have reached its decision without avoiding 
this finding and in doing so. it violated the first principle of 
appellate decision-making.

Second, it is clear that before the burden is shifted to the 
employer in an impact case to prove job relatedness or business 
necessity, the court must evaluate both petitioners’ attacks 
on plaintiffs’ evidence and petitioners ’ own rebuttal evidence, 
Dothard v. Rawlinson, 433 U.S. 321, 338 (1977) (Rehnquist, J., 
concurring); EEOC v. American Nat l Bank, 652 F.2d 1176, 
1189 (4th Cir. 1981); Pouncy v. Prudential Ins. Co., 668 F.2d 
795, 800-801, n.8 (5th Cir. 1982); Shidaker v. Carlin, 782 F.2d 
746, 750 (7th Cir. 1986). See B. Schlei & P. Grossman, Employ­
ment Discrimination Law, pp. 1325-26 (2d ed. 1983); p. 159, 
n.75 (suppl. 1984). The Ninth Circuit did not take into account, 
discuss, or even mention petitioners’ labor market evidence, 
statistical proof, or other evidence showing that the factual 
setting rendered plaintiffs’ comparative statistics virtually 
irrelevant.

21



12

Plaintiffs allege that petitioners utilized a practice of 
"nepotism" in filling job openings. This term is defined as 
"favoritism shown to. . relatives as by giving them positions 
because of their relationship rather than on their merits.” 
Webster s Third New International Dictionary of the English 
Language Unabridged, p. 1518. The trial court found that 
although relatives were hired, they were not hired because of 
that relationship, they were hired because they were skilled and 
qualified. App. 1-105-122. The district court found that no 
preference for relatives existed. In other words, nepotism was 
not established. Despite accepting the trial court s findings 
(see App. VI-20-21), the Ninth Circuit found that the practice 
of nepotism existed. 827 F.2d at 445. Moreover, the Ninth 
Circuit found that there were 349 “nepotistic hires” during 
1970-1975. Id.* The statistics come from tables prepared by 
plaintiffs that simply listed employees who were related. These 
tables were rejected by the trial court. App. I-105.* 9 Plaintiffs 
attempted to prove that the fact relatives were hired demon­
strated they were hired because they were relatives. The trial 
court found otherwisa This finding was not clearly erroneous.

Finally, as to housing and messing practices, the trial 
court found that even if it applied the impact analysis, the 
practices were justified by business necessity. This finding 
was not clearly erroneous and should not have been vacated 
by the Ninth Circuit under the rule of Anderson v. Bessemer 
City, supra.

* The Ninth Circuit panel's finding is even more curious when one 
recalls that this same panel had previously found that nepotism 
did not exist. See 768 F.2d at 1126, 1133 (App. 111-22-23, 56).

9 There were numerous methodological problems with plaintiffs’ 
so-called "nepotism tables.” A principal flaw was that they in­
cluded gross over-counting of employees and improperly treated 
some persons as related.

22



13

B. The N in th  Circuit D ecision is a R evolu tionary D epar­
ture from  the E sta b lish ed  R u les for the A lloca tion  o f  
P roof in a D iscrim ination  Case.

1. N ew  A lloca tion  o f  Proof.

The Ninth Circuit has invented a wholly unprecedented 
rule for cases that are tried under both the treatment and 
impact analysis. The Ninth Circuit held that since petitioners 
had, in their rebuttal to plaintiffs' treatment case, offered to 
"explain the disparity,"10 * they were precluded  from challenging 
plaintiffs’ im pact showing. App. VI-5. There is absolutely no 
Supreme Court precedent supporting this holding. The only 
decision cited by the Ninth Circuit is A lbem arle Paper Co. v. 
Moody, 422 U.S. 405 (1975). Albem arle  held that if the plaintiff 
has estab lished  disparate impact of an employment test, the 
employer must prove the job relatedness of that test. It did not 
hold that the employer was precluded from showing that there 
was no impact; nor did it hold that the employer was precluded 
from attacking pla in tiffs  ’ evidence purporting to show impact.

In effect, what the Ninth Circuit has done with this new 
"rule” is to avoid the clear burden of proof requirements in 
a treatment case set forth by this court in Texas D ep t o f  
C om m unity  A ffa irs  v. Burdine. 450 U.S. 248 (1981), and fol­
lowed by the majority of other circuits thereafter." B urd ine  
holds that once a plaintiff has established a prim a facie treat­
ment case, the employer may defend by articulating — not

10 By attacking plaintiffs’ statistics and by establishing the proper 
labor market, petitioners proved no disparity existed. In addi­
tion, petitioners articulated nondiscriminatory reasons for their 
conduct.

" E.g., St. Marie v. Eastern R. A ss n. 650 F.2d 395 (2d Cir. 1981); 
McNeil v. McDonough, 648 F.2d 178 (3d Cir. 1981); Robins v. 
White-Wilson Medical Clinic. 642 F.2d 153 (5th Cir. 1981). B ut 
see Segar v. Smith, 738 F.2d 1249 (D.C. Cir. 1984), cert. den. sub 
nom, Meese v. Segar, 471 U.S. 1115 (1985) and Griffin v. Carlin, 
755 F.2d 1516 (11th Cir. 1985).

23



14

proving — a legitimate nondiscriminatory reason for his action. 
450 U.S. at 258.12 The Ninth Circuit seems to hold that once 
the reason is articulated the employer may no longer attack 
plaintiffs’ statistics and prove lack of disparate impact; further, 
the employer must now not only articulate, he must prove the 
business necessity of the reason. The result of this new rule 
is to emasculate Burdine and make it impossible for an 
employer to defend a treatment case by articulating a reason 
for his action, unless he is prepared to prove the business 
necessity of the practice.

Combined with its holding that the proof necessary to 
establish a prima facie case under the treatment and impact 
theories is identical13 (App. VI-4-5), the Ninth Circuit has 
effectively held that burden of proof is shifted to the employer 
if plaintiffs meet the “not onerous” burden14 of establishing 
a prima facie treatment case.

A case cited that could support the Ninth Circuit’s holding 
on the burden of proof is Segar v. Smith, 738 F.2d 1249 (D.C. 
Cir. 1984), cert, denied sub nom, Meese v. Segar, 471 U.S. 1115
(1985). Segar also involved a disparate treatment attack on 
the cumulative effect of many alleged practices. The District 
of Columbia Circuit held that if an employer defends by artic­
ulating the reason for his conduct, he must identify which of 
the practices causes the disparity and then prove the business 
necessity of the practice.15 Segar was followed by the Eleventh

12 While the Ninth Circuit paid lip service to this requirement, it 
simply avoided it by equating "articulation” in a treatment case 
with an admission of impact and of causation in an impact case.

13 A holding that has little or no support and conflicts with Robin­
son v. Polaroid, 732 F.2d 1010 (1st Cir. 1984) (plaintiff established 
prima facie treatment case but not impact case).

14 Burdine, supra, 450 U.S. at 253.
15 In Segar, the employer explained his conduct, as is allowed by 

Burdine, but did not refute the statistical disparity. Here,
(footnote continued on next page)

24



15

Circuit in Griffin v. Carlin. 755 F.2d 1516 (11th Cir. 1985). 
No other circuits appear to have deviated from Burdine.

2. Hiring Criteria.

In applying the impact theory to hiring criteria, the Ninth 
Circuit also altered the burdens of proof and ignored the 
district court’s findings. Griggs v. Duke Power Co., 401 U.S. 
424 (1971) requires that if the plaintiff wishes to challenge a 
hiring criterion as having a disparate impact, he must prove 
that criterion causes the impact. In Griggs, plaintiffs estab­
lished the disparate impact of a high school diploma require­
ment with unrebutted evidence that a disproportionately 
smaller percentage of blacks had diplomas. 401 U.S. at 430, 
n.6. In Dothard, supra, plaintiff’s established the disparate 
impact of a height and weight requirement by showing that 
a disproportionate number of women were less than 5’2” feet 
tall and 120 lbs. 433 U.S. at 429-30. In neither case would 
the plaintiffs have been allowed to establish an impact case 
by simply alleging the practice was discriminatory without 
independent evidence that the qualification had an impact.

Yet, this is precisely what the Ninth Circuit has done here. 
It held that since plaintiffs “challenged” petitioners’ hiring 
criteria, they were not required to take those criteria into 
account. App. VI-17, 27. Plaintiffs chose not to do so,'* both 
in their labor market statistics and in their internal compara-

defendant did both: explained the facts that rebutted plaintiffs' 
prima facie showing {e.g., that defendants hired from an external, 
not internal, labor pool: that Local 37 dispatched a gross over­
representation of nonwhites), attacked the reliability of plaintiffs' 
statistics, and offered their own statistics that showed nonwhites 
were not underrepresented in the at-issue jobs.

19 Plaintiffs chose instead to rely on their argument that virtually 
all of the at-issue jobs did not require prior skills, experience, or 
other qualifications. The trial court found otherwise and plain­
tiffs offered no evidence that took the trial court's findings into 
account.

25



16

tive statistics. They did so at their peril, because the trial 
court did find that employers hired on the basis of job-related 
criteria. App. 1-45-75, 122.

It is not surprising that plaintiffs chose not to account 
for even the most basic qualifications of the “proxy” popula­
tion. Petitioners ’ did so with their labor market analysis and 
it established that qualified nonwhite availability was closer 
to 10% than to the 50% argued by plaintiffs.

In its discussion of hiring criteria, the Ninth Circuit stated 
that it was petitioners’ burden to prove the qualified nonwhite 
component in the labor market (App. VI-17, 26), but then 
ignored petitioners’ evidence doing just that. Instead of 
addressing petitioners’ evidence that showed not only the 
qualified nonwhite component in the labor market, but that 
nonwhites were not underrepresented in the at-issue jobs, the 
Ninth Circuit skipped over this evidence and held that the 
employers were first required to prove job relatedness of the 
criteria plaintiffs were challenging. Again, this is a totally 
inappropriate shifting of the burden of proof. Combined with 
its inappropriate crediting of plaintiffs’ statistics, this means 
that a plaintiff can simply allege that there is an “imbalance” 
between two job categories (i.e., something other than 50/50), 
“allege” that any qualifications required by the employer are 
discriminatory, and thereby force on the employer the burden 
of proving the job relatedness of its criteria without plaintiff 
ever having to make the threshold showing of impact of the 
qualification at issue.

3. Sources of Employees.

The Ninth Circuit’s allocation of the burden of proof in its 
treatment of the “hiring channels” and the word-of-mouth 
recruitment issues is particularly disturbing in light of the 
actual facts in this case. The Court of Appeals seems to 
conclude that plaintiffs’ comparative statistics combined with

26



17

word-of-mouth recruiting'7 was “discriminatory.” App. 
VI-28-29. No court has held that word-of-mouth recruiting is 
per se discriminatory; the court must look first at the results 
of that practice. See Markey v. Tenneco, 707 F.2d 172 (5th Cir.
1983) and Clark v. Chrysler Corp., 673 F.2d 921 (7th Cir. 1982), 
both of which hold that the employer can defeat an attack on 
word-of-mouth recruiting by establishing that the resultant 
hiring is in line with the external labor market. Even the 
Ironworkers Local 86 case cited by the Ninth Circuit17 18 did not 
conclude that plaintiffs had established their case without 
examination of the unrebutted stark racial statistics and the 
evidence as to racial composition of the local population. The 
Ninth Circuit did not do so here. Tb reach its conclusion on 
these practices, the Ninth Circuit not only ignored the trial 
court’s findings, it committed plain error in concluding that 
the companies did not argue the practices had “no impact.” 
App. VI-30. (This error was pointed out in the Petition for 
Rehearing, App. VII.)

The Ninth Circuit then placed the burden of proof on the 
petitioners to establish why they did not hire for the at-issue 
jobs through different sources. App. VI-30.19 In forcing
17 Word-of-mouth recruiting, the practice selected by the Ninth 

Circuit for consideration, was only one method by which potential 
employees came to the attention of management. For instance, 
the record also demonstrates that walk-in applicants and referrals 
from other unions having jurisdiction over the at-issue jobs were 
used. The trial court found that the company got far more 
applications than there were available non-cannery worker jobs.

18 United States v. Ironworkers Local 86. 443 F.2d 544, 548 (9th Cir.), 
cert, denied. 404 U.S. 984 (1971) cited at App. VI-29.

19 It appears that the Ninth Circuit has concluded, in the absence of 
evidence that minorities are underutilized in the at-issue jobs, 
that the employers should have hired carpenters, machinists, 
bookkeepers, etc., through Local 37 or done what the trial court 
held was unreasonable, that is, recruit for skilled personnel in the 
thousands of square miles of tundra in Western Alaska in the 
dead of winter.

27



18

petitioners to establish why they did not utilize the cannery 
worker crews as sources for at-issue jobs (i.e., promote from 
within), hire machinists through Local 37, or recruit for skilled 
jobs in remote regions of Alaska, the Ninth Circuit is doing 
nothing less than substituting its judgment for that of the 
employer as to the best way to operate its business. This is 
a flat violation of the admonitions of this court in Furnco 
Constr. Corp. v. Waters. 438 U.S. 567, 577*78 (1978) and 
reiterated in Burdine, supra, 450 U.S. at 259.

Where the employer has not underutilized minorities in 
the at*issue jobs, it is inappropriate to adopt a voluntary 
affirmative action plan to boost the number of minorities in
those jobs. Johnson v. Transportation Agency, 480 U.S_____
94 L. Ed. 2d 615 (1987). Yet, in that very situation here, the 
Ninth Circuit is demanding that petitioners prove why they 
have not taken the different and “affirmative” steps of utilizing 
different sources for employees. The underlying assumption 
is that these steps would “maximize the number of minority 
workers” hired. Again, this violates the principle of Furnco 
and Burdine.29

III. Allowing Plaintiffs to Challenge an Entire Range of 
“Named” Employment Practices Merely Because the 
Employers’ Work Force Reflects Uneven Racial Balance 
Is an Improper Application of the Disparate Impact 
Model, Places an Unfair Burden on the Employer, and 
Exacerbates an Existing Conflict of Authority in the 
Circuits.

The only showing plaintiffs made in support of their 
impact theory attack on petitioners’ hiring practice was the 20
20 It is worth noting that if petitioners here would be prohibited 

under Johnson v. Transportation Agency from adopting an 
affirmative action plan for minorities in the at-issue jobs, it can 
hardly be said that minorities have established a prima facie case 
of disparate impact against them. See Johnson v. Transportation 
Agency, supra. 94 L. Ed. 615, 631 and n.10.

28



19

allegation that the cumulative effect of the practices identified 
was the concentration of minorities in the cannery worker jobs. 
This was shown by their comparative statistics. With but two 
exceptions,21 they offered no other statistical evidence that 
even purported to show the impact of any one of the sixteen 
hiring practices they named, independent of the others. This 
is exactly what the plaintiff did in Pouncy v. Prudential Ins. 
Co., 668 F.2d 795 (5th Cir. 1982); but unlike the Ninth Circuit 
here, the Fifth Circuit refused to allow plaintiff to misuse the 
impact theory in this way. 668 F.2d at 800-802.22 The First 
Circuit agrees with the Fifth Circuit on this issue. Robinson 
u. Polaroid Corp., 732 F.2d 1010,1014,1016 (1st Cir. 1984). The 
Ninth Circuit has now joined the Eleventh Circuit, Griffin v. 
Carlin, 755 F.2d 1516, 1522-1525 (11th Cir. 1985), in conflict 
with the Fifth and First Circuits. This Court should resolve 
the conflict.

It is an important conflict to resolve. First, it is and will 
be a recurring problem. Many businesses have, for completely 
legitimate reasons, a concentration of a protected group in a 
particular job category. See, e.g., Ste. Marie, supra, 650 F.2d 
at 401-402; Hilton v. Wyman-Gordon, supra 624 F.2d at 380; 
Rivera supra 665 F.2d at 539-542. Employers need to know 
whether the imbalance will force them, like petitioners here, 
to prove the business necessity of every practice a plaintiff 
chooses simply because plaintiff alleges they “combined” to 
"cause” that imbalance or concentration.

21 Housing space charts and tables of relatives. The latter, along 
with plaintiffs’ labor market statistics, were rejected.

22 The Ninth Circuit unpersuasively tried to distinguish the facts 
in Atonio from Pouncy by saying that plaintiffs in Atonio 
“identified” tie., named) the practices. 810 F.2d at 1486, n.6. 
The plaintiff in Pouncy did the same thing. See 668 F.2d at 
801 (names three practices).

29



20

An obvious solution for an employer is to eliminate the 
imbalance as economically as possible. Tb the extent an 
overrepresentation of minorities produced the imbalance Ie.g., 
Hilton, supra, 624 F.2d 379), many employers will simply 
reduce the number of minority workers until overrepresenta­
tion disappears. If the petitioners here had adopted this 
“solution,” e.g., by refusing to cooperate with Local 37 unless 
it dispatched only 10% nonwhites, plaintiffs would not have 
a case.

For the employer who cannot (or will not) reduce its 
minority work force in lower-paying jobs, one solution is to use 
an in-house defacto racial quota in the upper jobs until the 
percentage of minorities in the two categories is the same. 
This is directly contrary to the spirit and intent of Title VII. 
See 42 U.S.C. § 2000e-2(j) (Title VII does not require preferences 
or quotas because there is a racial imbalance). It also risks 
liability in reverse discrimination suits — particularly where 
there was no underutilization in the upper jobs. See Hammon, 
supra, 826 F.2d 73 (D.C. Cir. 1987) (voluntary affirmative action 
plan set aside because no underutilization shown).

Second, the impact model was designed to focus on a 
particular requirement, usually a selection criterion, that can 
be measurably shown to cause an adverse impact, e.g., Pouncy, 
supra, 668 F.2d at 801; see discussion, infra, IV. Most of the 
practices that plaintiffs here allege combined to cause the 
imbalance {e.g., requiring cannery workers to cut the grass; 
restrictions on fraternization; failure to post) are far from this 
conception and can be, at best, only tangentially connected 
to the reasons minorities are overrepresented in the cannery 
worker jobs. Indeed, plaintiffs did not offer proof designed 
to show the impact of any one, independent of the others.

This leads to a third and very important reason this 
conflict should be resolved in favor of petitioners: The more 
practices plaintiffs can “name” or “identify” as allegedly

30



21

causing the concentration, the more impossible becomes the 
employers’ burden. For if the court finds that imbalance is 
sufficient to require the employer to prove business necessity, 
he could be forced to justify every practice identified. Courts 
may require “validation” under the EEOC Guidelines for 
Employee Selection Procedures — an enormously expensive 
proposition for one “procedure,” but prohibitive for several.

The unfair risk and burden the employer faces is best 
illustrated by petitioners’ situation: they have demonstrated 
to the satisfaction of district court and the Court of Appeals 
the business necessity of their rehire preference, an English 
language requirement, and (although the Ninth Circuit would 
disagree) of their hiring criteria. But they are still in court 
— because plaintiff named other practices that the Ninth 
Circuit says must also be justified, even though plaintiffs have 
not offered any evidence establishing that these remaining 
practices “caused” the imbalance, as opposed to the ones 
already proven to be a business necessity. This is exactly the 
situation the Fifth Circuit predicted in Pouncy would occur: 
allowing “disparate impact of one element to require validation 
of other elements having no adverse effects.” 668 F.2d at 801.

IV. There is a Substantial Conflict in the Circuits as to 
Whether the Disparate Impact Analysis May Be Applied 
to Subjective Decision Making and Other Practices That 
Do Not Act as “Automatic Disqualifiers.”21

The Ninth Circuit has now erroneously followed the Sixth, 
Tfenth, Eleventh, and the District of Columbia Circuits in 
applying the impact analysis to subjective practices and 
criteria. Rowe v. Cleveland Pneumatic Co., 690 F.2d 88 (6th 
Cir. 1982); Hawkins v. Bounds, 752 F.2d 500 (10th Cir. 1985); 
Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1985); Segarv. Smith, 23

23 See Dothard v. Rawlinson, supra, 433 U.S. at 338 (Rehnquist. J.. 
concurring).

31



22

738 F.2d 1249 (D.C. Cir. 1984), cert, denied sub nom. Meese v. 
Segar, 471 U.S. 1115 (1985). The Fourth, Fifth. Seventh, and 
Eighth Circuits do not apply the impact analysis to subjective 
practices, although there are some conflicts within some of 
those circuits. EEOC v. Federal Reserve Bank of Richmond, 
698 F.2d 633 (4th Cir. 1983), rev'd on other grounds sub nom, 
Cooper v. Federal Reserve Bank of Richmond, 467 U.S. 867
(1984); Pope v. City of Hickory, 679 F.2d 20 (4th Cir. 1982); 
Bunch v. Bullard, 795 F.2d 384 (5th Cir. 1986); Vuvanich v. 
Republic Nat'l Bank of Dallas, 723 F.2d 1195 (5th Cir. 1984), 
cert denied, 469 U.S. 1073 (1984); Pegues v. Mississippi State 
Employment Serv., 699 F.2d 760 (5th Cir.), cert, denied, 464 
U.S. 991 (1983); Pouncy v. Prudential Ins., 668 F.2d 795 (5th 
Cir. 1982); Griffin v. Board of Regents, 795 F.2d 1281 (7th Cir.
1986); Harris v. Ford Motor Co., 651 F.2d 609 (8th Cir. 1981).

This Court has presently granted certiorari and heard 
argument (January 21, 1988) in Watson v. Fort Worth Bank 
& Trust, 798 F.2d 791 (5th Cir. 1986), to review this important 
question. No. 86-6139.

This case illustrates a broader application of the issue than 
Watson, inasmuch as it poses several applications of the 
analysis, e.g., word-of-mouth recruitment, “separate hiring 
channels," labeling, and the effect of challenges to the cumula­
tive effect of multiple practices.

The decision in Watson may reach some of the issues raised 
by petitioners. While this Court may wish to consider ruling 
on this petition after that decision is issued, this case presents 
other important issues and the granting of the writ should not 
be delayed.

CONCLUSION

Although they mounted a broad scale attack, plaintiffs 
were unable to prove any instance of individual or of class­
wide disparate treatment of minorities in any aspect of the

32



23

employment relationship. Plaintiffs’ fallback position was to 
allege under the disparate impact theory that petitioners' 
practices combined to cause unintentional discrimination. 
Without significant evidence of unfair treatment, plaintiffs 
were left to prove their impact case with comparative statistics. 
These statistics did nothing more than show “imbalance’’ — 
that there was an overabundance of minority workers in the 
cannery worker jobs. They proved nothing as to the jobs 
at issue. Plaintiffs’ statistics were a simplistic reflection of 
the fact that Local 37 dispatched an oversupply of minority 
workers. In other words, but for the fact these petitioners 
fulfilled their collective bargaining responsibilities with Local 
37, plaintiffs would not have an impact case.

The trial court saw through plaintiffs' theory; the Ninth 
Circuit did not. Tb justify its decision, however, the Ninth 
Circuit issued an opinion that has ominous implications not 
only for petitioners, but for litigation of all discrimination 
cases and for the conduct of everyday business.

This case presents a meaningful opportunity for this Court 
not only to correct an erroneous decision, but to finally estab­
lish the proper boundaries of the impact analysis, to clarify 
the role of statistics and the proper allocation of the burdens 
of proof in applying that analysis, and to resolve numerous 
and longstanding circuit conflicts in this important area of law.

For the foregoing reasons, this petition should be granted.

Respectfully submitted,

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ik k elbo rg . B roz. 
W ells & F ryer

. „ , _ _ , Attorneys for Petitioners* Counsel of Record

33



No. 87-1387

IN THE SUPREME COURT 
of tho

UNITED STATES 

October Term, 1087

WARDS COVE PACKING COMPANY, INC. and 
CASTLE, & COOKE, INC.

Petitioner*.
vs.

FRANK ATONIO, EUGENE BACLIG, RANDY del FIERRO, 
CLARKE KIDO, LESTER KURAMOTO, ALAN LEW, 

CURTIS LEW, ROBERT MORRIS,,JOAQUIN ARRUIZA, 
BARBARA VIERNES, aa administratrix of the estate of 

Gene Alien Viernea, and all others similarly situated,
Respondents

BRIEF IN OPPOSITION TO 
PETITION FOR WRIT OF CERTIORARI TO 

THE UNITED STATES COURT OF APPEALS 
FOR THE NINTH CIRCUIT

Abraham A. Arditi
Northwest Labor and Employment Law Office 

900 Hoge Building 
Seattle, Washington 98104 

1-206-623-1590
Counsel of Record for Respondents

Bobbe Jean Bridge 
Garvey, Schubert &. Barer 
Waterfront Place Building 

1011 Western Avenue 
Seattle, WA 98101 

1-206-464-3939 
Counsel for Respondents

35



TABLE OF CONTENTS

Statutes................................................................................. 1
Statement of the Case.......................................................... 1
Summary of Argument.......................................................... 4
Reasons for Denying Writ.....................................................5
Conclusion.......................................................................... 14

TABLE OF AUTHORITY 

Table of Cases

Connecticut v. Teal, 457 U.S. 440 (1982)................ 5-6, 8, 9
Domingo u. New England Fish Co., 727 F.2d 1479
(9th Cir. 1984), modified 742 F.2d 520 (1984).............7, 11
Dothard v. Rawlinson, 433 U.S. 321 (1977)..................9, 12
Franks u. Bowman Transportation Co.,
424 U.S. 747 (1976).................................................... 10, 12
Griggs v. Duke Power Co., 401 U.S. 424 (1971)...........6, 12
Hazelwood School District v. United States,
433 U.S. 299 (1977)........................................................... 8
Satty v. Nashville Gas Co., 434 U.S. 136 (1977)................ 5
Teamsters v. United States, 431 U.S. 324 (1977)...........6, 7
Watson v. Fort Worth Bank and Trust,
___ U.S_____(1986) (No. 86-6139).................................13

Statutes

42 U.S.C. §2000e-2(a)....................................................1, 5
42 U.S.C. §2000e-2(h)....................................................... 6

i

36



1

STATUTES
This case involves 42 U.S.C. § 2000e-2(a), which appears in 

the Appendix to the petition for certiorari filed by the employees.

STATEMENT OF THE CASE

This class action challenges a pattern of racial segregation 
in jobs, housing and messing at several Alaska salmon canner­
ies. Two employers seek review of a decision by the court of 
appeals recognizing disparate impact claims under Title VII 
against them. The employees have also filed a petition for 
certiorari, but on other issues. For the Court’s convenience, the 
employees reiterate a portion of the statement of the case from 
their petition here.

The Alaska salmon canning industry has been heavily non­
white since the turn of the century. Because the canneries are 
generally located in remote areas, they hire migrant, seasonal 
workers, who live in company housing. The percentage of non­
white employees in the industry was 40% to 70% during 1906- 
1978, stabilizing at about 47% to 50% toward the end of this 
period. (Appendix, p. A-250.') The work force at the canneries 
here reflects industry-wide figures, for it has been about 43% 
non-white overall since 1970. (Ex. 583-87). But while the 
percentage of non-whites overall at the canneries is high, jobs 
are racially stratified. Non-whites are concentrated in the 
lowest paying jobs, while whites clearly dominate the higher 
paying jobs.

The degree of segregation varies somewhat, but the admin­
istrative, machinist, fisherman, tender, carpenter, beach gang, 
office and store departments are all white or heavily white. In 
contrast, the largest department—namely, cannery worker—is 
heavily non-white. At some canneries, the laborer department 
is also heavily non-white.

For example, at Bumble Bee cannery during 1971-80, 
seven departments—in which there were 342 new hires—were 
at least 90% white, although the cannery worker department

'Citations are to the appendix to the petition for certiorari filed by the 
employees rather than the employers.

37



2

was 52% non-white. (Exhibit A-278 SN, Table 4.) At Red 
Salmon cannery, four departments—in which there were 146 
new hires—were at least 94% white, although the cannery 
worker department was 64% non-white. (Exhibit A-278 RS, 
Table 4.) At Wards Cove cannery, six departments—in which 
there were 228 new hires—were at least 93% white, although 
the cannery worker department was 31% non-white. (Exhibit 
A-278 WC, Table 4.) At Ekuk cannery, five departments—in 
which there were 111 new hires—were at least 90% white, 
although the cannery worker deaprtment was 67% non-white. 
(Exhibit A-278 EK, Table 4.) At Alitak cannery, seven depart­
ments—in which there were 299 new hires—were at least 62% 
white, while the cannery worker department was 65% non­
white. (Exhibit A-278 AK, Table 4.)

The pattern of segregation is matched by express race 
labelling of jobs, bunkhouses and messhalls. Company records 
refer to “Filipino cannery workers,” “Native cannery workers,” 
the “Filipino union,” “Philippine Bunkhouse,” “Native Galley 
Cook ” and “Filipino Mess.” (Appendix, p. A-283-284.) A mail 
slot in the office at one cannery is marked “Oriental bunk- 
house.” (Appendix, A-285-286.) The president of one employer 
testified Alaska Native crews are race labelled “for mere ease 
or habit of identification.” (RT 1143.) Employee badge num­
bers are assigned along racial lines. (Appendix, p. A-284-285.) 
Even the salmon butchering machine has a name with racial 
overtones, the “Iron Chink.” (Appendix, p. A-37.) Far from 
being incidental, the “(rjace labelling is pervasive at the salmon 
canneries.” (Appendix, p. A-37.)

The pattern of job segregation is enforced by several 
practices.1

First, race labelling “operates as a headwind” to advance­
ment of non-whites, because it conveys a “message” that they 
need not apply for upper level jobs. (See Appendix, p. A-37 and 
43.)

Second, the employers use essentially segregated hiring 
channels for different jobs, which prevent non-whites from 
competing on the basis of qualifications with whites. The *

*The employers say the employees challenged 16 practices which fostered 
job segregation, but only eight were treated by the court of appeals.

38



3

employers recruit from non-white sources such as Alaska 
Native villages, foremen of Asian descent and the heavily 
Filipino Local 37, ILWU, but only for the lowest paying jobs. 
(Appendix, p. A-31-32; Revised Pretrial Order, p. 14.)

Third, to fill higher-paying jobs, the employers rely on 
informal word-of-mouth recruitment among friends and rela­
tives of white foremen and superintendents. (Appendix, p. A-32.) 
The employers neither publicize vacancies for upper-level jobs 
nor promote from non-white to white jobs, so the effect of 
segregated hiring channels is aggravated. (Appendix, p. A-241- 
242; Revised Pretrial Order, p. 16-19.)

Fourth, the employers do not use objective job qualifica­
tions for jobs at issue. They retained an expert who prepared 
qualifications for litigation, which the district court found could 
be “reasonably required.” (Appendix, p. A-26.) But the expert 
testified the qualifications were never actually applied. (RT 
3113; see also RT 3067.) The anecdotal evidence also indicates 
they were never actually applied. The use of subjective 
qualifications has a disparate impact on non-whites, which 
prevents them from competing on equal terms for upper-level 
jobs. (See Appendix, p. A-28-30.)

Fifth, there is pervasive nepotism at the canneries, which 
contributes to the racial segregation, since white supervisors 
control the upper level jobs. “(0|f 349 nepotistic hires in four 
upper-level jobs during 1970-75,332 were of whites, 17 were of 
non-whites.” (Appendix, p. A-24.)

Sixth, the employers give re-hire preference to employees 
in their old jobs, a practice which perpetuates the segregation 
by race in jobs. The court of appeals affirmed a finding the 
practice was justified by business necessity, although no 
evidence of business necessity was offered in the district court. 
(See Appendix, p. 36.)

Seventh, the employers maintain racially segregated bunk- 
houses, a practice which “aggravate(s| the isolation of the non­
white workers from the ‘web of information’ spread by word of 
mouth among white people about the better paying jobs.” 
(Appendix, p. A-41.)

Eighth, the segregated messhalls have a similar effect, 
again enforcing the pattern of job segregation. (Appendix, p. A- 
42-32.)

39



4

The employers justify the job segregation by arguing they 
hire too many non-whites in the lower paying jobs, rather than 
too few in the upper-level jobs. The centerpiece of their 
approach is a labor market comparison, which assumes non­
white availability of only about 10%, even though non-whites 
comprise about 48% of employees in the industry. (RT1870 et 
seq.) The job segregation is so graphic even the economist 
hired by the employers testified non-whites were absent from 
certain jobs at statistically significant levels. (RT 1871-73; see 
also RT 1875.) But overall, he testified there was no pattern or 
practice of discrimination, because recruiting through largely 
non-white sources for low paying jobs distorted the racial 
composition of the labor pool

Following trial, the district court dismissed all claims. A 
panel of the court of appeals initially affirmed, but its opinion 
was withdrawn, when rehearing en banc was granted. Sitting en 
banc, the court of appeals held the disparate impact approach 
could be applied to the challenged practices. It then returned 
the appeal to the panel to apply this ruling.

The panel reversed dismissal of disparate impact challenges 
to nepotism, lack of objective qualifications, use of separate 
hiring channels, word of mouth recruitment, segregation in 
housing, segregation in messing and race labeling. Because it 
was not clear what—if any—job qualifications were actually 
applied, the court of appeals remanded for findings on this 
issue. It affirmed dismissal of a disparate impact challenge to 
use of certain re-hire preferences, because of a finding they 
were justified by business necessity.

SUMMARY OF ARGUMENT
The Court should deny the writ, since virtually every issue 

raised by the employers has been foreclosed by controlling 
decisions of the Court. The departures the employers cite from 
decisions of the Court are based on a mis-reading of the 
opinions below.

40



5

REASONS FOR DENYING THE WRIT

1.
The Ruling of the Court of Appeals on the

Prima Fade Case is Consistent With Controlling 
Decision of this Court

The court of appeals correctly held that the job segregation 
statistics—as well as other evidence—established disparate 
impact, regardless of findings on the percentage of non-whites 
in the labor market area. Because this holding is consistent 
with controlling decisions of this Court, there is no reason to 
grant certiorari.

First, by its terms, Title VII makes segregation in jobs by 
race unlawful.

It shall be an unlawful employment practice for an 
employer—
to limit, segregate, or classify his employees or applicants 
for employment in any way which would deprive or tend to 
deprive any Individual of employment opportunities or 
otherwise adversely affect his status as an employee, 
because of such individual’s race . . .

42 U.S.C. §2000e-2(a)(2). (Emphasis added.) This Court has 
recognized claims of disparate impact under this section. 
Connecticut v. Teal, 457 U.S. 440, 445-46 (1982); Satty v. 
Nashville Gas Co., 434 U.S. 136, 144 (1977).

Second, because the disparate impact approach focuses 
on lost opportunities rather than simply lost jobs, labor 
market comparisons with those hired are irrelevant. This 
Court has already rejected the "bottom line” defense in 
disparate impact cases in Connecticut v. Teal, supra. 
Since a labor market argument is one form of a "bottom 
line” defense, Teal is controlling here.

In considering claims of disparate impact under § 703(a)(2) 
this Court has consistently focused on employment and 
promotion requirements that create a discriminatory bar 
to opportunities. This Court has never read § 703(a)(2) as 
requiring the focus to be placed instead on the overall 
number of minority or female applicants actually hired or

41



6

promoted. Thus Dothard v. Rawlinson, 433 U.S. 321
(1977), found that minimum statutory height and weight 
requirements for correctional counselors were the sort of 
arbitrary barrier to equal employment opportunity for 
women forbidden by Title VII. Although we noted in
{>assing that women constituted 36.89 percent of the 
abor force and only 12.9 percent of correctional counselor 
)ositions, our focus was not on this “bottom line.” We 
ocused instead on the disparate effect that the minimum 
leight and weight standards had on applicants: classify- 
ng far more women than men as ineligible for employ­
ment. Id., at 329-330 and n. 12. Similarly, in Albermarle 
Paper Co. v. Moody, supra, the action was remanded to 
allow the employer to attempt to show that the tests that 
he had given to his employees for promotion were job 
related. We did not suggest that by promoting a sufficient 
number of the black employees who passed the examina­
tion, the employer could avoid this burden. See 422 U.S., 
at 436. See also New York Transit Authority v. Beezer, 440 
U.S. 568, 584, (1979).

In short, the District Court’s dismissal of respondent’s 
claim cannot be supported on the basis that respondents 
failed to establish a prima facie case of employment 
discrimination under the terms of § 703(a)(2). The sug­
gestion that disparate impact should be measured only at 
the bottom line ignores the fact that Title VII guarantees 
these individual respondents the opportunity to compete 
equally with white workers on the basis of job related 
criteria.

Connecticut v. Teal, supra at p. 450-51. (Emphasis in original.)
Third, beginning with Griggs, this Court has endorsed 

disparate impact attacks on practices which foster job segrega­
tion. Griggs involved education and testing requirements, 
which operated as “built in headwinds” to transfers in a plant 
where “Negroes were employed only in the labor department,” 
while “only Whites were employed” in the other four depart­
ments. Griggs v. Duke Power Co., 401 U.S. 424, 427, 432
(1971). Similarly, in Teamsters v. United States, 431 U.S. 324
(1977), this Court observed that a seniority system which 
enforced a pattern of job segregation would have been subject 
to a disparate impact challenge but for the exemption for 
seniority systems in 42 U.S.C. §2000e-2(h).

42



7

The vice of this (seniority! arrangement, as found by 
the District Court and the Court of Appeals, was that it 
“locked” minority workers into inferior jobs and perpetu­
ated prior discrimination by discouraging transfers to 
jobs as line drivers.

• • • •

Were it not for § 703(h), the seniority system in this 
case would seem to fall under the Griggs rationale.

Teamsters v. United States, supra at 344, 349.

Fourth, while the job segregation statistics illustrate the 
effect of the challenged practices, they are not the only 
evidence of disparate impact Plaintiffs offered separate statis­
tics on nepotism, which showed that fully 332 of 349 nepotistic 
hires in four upper-level departments during 1970-75 went to 
whites. (Appendix, p. A-24.) Similarly, the effect of segregated 
hiring channels is obvious, as the court of appeals pointed out.

(T|he companies sought cannery workers in Native villages 
and through dispatches from ILWU Local 37, thus 
securing a work force for the lowest paying jobs which was 
predominantly Alaska Native and Filipino. For other 
departments the companies relied on informal word-of- 
mouth recruitment by predominantly white superinten­
dents and foremen, who recruited primarily white em­
ployees. That such practices can cause a discriminatory 
impact is obvious.

(Appendix, p. A-31-32.) See also Domingo v. New England Fish 
Co., 727 F.2d 1479,1435-36 (9th Cir. 1984),modified!42 F.2d 
520 (1984). Beyond this, the employers conceded the effect of 
the practice in the court of appeals by arguing that the 
concentration of non-whites in lower level jobs results from 
recruiting in Alaska Native villages and through Local 37 for 
cannery worker jobs.5 Similarly, the employers acknowledged 
in the court of appeals that the racial imbalance in jobs results 
from the inability of non-whites to meet the undisputedly 
subjective qualifications they impose.* 4 But this is simply

’Brief of Appellees, p. 8 and 29.

4Brief of Appellees, p. 27-28.

43



8

another way of saying that the qualifications disqualify non­
whites at a higher rate than whites, an observation which lies at 
the heart of a disparate impact violation.

Fifth, whether the employers fill the upper level jobs with 
new hires rather than promoting from within is irrelevent to the 
existence of disparate impact. A non-white who is condemned 
to a menial job because of an employer’s recruitment practices 
is no less a victim of discrimination than a non-white who is 
denied a promotion once hired. Segregated hiring channels cut 
non-whites off from opportunities in the better jobs. Because 
the abilities of those recruited through different channels are 
not compared, non-whites are foreclosed from competing 
effectively on the basis of qualifications for upper-level jobs.

2.
The Court of Appeals Correctly Applied the Disparate 
Impact Approach in a Way Which is Consistent with 

Controlling Decisions of this Court

Because the court of appeals applied the disparate impact 
approach in a way which is consistent with controlling decisions 
of this Court, certiorari is inappropriate.

a.
The Court of Appeals Did Not Override 

the Findings of the District Coui;t
The employers argue that the court of appeals ignored the 

district court’s findings on the labor market. But because a 
labor market defense is precisely the sort of “bottom line” 
argument this Court has rejected in disparate impact cases, the 
findings were irrelevant. {See page 5-6, supra.) Connecticut v. 
Teal, supra. Beyond this, statistics offered by the employers 
establish disparate impact, as is apparent from the job 
segregation figures—all of which are taken from defense 
exhibits—which are cited above. (See page 1-2, supra.) Even 
the economist who testified for the employers testified that 
there was statistically significant underrepresentation of non­
whites in certain upper-level jobs. (See page 3, supra.) The 
employers argue Hazelwood School District v. United States, 
433 U.S. 299 (1977), requires a labor market comparison. But 

44



9

since it involves neither a disparate impact nor a job segrega­
tion claim, it is not pertinent here.

The employers also maintain the court of appeals ignored 
the district court’s findings on nepotism. However, the district 
court found there was a“pervasive incidence of nepotism in the 
canneries.” (Appendix, p. A-315.) Far from overturning this 
finding, the court of appeals simply corrected the district 
court’s misunderstanding of the term “nepotism.” (Appendix, 
p. A-23-24.) Since the only authority the employers cite is 
Webster’s, this issue hardly merits a grant of certiorari.

b.
The Court of Appeals Followed the Allocation of 

the Burden of Proof Set By Controlling 
Decisions of this Court

The court of appeals held that the employees made a prima 
facie case of disparate impact by:

(1) show(ing| a significant disparate impact on a protected 
class, (2) identifyjingj specific employment practices or 
selection criteria ana(3) showjingj the casual relationship 
between the identified practices and the impact

(Appendix, p. A-71, A-81, A-87-88.) These elements track 
precisely the guidelines this Court has set for a prima facie case 
on a disparate impact claim. Connecticut v. Teal, supra at 446 
(“|t)o establish a prima facie case.. .  a plaintiff must show that 
the facially neutral employment practice had a significant 
discriminatory impact”); Dothard v. Rawlinsort, 433 U.S. 321, 
329 (1977) (“to establish a prima facies case... a plaintiff need 
only show that the facially neutral standards select applicants 
for hire in a significantly discriminatory pattern”). Any claim 
that the court of appeals required a lesser showing is based on a 
misreading of its opinions.

c.
The Court of Appeals Correctly Allocated the 

Burden of Proof on Qualifications Here

45

The court of appeals correctly allocated the burden of proof 
on qualifications here. While the employers maintain the issue



10

is one of general importance, it in fact arises largely from 
circumstances peculiar to this case.

The employers argue that the employees bear the burden of 
offering statistics on the percentage of qualified non-whites. 
But this Court has held that only “non-discriminatory stan­
dards actually applied’ by the employer are pertinent in a Title 
VII case. Franks v. Bowman Transporation Co., 424 U.S. 747, 
773 n. 32 (1976). (Emphasis in the original.) Since the 
employers never articulated what—if any—qualifications they 
actually applied, the employees could not offer statistics on 
qualified non-whites.

The employers called an expert witness to testify at trial 
about qualifications, but he admitted the qualifications he 
devised had never been applied.

THE COURT: All right. Mr. DeFrance, in this case, I 
believe you have already testified that the Defendants 
have not adopted, to your knowledge, the minimum 
qualifications that you recommended; is that correct?

THE WITNESS: That’s correct. I don’t know that they 
have ever been adopted.

(RT 3113; see also RT 3067.) One management employee who 
recruited in nearly all upper level jobs testified,

Q. But there were not set qualifications a person had to 
meet?
A. No.
Q. Pardon?
A. No.

(RT 622-23, 627-30 and 637.) A cannery superintendent 
conceded in deposition testimony offered at trial,

Q. You don’t have any written job qualifications at 
Bumble Bee, do you?
A. No.
Q. Have you ever had them?
A. No.
Q. You just rely on your own judgment and the judgment 
of the foreman who is hiring?

46



11

(Dep. Leonardo-1978 p. 2 and 46-47.) The employers neither 
give tests nor impose education requirements.5 They make no 
attempt to hire on standardized qualifications, even when 
several people hire for the same job.* * They have no written 
qualifications for any job at issue/ Nor with rare exceptions, 
do they have written job descriptions.* Witness after witness 
called by the employers acknowledged on cross-examination 
that there were no objective qualifications.9 The district court 
cited evidence of a “general lack of objective job qualifica­
tions.” (Appendix, p. A-317.) The employees offered anecdotal 
evidence that individuals were hired on much lower qualifica­
tions than the employers asserted at trial. (See Appendix, p. A- 
29.) Understandably concerned, the court of appeals remanded 
for findings on what—if any—objective qualficiations were 
actually applied. (Appendix, p. A-28.) Where job qualifications 
are unknown, it is unrealistic to require the employees to prove 
the percentage of qualified non-whites. Domingo v. New 
England Fish Co., supra at 1437 n.4.

Beyond this, the employees were not required to show that 
they were qualified under selection criteria which were them­

*Ex. 156-159; Dep. A-W. BrindIe-1975, p. 30; Dep. Jorgenson, p. 7-8; 
Dep. Snyder, p. 12-13; Dep. Rohrer, p. 43; Dep. Leonardo-1975, p. 21; Dep. 
Leonardo-1978, p. 46-47; Dep. W.F. Brindle-1978, p. 49. One cannery has 
given home-made tests but has neither kept scores nor recorded the results, 
since “the proof is in the pudding.” (Ex. 160.)

*Dep. Rohrer, p. 43; Dep. Jorgesen, p. 8; Dep. Snyder, p. 13.
:R T  2365,2548,2569,2758,2805,2819 and 3316; Dep. Gilbert-1980, p. 

10; Dep. W.F. Brindle-1978, p. 49; Dep. Snyder, p. 13.
*Dep. Gilbert-1980, p. 11; Dep. A. Brindle-1975, p. 29.
*RT 2365, 2548, 2569, 2617, 2642 and 3181; Dep. Aiello, p. 19. Aside 

from expert testimony, the only arguable listing of qualifications came in 
certain interrogatory answers. (Ex. 68-72.) However, the qualifications given 
are more modest than those asserted at trial, are often purely subjective and 
in any case were not actually imposed. As one cannery superintendent 
testified,

Q. So (the interrogatory answers| are your ideal for qualifications? 
A. Yes.
Q. And Alitak may have hired on lower qualifications?
A. Oh, Yes. W e  always shot for the best
Q. So your answer to Interrogatory 20.C does not get job qualifications 

as they were actually imposed at Alitak from 1970 onward?
A. Right

(Dep. W.F. Brindle-1978, p. 12 and 14.)

47



12

selves discriminatory. Once they made a prima facie case of the 
disparate impact of subjective criteria, they were relieved of 
proving that they met these qualifications. To hold otherwise 
would have meant the plaintiff in Griggs had to have a high 
school diploma before challenging the discriminatory nature of 
the high school diploma requirement. Once again, only a failure 
to meet “non-discriminatory” standards is pertinent under 
Title VII. Franks v. Bowman Transportation Co., supra at 773 
n. 32.

d.
The Court of Appeals Correctly Allocated the Burden 

of Showing Business Necessary for 
Separate Hiring Channels

The employers complain that the court of appeals unfairly 
placed on them the burden of justifying their practice of using 
segregated hiring channels. But because the employees estab­
lished the disparate impact of separate hiring channels, it fell 
to the employers to establish the business necessity of the 
practice. Dothard u. Rawlinson, supra at 332 n. 14; Griggs u. 
Duke Power Co., supra at 431.

3 .
The Employees Have Shown the Impact of Each 

Practice They Challenge

The employers maintain that the employees challenge the 
cumulative effect of a variety of practices with no more 
evidence than job segregation statistics. But in so doing, the 
employers mis-characterize the rulings below. This Court 
should not grant certiorari to review an issue which is absent 
from the case.

Sitting en banc, the court of appeals wrote,

We note that a related concern is that the “impact 
model is not the appropriate vehicle from which to launch 
a wide ranging attack on the cumulative effect of a 
company’s employment practices.” Spaulding, 740 F.2d 
at 707. However, this is not such a case. The class has not

48



13

simply complained about the overall consequences of a 
collection of unidentified practices; rather it has identified 
specific employment practices which cause adverse im­
pact. These specific practices which cause adverse impact 
may be considered individually and collectively.

(Appendix, p. A-91 n. 6.) (Emphasis added.) It also noted the 
employers conceded that the challenged practices caused the 
disparate impact.

The statistics provide evidence of a significant disparate 
impact and the challenged practices are agreed to cause 
disparate impact

(Appendix, p. A-81.) (Emphasis added.)
On return to the panel, the court of appeals cited proof of or 

a concession about the disparate impact of each challenged 
practice. From statistics showing 332 of 349 hires in upper- 
level departments of relatives were of whites it concluded that 
nepotism had a disparate impact (Appendix, p. A-24.) When 
treating subjective criteria, it observed, “The companies 
concede the casual relation between their hiring criteria and
the number of nonwhites in the at-issue jobs__” (Appendix, p.
A-28.) Similarly, it commented it was “obvious” that recruiting 
for menial jobs from non-white sources while soliciting white 
applicants by word of mouth had a disparate impact. (Appen­
dix, p. A-32.)

4 .
The Court Need Not Grant Certiorari on the Application 
of the Disparate Impact Approach to Subjective Practices

While the Court has granted certiorari in Watson v. Fort
Worth Bank and Trust___ U.S-------(1986) (No. 86-6139), to
decide the suitability of the disparate impact approach to 
subjective practices, the employers say this case involves 
somewhat different issues. Under these circumstances, the 
Court should not grant certiorari here.

49



14

CONCLUSION

The Court should deny the writ of certiorari.

Respectfully submitted,

Abraham A. Arditi 
Attorney for Respondents

50



No. 87-1387

IN THE

Supreme Court of the United States

October Term, 1988

W a r d s  Cove P acking Company. Inc., 
Castle & Cooke, Inc.,

Petitioners,

v.

F rank  A tonio. et al., 
Respondents.

BRIEF OF PETITIONERS

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ikkelborg. B roz, 
W ells & Fryer 
Suite 3300
1001 Fourth Avenue Plaza 
Seattle, Washington 98154 
(206) 623-5890
Attorneys for Petitioners

* Counsel of Record 

September 9, 1988

51



1

1. Does statistical evidence that shows only a concen­
tration of minorities in jobs not at issue fail as a matter of law 
to establish disparate impact of hiring practices where the 
employer hires for at-issue jobs from outside his own work 
force, does not promote-from-within or provide training for such 
jobs, and where minorities are not underrepresented in the 
at-issue jobs?

2. In applying the disparate impact analysis, did the 
Ninth Circuit improperly shift the burden of proof to 
petitioners?

3. Did the Ninth Circuit commit error in allowing plain­
tiffs to challenge the cumulative effect of a wide range of non- 
racially motivated employment practices under the disparate 
impact model?

QUESTIONS PRESENTED

52



11

Petitioners are Wards Cove Packing Co., Inc., and Castle 
& Cooke, Inc., who were defendants in the trial court pro­
ceeding. (Claims against a third defendant, Columbia Wards 
Fisheries, were dismissed. This was affirmed on appeal. 
See fn. 1 infra.)

Respondents are Frank Atonio, Eugene Baclig, Randy del 
Fierro, Clarke Kido, Lester Kuramoto, Alan Lew, Curtis Lew, 
Joaquin Arruiza, and Barbara Viernes (as administratrix of 
the Estate of Gene Allen Viernes), who were individual plain­
tiffs and representatives of a class of all nonwhite employees 
in the trial court proceeding.

LIST OF PARTIES

53



Ill

Questions Presented. . . . . . . . . . .  i

List of Parties . . d

Table of Contents....................................................................Id

List of Authorities.................................................................vd

Opinions Below.............................  1

Jurisdiction ...........................  . . . . . . .  2

Pertinent S ta tu te .........................     2

Statement of The C ase ......................  3

A. Nature of the Case...........................^ .................... 3

B. Material F a c ts ...................   4

C. Court of Appeals Rulings................  11

Summary of A rgum ent.............................................   13

Argument................................................................................ 15

I. Respondents Failed to Prove Discrimination Under
the Disparate Impact Theory ...............................  • • 15

A. Statistical Evidence That Shows Only a Con­
centration of Minority Employees in Jobs Not 
at Issue Fails As a Matter of Law to Estabdsh 
Disparate Impact of Hiring Practices Where the 
Employer Fills the At-Issue Jobs From Outside 
His Own Work Force, Does Not Promote From 
Within or Provide TVaining for Such Jobs, and 
Where Minority Employees Are Not Underrepre­
sented in the At-Issue Jobs According to a Labor 
Market Analysis Accepted by the TYial Court. . . . 15

TABLE OF CONTENTS

Page

54



iv

TABLE OF CONTENTS, (continued)

1. Petitioners’ Labor Market Analysis Was
More Probative Than Respondents’ Com­
parative Statistics, Refuted Any Showing 
of Disparate Impact, and Should Not Have 
Been Ignored by the Ninth C ircu it............ 16

2. Allowing Proof of Racial Imbalance to
Establish Disparate Impact is Inimical to 
the Desirable Purposes of Title VII, and 
Provides an Unreasonable and Unworkable 
Standard in Practice..................................... 21

3. Respondents’ "Separate Hiring Channels’’ 
Argument is a Red Herring: Where Re­
spondents Have Failed to Show That the 
Petitioners’ Practices for Filling Jobs Not 
at Issue Either Intentionally or in Effect 
Excluded Nonwhites from Jobs At Issue,
Those Practices Are Irrelevant...................... 22

B. Respondents Failed to Prove a Practice of
"Nepotism" Existed, Their Statistics Pur­
porting to Show Its Existence Were Properly 
Rejected, and The Ninth Circuit Committed 
Error in Finding Otherwise................................ 24

C. Respondents Failed to Prove Unlawful Dis­
crimination Under the Disparate Impact 
Theory in Housing, Messing, or So-Called 
"Racial Labeling".........................................  27

55



V

TABLE OF CONTENTS, (continued)

Page

II. Allowing Respondents to Challenge the Cumulative 
Effect Of An Entire Range of Non-Racially Moti­
vated Employment Practices Based Merely On A 
Showing The Petitioners’ Work Force Reflects An 
Uneven Racial Balance Is An Improper Applica­
tion Of The Disparate Impact Model, Unfairly 
Allocates The Burdens of Proof And Encourages 
Conduct At Odds With The Purposes of Title VII . 30

A. Respondents Are Required to Prove the 
Causal Effect of Each Practice They Choose
to Challenge Under the Impact Model............ 30

B. Dispensing With the Causation Requirement
for Plaintiffs Places an Unfair Burden on 
Employers, and Encourages Conduct at Odds 
With Title VII’s Purpose...................................34

III. The Ninth Circuit Improperly Shifted The Burden
Of Proof Tb Petitioners...............................................37

A. Respondents Did Not Meet the Initial Re­
quirements to Establish An Impact Case . . . . 37

B. The Ninth Circuit’s Decision Conflicts With
the Order of Proof Requirements of Burdine . . 38

C. Respondents' Evidence Was Not Adequate
Under Either Impact or TVeatment Order of 
Proof Requirements . ......................................... 41

56



vi

D. It Was Not Petitioners’ Burden to Show Lack
of Qualifications of the Respondents Nor 
Should Petitioners Be Required to Target 
Labor Sources Chosen by the Court of Ap­
peals to Maximize Minority H iring ................ 43

E. Clarification of the Order of P roof.................. 46

Conclusion..........................  .............................................48

TABLE OF CONTENTS, (continued)

Page

57



Table of Cases

American Federation of State, County & Municipal 
Employees (A.F.S.C.M.E.) AFL-CIO v. State of 
Wash., 770 F.2d 1401 (9th Cir. 1985).......................... 30

Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975). .22, 39

Allen v. Prince George's County, Md., 737 F.2d 1299
(4th Cir. 1984)..............................................................  21

Anderson v. Bessemer City, 470 U.S. 564 (1985) . . . .  21, 26

Asbestos Workers, Local 53 v. Vogler, 407 F.2d 1047
(5th Cir. 1969)..............................................................  25

Bonilla v. Oakland Scavenger Co., 697 F.2d 1297
(9th Cir. 1982)............................................................... 25

Carroll v. Sears. Roebuck & Ca,708 F.2d 183 (5th
Cir. 1983)....................................................  18, 24, 30, 31

Castaneda v. Partida, 430 U.S. 482 (1977)......................... 20

Clady v. County of Los Angeles, 770 F.2d 1421 (9th
Cir. 1985), cert, denied, 475 U.S. 1109 (1986)............  35

Clark v. Chrysler Corp.,
673 F.2d 921 (7th Cir. 1982)............................ 19, 20, 32

Connecticut v. Teal, 457 U.S. 440 (1982)....................  31, 32

Contreras v. City of Los Angeles, 656 F.2d 1267 (9th
Cir. 1981), cert, denied, 455 U.S. 1021 (1982). . . . 17, 29

Vll

TABLE OF AUTHORITIES

Page

58



Vlll

Cases, (continued) Page

Coser v. Moore, 739 F.2d 746 (2d Cir. 1984)..............  20, 45

De Medina v. Reinhardt, 686 F.2d 997 (D.C.
Cir. 1982)......................................    19

Dothard v. Rawlinson, 433 U.S. 321 (1977). .17, 31, 44, 47

E.E.O.C. v. American N a tl Bank, 652 F.2d 1176 (4th
Cir. 1981)........................................................................ 17

E.E.O.C. v. Federal Reserve Bank of Richmond, 698 
F.2d 633 (4th Cir. 1983), rev'd on other grounds 
sub nom Cooper v. Federal Reserve Bank of 
Richmond, 467 U.S. 867 (1984)......................  19, 20, 45

E.E.O.C. v. Sheet Metal Workers, Local 122, 463 F.
Supp. 388 (D. Md. 1978)................................................ 25

Fumco Constr. Corp. v. Waters, 438 U.S. 567 (1978) . . 46

Garcia v. Gloor, 618 F.2d 264 (5th Cir. 1980), cert.
denied, 449 U.S. 1113 (1981).......................................... 29

Gibson v. Local 40 Supercargoes & Checkers, etc.,
543 F.2d 1259 (9th Cir. 1976)................ ................24, 25

Green v. USX Corp., 843 F.2d 1511 (3d Cir. 1988)..........  40

Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1985)........35, 40

Griggs v. Duke Power Co.,
401 U.S. 424 (1971)....................  16, 31, 35, 42, passim

Hammon v. Barry, 826 F.2d 73 (D.C. Cir. 1987)..............  19

TABLE OF AUTHORITIES, (continued)

59



IX

Cases, (continued) Page

Hazelwood School Dist. v. United States, 433 U.S.
299 (1977)....................................  18, 19. 20, 21, passim

Hilton v. Wyman-Gordon Co., 624 F.2d 379 (1st Cir.
1980)................................................................................20

Johnson v. Transp. Agency, 480 U.S------- 108 S.Ct.
___ 94 L. Ed. 2d 615 (1987)..........................  19, 20, 46

Johnson v. Uncle Ben's, Inc., 628 F.2d 419 (5th 
Cir. 1980), vacated and remanded, 451 U.S. 902 
(1981), aff'd on remand, 657 F.2d 750 (5th Cir.),
(1981), cert denied, 459 U.S. 967 (1982)........  18, 19, 38

Kaplan v. Int'l Alliance of Theatrical & Stage
Employees, 525 F.2d 1354 (9th Cir. 1975).----------- 43

Markey v. Tmneco Oil Co., 707 F.2d 172 (5th Cir. 1983) . 20

Moore v. Hughes Helicopters, Inc., 708 F.2d 475 (9th
Cir. 1983)....................................................  18, 20, 38, 44

New York City Transit Authority v. Beazer, 440 U.S.
568 (1979)...............................................................  29, 37

Pack v. Energy Research & Dev. Admin., 566 F.2d
1111 (9th Cir. 1977)..............................................  20, 44

Personnel Administrator of Mass. v. Feeney, 442 U.S.
256 (1979)................................................  15

Pouncy v. Prudential Ins. Co. of America, 668 F.2d
795 (5th Cir. 1982)....................................  17, 30, 31, 34

TABLE OF AUTHORITIES, (continued)

60



X

Cases, (continued) Page

Presseisen v. Swarthmore College, 442 F.Supp. 593
(E.D. Pa. 1977), aff'd, 582 F.2d 1275 (3rd Cir. 1978). 27

TABLE OF AUTHORITIES, (continued)

Rivera v. City of Wichita Falls, 665 F.2d 531 (5th
Cir. 1982)..........................................................  18, 19, 20

Robinson v. Polaroid Corp., 732 F.2d 1010 (1st Cir.
1984)........................................................................ 30, 31

Scott v. Pacific Maritime Ass'n, 695 F.2d 1199 (9th
Cir. 1983)............................................................   27

Segar v. Smith, 738 F.2d 1249 (D.C. Cir. 1984), cert, 
denied sub nom Meese v. Segar, 471 U.S. 1115
(1985)......................................................................  39, 40

Shidaker v. Carlin, 782 F.2d 746 (7th Cir. 1986)................  17

Spaulding v. Univ. of Wash., 740 F.2d 686 (9th Cir.
1984)................................................................................30

Speasu v. Merchants Bank & Trust Co., 188 N.C. 524,
125 S.E. 398 (1924)......................................................  42

Ste. Marie v. Eastern R. Ass'n, 650 F.2d 395, (2d Cir.
1981)....................................................   45

Teamsters v. United States, 431 U.S. 324 (1977) . . . .  18, 26

Texas Dep't of Community Affairs v. Burdine, 450
U.S. 248 (1981)............................  35, 37, 38, 39, passim

61



Cases, (continued) ^ aSe

United States v. Jacksonville Terminal Co., 316 F.
Supp. 567 (M.D. Fla. 1970), affd in relevant part, 
rev'd & remanded on other grounds, 451 F.2d 
418 (5th Cir. 1971), cert denied 406 U.S.
906 (1972)........................................................................25

United States v. Ironworkers Local 1, 438 F.2d 679
(7th Cir.) cert, denied, 404 U.S. 830 (1971) 24

Wang v. Hoffman, 694 F.2d 1146 (9th Cir. 1982) 43, 44

Watson v. Fort Worth Bank & Trust, 487 U.S-------
108 S.Ct____101 L. Ed. 2d
827 (1988)....................................  16, 17, 22, 29, passim

Statutes
28 U.S.C. § 1254 (1)................................................................ 2

§ 1331..................................................................... 2

42 U.S.C. § 2000e, et seq., Civil Rights Act of 1964
(1981).......................................................................2-3, 14

42 U.S.C. § 1981, Civil Rights Act of 1866 (1981)........  3, 4

Textbooks
Fed. R. Evid. 301..........................................................37, 41

9 Wigmore, Evidence, § 2487 (1981) 42
§ 2491 (3d ed. 1940)........................ 41
§ 2493C (1981).................................41

xi

TABLE OF AUTHORITIES, (continued)

62



Other Authorities Page

B. Schlei & P. Grossman, Employment Discrimina­
tion Law (2d ed. 1983)................................................  24

Miscellaneous
Webster's Third New International Dictionary of the

English Language Unabridged.................................... 24

TABLE OF AUTHORITIES, (continued)

63



IN THE
Supreme Court of the United States

October Term, 1988

W a r d s  Cove P acking Company, Inc., 
Castle  & Cooke. Inc.,

Petitioners,

v.

F rank  A tonio, et al., 
Respondents.

BRIEF OF PETITIONERS

OPINIONS BELOW

On October 31,1983, the United States District Court for 
the Western District of Washington (Quackenbush, J.) entered 
an opinion following a nonjury trial. Pet. App. I. See n. 3, 
infra. An order correcting the opinion and judgment in favor 
of petitioners was entered December 6,1983. Pet. App. II. The 
District Court’s decision was published at 34 E.P.D. 1 34, 437 
(Commerce Clearing House, Inc.). The opinion of the Court 
of Appeals affirming the judgment was published at 768 F.2d 
1120. Pet. App. III. An order that withdrew the opinion and 
ordered rehearing en banc was published at 787 F.2d 462. Pet. 
App. IV. An opinion of the en banc Court of Appeals was 
published at 810 F.2d 477. Pet. App. V. An opinion of the 
panel of the Court of Appeals on remand from the en banc 
panel was published at 827 F.2d 439. Pet. App. VI. An order 
clarifying the opinion was entered on November 12,1987, Pet. 
App. VIII, and a petition for rehearing denied. Pet. App. IX.

64



2

JURISDICTION

Federal jurisdiction in the trial court was invoked under 
28 U.S.C. § 1331. The decision of the Court of Appeals sought 
to be reviewed was entered on September 2, 1987. Pet. App. 
VI. A timely petition for rehearing was filed on September 16, 
1987, Pet. App. VII, and the petition was denied on November 
12, 1987. Pet. App. IX. Jurisdiction in this Court is invoked 
under 28 U.S.C § 1254(1).

PERTINENT STATUTE

Plaintiffs’ claims arise under Title VII of Civil Rights Act 
of 1964, as amended, 42 U.S.C. § 2000e-2:

(a) It shall be an unlawful employment practice for an 
employer—

(1) to fail or refuse to hire or to discharge any indi­
vidual, or otherwise to discriminate against any individual 
with respect to his compensation, terms, conditions, or 
privileges of employment, because of such individual’s 
race, color, religion, sex, or national origin; or

(2) to limit, segregate, or classify his employees or 
applicants for employment in any way which would de­
prive or tend to deprive any individual of employment 
opportunities or otherwise adversely affect his status 
as an employee, because of such individual’s race, color, 
religion, sex, or national origin.

***

(j) Nothing contained in this subchapter shall be inter­
preted to require any employer, employment agency, labor 
organization, or joint labor-management committee sub­
ject to this subchapter to grant preferential treatment to 
any individual or to any group because of the race, color, 
religion, sex, or national origin of such individual or group

6 5



3

on account of an imbalance which may exist with respect 
to the total number or percentage of persons of any race, 
color, religion, sex, or national origin employed by any 
employer, referred or classified for employment by any 
employment agency or labor organization, admitted to 
membership or classified by any labor organization, or 
admitted to, or employed in, any apprenticeship or other 
training program, in comparison with the total number 
or percentage of persons of such race, color, religion, sex, 
or national origin in any community, State, section, or 
other area, or in the available work force in any community, 
State, section, or other area.

STATEMENT OF THE CASE

A. Nature of the Case.

The respondents in this class-action suit are former em­
ployees at several salmon canneries in Alaska. They brought 
this action against their former employers, petitioners Wards 
Cove Packing Company, Inc., and Castle & Cooke, Inc.1 charg­
ing employment discrimination on the basis of race in violation 
of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e, 
et seq., and the Civil Rights Act of 1866, 42 U.S.C. § 1981. 
The class is defined as all nonwhites who are now, will be, or 
have been at any time since March 20,1971, employed at any 
one of five canneries.* *

1 Title VII claims against a third defendant, Columbia Wards 
Fisheries, were dismissed and a judgment was entered in its favor 
on the claims brought under 42 U.S.C. § 1981. Plaintiffs’ Petition 
for Writ of Certiorari on these claims was denied April 4, 1988 
(No. 87-1388).

* These canneries are Bumble Bee at South Naknek (owned by 
Castle & Cooke, Inc); Wards Cove and Red Salmon (owned by 
Wards Cove Packing Co., Inc), and Alitak and Ekuk (operated by 
dismissed-defendant Columbia Wards Fisheries).

6 6



4

Following a lengthy non-jury trial, the trial court found 
that plaintiffs had not established discrimination under § 1981 
or Title VII and judgment was entered for petitioners. The 
Ninth Circuit affirmed this decision, but on rehearing en banc 
resolved a conflict within the circuit by determining that the 
impact analysis could be applied to subjective employment 
practices and remanded to the original panel. The subsequent 
panel decision vacated the judgment and remanded to the 
district court with directions to apply the disparate impact 
analysis in a manner inconsistent with decisions of this Court 
and in conflict with other circuits.

B. Material Facts.

The salient facts may be found in the detailed findings of 
the District Court (Pet. App. 1:1-43).* See also the Court of 
Appeals summary (Pet. App. 111:3-9). Petitioners operate sal­
mon canneries and fish camps in remote and widely separated 
areas of Alaska. Of eleven facilities, five were certified for this 
class action, and three remain in the litigation. See n. 1, 2, 
supra. The canneries operate only during the summer salmon 
run. For the remainder of the year they are vacant. Peti­
tioners' head office and support facilities are located at Seattle, 
Washington, and Astoria, Oregon.

Throughout the case period, minorities have held top posi­
tions with petitioners, including three at the superintendent

* The following abbreviations are used herein for citations to the 
record: "Pet. App." refers to the Appendices attached to the 
Petition for Writ of Certiorari; "J.A.” refers to the Joint Appendix 
filed herewith; "E.R.” refers to the Excerpt of Record lodged with 
the Court which contains certain trial exhibits that did not lend 
themselves to the Joint Appendix format; "Ex." refers to trial 
exhibits; "R.T.” refers to the Reporter’s TVanscript of the trial; 
"FF” refers to the District Court's Findings of Fact; "Dep." refers 
to a Deposition transcript offered in lieu of live testimony at trial: 
and “Tbl.” refers to a Thble in a composite statistical exhibit.

67



5

level.4 At petitioners’ Alaska facilities combined, for the period 
1971-80, minorities were nearly 24% of the new hires in the 
at-issue jobs — the jobs from which respondents claim minor­
ities were excluded.8 The five class facilities combined hired 
nearly 21% nonwhites, and each of the three remaining can­
neries hired from 10%-18% nonwhites in those jobs.6

The manner in which petitioners operate is dictated in 
large part by geography and nature: until the long Alaska 
winter is over in late April and early May, petitioners can do 
little at the canneries to prepare them for the salmon run.

Based upon the size of the predicted run for the coming 
season, management will decide whether to open a particular 
facility for canning and, if so, how many canning lines to run, 
tenders to operate, and employees to hire. Pet. App. 1:16-18. 
Three canneries (Red Salmon, CWF-Ekuk, and South Naknek) 
are located in Bristol Bay in the southeast corner of the Bering 
Sea, north of the Aleutian Islands. One cannery is located on 
Kodiak Island (CWF-Alitak) and one at Ketchikan (Wards 
Cove). Pet. App. 1:5, 6.

Of the class facilities, Red Salmon, South Naknek, and 
Wards Cove were closed for canning during certain years in 
the case period but did operate as fish camps.7 Pet. App. 1:18, 
FF 18.__________
4 R.T. 1122, 2862, 2889-90, 2439-40, 3271-72.
8 Ex. A-403, Tbl. 22 (E.R. 13).
6 I d  Tbl. 23, 1, 2, 5 (E.R. 14, 10-12).
7 A fish camp is a support base for tenders and fishing vessels. It 

performs no processing. Many at-issue jobs are held at the fish 
camps: eg., beachgang, carpenters, cooks, tender crews, and fish­
ermen. There is no racial stratification between cannery and at- 
issue jobs because no canning is performed and Local 37 has no 
contract for any jobs. Accordingly, respondents did not name the 
fish camps at Egegik, Moser Bay, Craig, and Chignik as class 
facilities.

68



6

Each facility is a self-supporting installation where the 
employees are housed and fed by the company. The canneries 
must rely almost entirely upon their own on-site employees 
to maintain and repair the cannery buildings and equipment.

There are two general categories of jobs: cannery worker 
and laborer jobs which are not at issue, and non-cannery jobs 
which are at issue. (Pet. App. 1:28, FF 82.) The non-cannery 
workers are hired during the winter and early spring and sent 
to Alaska during the preseason. The cannery workers are not 
needed until several weeks later when the salmon run actually 
commences.

The non-cannery workers include such jobs as carpenters, 
machinists, tender crews, and a beachgang. During the pre­
season, these personnel drive piling, launch boats, get the 
machinery running, and repair and de-winterize the cannery. 
They are housed in bunkhouses insulated and heated for the 
cold spring weather. The preseason is an intense period be­
tween the winter “break up” and the commencement of the 
salmon run, and there is no time to train unskilled workers 
for skilled jobs. (Pet. App. 1:18-19). The non-cannery workers 
are hired laterally from an external labor market. (Pet. App. 
1:39, FF 112.) This hiring is done during the first three months 
of the year and requires availability by the end of April. (Pet. 
App. 1:30, FF 86.)

In June, after the canneries are in operating condition, 
nonresident (outside Alaska) cannery workers are mostly hired 
through Local 37, I.L.W.U.8 They are transported to Alaska 
in time to process the first of the salmon caught. Resident 
cannery workers were hired in the early case period from the

8 Local 37 did not have a contract with Ekuk. Pet. App. 1:32. 
FF 91.

69



7

areas near the canneries.9 When cannery workers arrive, addi­
tional bunkhouses are opened for them. These are bunkhouses 
suited for the summer weather. (Pet. App. 1:83-84, FF 149 
A-B.) During the season, which will last from three weeks to 
two months, most employees, both cannery and non-cannery 
workers, have season guarantees in their union contracts, a fact 
which inhibits transfers across union lines during the season 
because it would require the employer to pay an additional 
guarantee. (Pet. App. 1:39, FF 111.) There are few midseason 
vacancies, and transfers across departmental lines during the 
season, and even between seasons, is rare. (Pet. App. 1:34, 
FF 98, 99.)

As soon as the salmon run ends, the cannery workers are 
discharged and sent home; the non-cannery workers then haul 
out the boats, sail the tenders south,10 11 and winterize the 
canneries.

The union contracts for the carpenters, machinists, ten­
der crews, culinary crews, and also for Local 37 have rehire 
preference clauses which operate like a seniority system. (Pet. 
App. 1:29-31, 35, FF 85, 87,101.) This rehire preference clause 
obligates the petitioners to rehire satisfactory employees in 
the same job for the upcoming season and this accounts for 
nearly one-half the hires for at-issue jobs."
9 As pointed out by the District Court, the racial composition of 

the hires of Alaska residents is largely dictated by geography (Pet. 
App. 1:38). Some canneries hired few local residents as cannery 
workers — Red Salmon hired only 18 nonwhite Alaska residents 
out of 338 total new hires during the entire case period. Ex. 497, 
Tbl. 3(b) (E.R. 16).

10 Some tenders move on to other canneries for later seasons. 
Respondents’ comparative statistics sought to count each tender 
crew member on arrival at each cannery as a new hire. The 
District Court was not impressed with this. Pet. App. 1:120.

11 See Ex. A-320 (a), p.5, col. "same dept., same job” (40% of at issue 
jobs filled by rehires).

70



8

Although the labor market for cannery worker and laborer 
jobs is 90% white, nonwhites are greatly overrepresented in 
these jobs because of institutional factors: Local 37 is domi­
nated by Filipinos as are the crews it dispatches, and the 
geographic areas surrounding most of the canneries are heavily 
nonwhite. (Pet. App. 1:36-39, FF 105-109.)'2

Petitioners, however, do not exclude members of Local 37 
or Alaska Natives from consideration for at-issue jobs.11 * 13 For 
instance, respondent Atonio originally obtained a cannery 
worker job through Local 37; later, after two untimely oral 
requests for other jobs, he made a timely application and was 
hired first in the beachgang and later as a tender deckhand. 
He was rehired for a job as deckhand in 1981, but quit before 
the boat departed. (Pet. App. 1:87-88, FF 159.)

For at-issue jobs, petitioners obtain many more applica­
tions than there are vacancies. (Pet. App. 1:31, FF 89). Conse­
quently, petitioners do not advertise; however, the Alaska

11 As the District Court found (Pet. App. 1:38) and the panel also 
recognized (Pet. App. 111:32), Alaska Natives comprise a high 
percentage of local labor market for resident cannery workers at 
the remote canneries. For example, at Ekuk, the most remote 
cannery, of the Alaska residents hired as cannery workers. 97% 
were nonwhite. Ex. A-497, Tbl. 3(b) (E.R. 16); at the same cannery 
for Alaska residents hired for at-issue jobs, 91.6% were nonwhite 
(J.A. 290-91; Ex. A-501, Tbl. 2(A)).

By comparison, at Wards Cove, located near the city of 
Ketchikan, the majority of Alaska residents hired for cannery 
worker jobs were white (402 out of 471 openings). Ex. A-497. 
Tbl. 3(b) (E.R. 16).

13 Of Alaska residents hired, Alaska Natives filled 60% of the at-issue 
jobs and 60% of the cannery worker and laborer jobs overall. 
Ex. A-501, Tbl. 1(A); J.A. 290-91, 1 45. The hiring of nonwhites 
for at-issue jobs is far in excess of their availability in the labor 
market for Alaska. Nonwhites only comprise 15.6% of that 
market. J.A. 290-91; Ex. A-501, Tbl. 18.

71



9

Unemployment Service has been called. (Pet. App. 1:28-29, 
FF 83.) Petitioners do accept walk-in applicants and referrals 
from unions. R.T. 2769, 2771; Dep. Lessley, p. 7, J.A. 15. 
There is not time to post openings during the season because 
the job needs to be filled immediately and management can­
not wait for an interview structure. R.T. 1135; 2772; Pet. 
App. 1:34, FF 96.

In 1974 respondents commenced a class action against 
petitioners. The suit mounted a broad-scale attack against the 
gamut of petitioners’ employment practices. Respondents 
identified 16 “practices”14 which they contended caused an 
imbalance and thus a "concentration” of nonwhites in the 
lower-paying cannery worker jobs. Respondents used com­
parative statistics to argue that of the total work force, the 
majority of the nonwhites were concentrated in the lower- 
paying jobs and that there should have been a balance of 50% 
white/nonwhite employees in all job classifications.

After 12 trial days, in which more than 100 witnesses 
testified, over 900 exhibits were admitted, and over 1,000 
statistical tables were submitted, the trial court entered 
extensive findings of fact in a 73-page opinion. Pet. App. I.15 * * 
The findings determined that respondents’ comparative statis­
tics were of little probative value; that the labor supply for peti­
tioners’ facilities is approximately 90% white; that minorities 
were not underrepresented in the at-issue jobs; that cannery 
workers are not the appropriate comparison labor pool for
14 In the Revised Pretrial Order, plaintiffs listed word-of-mouth 

recruitment, separate hiring channels, nepotism, termination of 
Alaska Natives, rehire preference, retaliatory terminations, menial 
work assignments, fraternization restrictions, housing, messing, 
English language requirement, race labeling, subjective hiring 
criteria, lack of formal promotion practices, failure to post jobs, 
and discrimination in pay in certain jobs.

15 Although respondents contended there was discrimination against 
Alaska Natives, not a single Alaska Native testified in plaintiffs’
case. Only petitioners called Alaska Natives to the witness stand.
E.g., J.A. 414.

72



10

at-issue jobs; that petitioners hire from an external labor sup­
ply and do not either promote-from-within or train inexperi­
enced, unskilled workers for at-issue jobs: that the jobs are not 
fungible and most jobs at issue require skill and prior experi­
ence that is not readily acquirable at the canneries; that Local 
37 provides an oversupply of nonwhite cannery workers and that 
this overrepresentation is an institutional factor in the industry.

In addition, the trial court found that no individual in­
stances of discrimination were proven: that petitioners did not 
give job preference to friends and relatives; that respondents’ 
"nepotism” statistics were distorted and unreliable; that hiring 
was "on the basis of job-related criteria”;18 that giving experi­
enced personnel a preference in hiring was a business necessity; 
that the rehire preference clauses in the union contracts oper­
ated like a seniority system; that housing is not racially segre­
gated, that housing and rehire policies were dictated by busi­
ness necessity; and that Local 37 was responsible for messing.

The trial court found that respondents had failed to estab­
lish intentional discrimination and the disparate impact anal­
ysis was not appropriate for application to respondents’ wide- 
ranging multiple practice challenge nor to subjective hiring 
practices. In applying the impact analysis individually to five 
of petitioners’ practices (rehire preference, English language, 
"nepotism,” housing, and messing), the District Court again 
found in favor of petitioners. (Pet. App. 1:102-107, 124-129.)

The court found that petitioners had not discriminated 
on the basis of race and entered judgment in their favor. (Pet. 
App. 1:130.)
18 This finding was supported by substantial evidence. E.g., peti­

tioners’ skill expert DeFrance analyzed the skills necessary for 
several job classifications. He then compared a survey of incum­
bents and found that of 139 persons for which adequate informa­
tion was available, 131 did possess the requisite skills and 8 were 
not qualified. R.T. 2988. This finding was also uniformly sup­
ported by testimony of management, supervisors, and incumbents, 
who testified to the need to hire persons with prior skill and 
experience in the at-issue jobs. E g .  J.A. 161; 439-443; 596-607.

73



11

C. Court of Appeals Rulings.

On appeal a panel of the Ninth Circuit affirmed the 
judgment. Pet. App. 111:56.

The Court of Appeals recognized that respondents had 
failed in their labor market proof, that respondents’ compara­
tive statistics were of little probative value, that nonwhites 
were overrepresented in cannery worker jobs, and that institu­
tional factors distorted the racial composition of the work force. 
Pet. App. 111:20-36. The petitioners’ labor market statistics 
and findings thereon by the District Court were affirmed,17 as 
were the findings that respondents had not given friends or 
relatives a preference in hiring and that petitioners hired 
according to job-related criteria. The panel concluded:

The [district] court stated, “regardless of the manner in 
which a prospective employee came to the attention of the 
hiring personnel, the person was evaluated according to 
job-related criteria.’’ Thereafter, in concluding the case, 
the [district] court encompassed all of the claims when it 
said “defendants did not discriminate in the hiring, firing, 
promoting, or p a y i n g T h e  decision of the District 
Court will not be disturbed.’’

Pet. App. III:39-40.18

This holding would seem to have disposed of respondents’ 
claims regardless of the analytical theory on which presented. 
The panel noted, however, a conflict in decisions of several 
circuits and within the Ninth Circuit itself as to whether 
the disparate impact analysis could be applied to subjective

17 None of the findings of fact by the District Court were overturned 
as clearly erroneous.

u The court also held “the ultimate fact, that there existed no 
pattern or practice of discrimination in hiring, promoting, paying, 
and firing, is supported by the numerous subsidiary findings of 
the District Court.” Pet. App. 111:38.

74



12

practices. Pet. App. III-46-55. A petition for rehearing en banc 
was granted, Pet. App. IV, and the en banc court subsequently 
held that the impact analysis could be applied to subjective 
employment practices. The case was returned to the original 
panel. Pet. App. V:39.

On remand the Court of Appeals panel affirmed the Dis­
trict Court on rehire preferences, did not discuss the English 
language requirement, but held that plaintiffs’ “comparative 
statistics,’’ which showed only a concentration of minorities 
in the cannery worker jobs, were nonetheless adequate to force 
petitioners to prove their hiring practices were justified on 
grounds of business necessity. In doing so, the Court of 
Appeals did not hold that any practice caused disparate 
impact, and ignored the District Court’s findings that re­
spondents’ statistics were distorted and unreliable, that peti­
tioners hired more nonwhites than the proportion available in 
the labor supply, and that institutional factors, not the peti­
tioners’ practices, caused an overrepresentation of minorities 
in cannery worker jobs. Pet. App. VI.

The court also held, contrary to trial court findings, that 
a preference for relatives ("nepotism’’) existed and had an 
adverse impact on nonwhites. Finally, the court questioned 
the District Court’s findings of business necessity for peti­
tioners’ housing and messing practices, but did not hold them 
to be clearly erroneous. The Court of Appeals selected seven 
of the 16 practices complained of by respondents to be exam­
ined under a business necessity standard. (Word-of-mouth 
recruitment, nepotism, subjective criteria, separate hiring 
channels, labeling, housing, and messing.) The Court of 
Appeals vacated judgment for petitioners and remanded. 
Pet. App. VI.

75



13

SUMMARY OF ARGUMENT

Although they mounted a broad scale attack on behalf of 
over 2,000 class members, respondents were unable to prove 
any instance of individual or class-wide disparate treatment 
of nonwhite employees in any aspect of the employment rela­
tionship. Respondents’ fall back position was to allege under 
the disparate impact theory that their same marginal evidence 
proved petitioners’ practices combined to cause unintentional 
discrimination.

Respondents’ impact case was centered on comparative 
statistics showing internal work force comparisons. The Court 
of Appeals held that these statistics were sufficient to raise 
an “inference” of discrimination under the disparate impact 
model. The Court of Appeals fashioned a new allocation of 
the order of proof. This order of proof erroneously establishes 
a much lower threshold for a plaintiff in Title VII litigation 
than has been developed under decisions of this Court and 
the courts of appeal. It deprives the employer of the usual 
defenses, e.g., that the plaintiff’s statistics are flawed, that 
the relevant labor market shows minorities are not under­
represented in at-issue jobs, and that the inferences urged by 
plaintiff are less probative than those urged by the employer.

The District Court properly considered the structure and 
practices of respondents’ business and in a carefully reasoned 
opinion found that the imbalance was nothing more than the 
result of institutional factors which produce an overrepresenta­
tion of minorities in cannery worker jobs.

The District Court properly rejected respondents’ statis­
tics in favor of petitioners’ labor market analysis that showed 
that class members were not underrepresented in the jobs at 
issua The District Court also found that respondents had 
failed to prove a discriminatory preference for relatives existed 
and rejected respondents’ statistical evidence on that issue as 
flawed. In rehabilitating respondents’ case under the impact

76



14

theory, the Ninth Circuit ignored the foregoing findings, as well 
as a long line of decisions of this Court and the circuit courts 
that supported the District Court’s action.

The Ninth Circuit improperly allowed respondents to 
extend the reach of the disparate impact analysis to challenge 
the cumulative effect of a wide range of practices respondents 
chose to name. Respondents have the burden of proving the 
causal connection between any challenged practice and the 
alleged disparity, but the Court of Appeals decision effectively 
dispenses with that requirement. Combined with its accep­
tance of respondents’ evidence of racial imbalance in job cate­
gories, the Ninth Circuit has forced the employer to shoulder 
the burden of justifying each practice the respondents choose 
to name based on a mere showing that the employers’ work 
force is not racially balanced. This is at odds with the Con­
gressional purpose stated in Title VII, 42 U.S.C. § 2000e-2(j), 
and all but compels employers to engage in quota hiring and 
other activities that reduce job opportunities for minorities.

Any doubt that the Ninth Circuit was revolutionizing the 
allocation of burdens of proof was removed when it held, 
without relevant authority, that any attempt by an employer 
to explain or justify his practices in response to respondents’ 
disparate treatment claim, precluded the employer from chal­
lenging respondents’ impact case; by ruling that employers 
must prove the business necessity of job qualifications with­
out requiring respondents to prove the qualifications had a 
disparate impact; and requiring respondents to justify why 
they did not use certain labor sources that the Court of 
Appeals apparently decided might result in increased minority 
hiring — in the face of the fact that minorities were not under­
represented in the jobs at issue. This misallocation of the 
burdens of proof conflicts with decisions of this Court and 
the circuit courts and should be rejected.

77



15

ARGUMENT

I. Respondents Failed to Prove Discrimination Under the 
Disparate Impact Theory

The linchpin of respondents’ case is the undisputed fact 
that nonwhites are overrepresented in the cannery worker 
category. That one fact, reflected in various forms in respon­
dents’ statistics, is the foundation of their claim that nonwhites 
are disproportionately excluded from at-issue jobs and are ra­
cially segregated in housing and messing. That overrepresen­
tation is caused by institutional factors in the industry and is 
without legal significance Because respondents’ other evidence 
failed to establish that nonwhites were excluded from the at- 
issue jobs by any identified practice their impact case must fail 
— particularly since petitioners met their burden of justifying 
many of the practices respondents challenged in the aggregate

A. Statistical Evidence That Shows Only a Concentration 
of Minority Employees in Jobs Not at Issue Fails As 
a Matter of Law to Establish Disparate Impact of 
Hiring Practices Where the Employer Fills the At- 
Issue Jobs From Outside His Own Work Force, Does 
Not Promote From Within or Provide Training for 
Such Jobs, and Where Minority Employees Are Not 
Underrepresented in the At-issue Jobs According to 
a Labor Market Analysis Accepted by the Trial Court.

The trial court found against respondents on the treat­
ment theory, i.e., the petitioners did not intentionally dis­
criminate against the class or any individual class member in 
the adoption of or application of any of the employment 
practices challenged by respondents here.19 This decision was 
affirmed on appeal.
19 Thus, for instance, plaintiffs alleged, but did not prove that 

employers hired nonresident cannery workers through Local 37, 
ILWU ‘‘because of, rather than in spite of” the predominantly 
Filipino composition of that union, Personnel Administrator of 
Mass. v. Feeney, 442 U.S. 256, 279 (1979), and they failed to 
establish that any similarly situated employees were treated 
differently on the basis of race under any practice challenged.

78



16

Nonetheless, certain practices that are fair in form and are 
equally applied may have a "disparate impact" on Title VII 
protected class. That is, they "may in operation be functionally 
equivalent to intentional discrimination.” Watson v. Ft. Worth
Bank & Ihist, 487 U.S.___ , 101 L.Ed. 2d 827,840 (1988). This
“disparate impact" analysis was first adopted in Griggs v. 
Duke Power Co., 401 U.S. 424 (1971). Under it, the plaintiff 
can establish a prima facie impact case if the evidence estab­
lishes that an employer’s practice causes a "substantial dis­
parate impact," i.e., that the practice has the effect of dispro­
portionately denying job opportunities on the basis of race. 
Id. Failure to justify a practice in the face of such evidence 
will subject the employer to liability. Id.

1. Petitioners' Labor Market Analysis Was More 
Probative Than Respondents' Comparative Statistics, Refuted 
Any Showing of Disparate Impact, and Should Not Have 
Been Ignored by the Ninth Circuit.

The parties offered starkly contrasting statistical evidence 
on the issue of whether a disparate impact in hiring existed. 
Respondents argued that the petitioners’ actual hiring results 
in the at-issue jobs should be compared "internally," i.e., com­
pared to the racial composition of the cannery worker jobs.20 21 
Because this showed a "stratified” work force, i.e., nonwhites 
were concentrated in the cannery worker jobs, respondents 
contend impact has been proven.2'
20 A variant of this same theme was to make an internal comparison 

between the petitioners’ hiring results for its entire work force 
(at-issue and not at-issue jobs combined) with the hiring results 
in the at-issue jobs.

21 In essense, plaintiffs are arguing either that the cannery workers 
are the available labor supply or the racial composition of the 
cannery workers is a reasonable proxy for the available labor 
supply for the at-issue jobs. The trial court found against them 
on both points: the cannery workers did not form the labor supply 
for the at-issue jobs, the company does not promote-from-within 
in any jobs, and the race of cannery workers is not representative 
of the relevant labor supply.

79



17

Petitioners’ countervailing evidence compared their hiring 
results in the at-issue jobs with the racial composition of an 
"external” labor market for the jobs at issue. This evidence 
showed that nonwhites were overrepresented in the cannery 
worker category and not significantly underrepresented in 
the at-issue jobs.22 23 Although the trial court explicitly found 
petitioners’ statistical evidence more probative, the Ninth Cir­
cuit credited respondents’ statistics as raising an "inference” 
of disparate impact. In so doing, the Ninth Circuit also ignored 
decisions from this Court and the circuit courts that compelled 
a finding that disparate impact was not proven.

In determining whether the evidence established disparate 
impact, the District Court properly considered the evidence 
and arguments of both parties. The Ninth Circuit did not. 
This was serious error. Watson, supra, 101 L.Ed. 2d at 846 
(plurality), citing Dothard v. Rawlinson, 433 U.S. 321, 331 
(1977), and id. at 338-39 (Rehnquist, J., concurring in result 
and concurring in part) (must examine plaintiffs’ evidence 
of impact in light of the facts, defendants’ attack on that 
evidence, and defendants’ own evidence).2*
22 Ex. A-278, Tbl. 4 (each facility) (E.R.2-7). Nonwhites were over­

represented in the at-issue jobs combined whether considered on 
a cannery by cannery basis or combination of facilities. See table 
at J.A. 279; Ex. A-278, Tbl. 4, col. "At Issue” row "CMPS DEV” 
for each facility (E.R.2-7). It was only when the statistics were 
disaggregated and analyzed job family by job family on an in­
dividual cannery basis that any evidence of underrepresentation 
of nonwhites surfaced; however, even then, there were only three 
instances out of a possible 65 at the five class canneries where the 
underrepresentation was significant and in each of those three 
instances, the underrepresentation was less than three standard 
deviations. J.A. 280, summarizing J.A. 266-78.

23 Accord Shidaker v. Carlin, 782 F.2d 746, 750 (7th Cir. 1986); EEOC  
v. American Nat'l Bank, 652 F.2d 1176, 1189 (4th Cir. 1981); 
Contreras v. City o f Los Angeles, 656 F.2d 1267, 1273 (9th Cir. 
1981), cert, denied, 455 U.S. 1021 (1982); Pouncy v. Prudential Ins. 
Co., 668 F.2d 795. 801, n. 8 (5th Cir. 1982).

80



18

In analyzing such evidence, this Court cautioned ten years 
ago that statistics come in an "infinite variety” and their 
usefulness "depends on all of the surrounding facts and cir­
cumstances” 7Teamsters v. United States, 431 U.S. 324, 340 
and n. 20 (1977).24 Here, the key factual findings (supra, pp. 
9-10) plainly undermined whatever probative value respon­
dents’ imbalance evidence had and just as plainly supported 
the petitioners’ labor market analysis.

Where the plaintiff, as here, alleges that the employer's 
recruiting practices and hiring criteria have caused a dispro­
portionate exclusion of a Title VII protected class from certain 
jobs, identifying the relevant labor market for those jobs and 
determining its racial composition is "usually the starting 
point for impact analysis.” Moore v. Hughes Helicopters, Inc., 
708 F.2d 475, 482 (9th Cir. 1983).“ * &

u  Failure to heed this simple, but crucial admonition has often been 
fatal to a party’s statistical case. E.g., Carroll v. Sears, Roebuck
& Co., 708 F.2d 183, 191 (5th Cir. 1983) (plaintiffs’ applicant 
flow statistics disapproved because they fail to account for re­
cruiting efforts that resulted in artificially high number of black 
applicants); Johnson v. Uncle Ben's, Inc., 628 F.2d 419, 426 (5th 
Cir. 1980) (in promotion case, employer's "statistics comparing 
Uncle Ben’s work force to the external labor market are irrele­
vant”), vacated and remanded 451 U.S. 902 (1981), aff'd on 
remand, 657 F.2d 750 (5th Cir.), cert, denied. 459 U.S. 967 (1982).

“ This “often-decisive. . . labor pool definition” requires findings 
as to the source from which the employer normally fills such jobs 
and the qualifications of potential applicants for such positions. 
Rivera v. City o f Wichita Falls, 665 F.2d 531. 540 (5th Cir. 1982), 
citing Hazelwood School Dist. v. United States, 433 U.S. 299, 
308-312 (1977).

81



19

Because petitioners’ vacant positions are “filled by lateral 
hires’’ from outside their work force, then the “external labor 
market” is the relevant one. Rivera v. City of Wichita Falls, 
665 F.2d 531, 540-545.* 28 Using accepted methodology, peti­
tioners’ experts determined the proper geographical boundaries 
of and the racial composition of the persons in that market 
most likely to possess the qualifications for the jobs at issue.27 
Hazelwood School Dist. v. United States, 433 U.S. 299, 308; 
Rivera, supra; EEOC v. Federal Reserve Bank of Richmond, 
698 F.2d 633, 658-62 (4th Cir. 1983), rev'd on other grounds 
sub nom Cooper v. Federal Reserve Bank of Richmond, 467 
U.S. 867 (1984); De Medina v. Reinhardt, 686 F.2d 997, 1004- 
1009 (D.C. Cir. 1982); Clark v. Chrysler Corp., 673 F.2d 921, 
927-929 (7th Cir. 1982).28 They then compared petitioners’
28 On the other hand, the employer s existing work force or “internal 

labor pool” is most appropriate where the employer fills the jobs 
at issue from lower level positions by promotion-from-within. 
Rivera, supra, 665 F.2d at 540-41; Uncle Ben's, supra, 628 F.2d at 
425-426.

Indeed, plaintiffs themselves, in recognition of these facts, 
offered their own external labor market analysis, but it was re­
jected by the trial court. Plaintiffs do not challenge that finding 
here.  ̂Although plaintiffs initially alleged there was promotion 
discrimination, they offered no statistics purporting to show 
promotion bias.

Petitioners labor market theory and hiring analysis was explained 
by their expert labor economist, Dr. Albert Rees, J.A. 250-303; 
the statistical theory was explained by expert statistician Dr. 
Donald Wise, R.T. 1688-1726 (see excerpt discussing Ex. A-278 
at J.A. 237-246); and the terms in the statistical tables (e.g., 
Ex. A-278) and other foundational material are explained by Dr. 
William Price, an expert computer programmer, R.T. 1553-1662 
(index of terms set forth at R.T. 1674-77).

28 This approach is also used to assess the validity of the voluntary 
adoption of affirmative action plans where such adoption is chal­
lenged in "reverse" discrimination cases. See Johnson v. TYansp 
Agency, 480 U.S.__, 94 L.Ed. 2d 615, 631 (1987) (citing Hazelwood 
with approval); Hammon v. Barry, 826 F.2d 73 (D.C. Cir. 1987) (plan 
disallowed where minority employees not underrepresented in the 
jobs at issue in comparison to the area labor market).

82



20

actual hiring results in filling vacancies over the relevant time 
span using the "standard deviation” analysis approved by 
this Court in Castaneda v. Partida, 430 U.S. 482 (1977) (jury 
selection case), and in the employment discrimination context 
in Hazelwood, supra, 433 U.S. at 311, n. 17; Rivera, supra 665
F.2d at 536, n. 7. The District Court accepted this evidence.

At least four post -Hazelwood circuit court decisions hold 
that comparative statistics like those offered by plaintiffs will 
be refuted by credible external labor market evidence that shows 
no underrepresentation of minorities in the jobs at issue.29 
Hilton v. Wyman-Gordon Co., 624 F.2d 379, 380 (1st Cir. 1980); 
Coser v. Moore, 739 F.2d 746, 752 (2d Cir. 1984); Federal Reserve 
Bank of Richmond, supra 698 F.2d at 658-62 (4th Cir.); Rivera 
supra 665 F.2d at 539, 544-45 (5th Cir.). See Clark, supra 673 
F.2d at 929 (7th Cir.) (external labor market data relied on to 
show no disparate impact in hiring).30 If nonwhites are not 
underrepresented in at-issue jobs, it can hardly be said that they 
have established a prima facie case of disparate impact against 
them in those jobs. See Johnson v. 7Yansp. Agency, supra n. 
28, 94 L.Ed. 2d at 631, n. 10.

The District Court’s determination as to the racial composi­
tion of the relevant labor market was undoubtedly factual and 
reviewable only under the clearly erroneous standard,31 as was
“ Moreover, even if plaintiff’s objections to Dr. Rees’ analysis were 

accepted and general labor force figures were not adjusted for 
qualifications or availability, seasonal nonwhites are still over­
represented in the at-issue jobs combined and there is no change 
in the findings as to the few instances of underrepresentation. 
Ex. A-278, Thble 5 (for each cannery or combination of facilities)

30rI\vo Ninth Circuit cases rejected contentions of discrimination 
where plaintiffs relied on comparative statistics without a showing 
of relative qualifications. Moore v. Hughes Helicopters, Inc. 708 
F.2d 475 (9th Cir. 1983) (impact); Pack v. Energy Research & Dev. 
Admin., 566 F.2d 1111 (9th Cir. 1977).

31 Markey v. Tenneco Oil Co.. 707 F.2d 172. 174 (5th Cir. 1983); Clark 
v. Chrysler Corp., supra, 673 F.2d at 928.

83



21

its determination of the probative weight of the parties’ 
statistics.32. Allen v. Prince George's County, Md., 737 F.2d 
1299,1303 (4th Cir. 1984). Like the circuit court in Hazelwood, 
the Ninth Circuit “substituted its judgment for that of Dis­
trict Court” in accepting respondents’ proof and ignoring 
petitioners’ evidence that told “a totally different story.” 
Hazelwood, supra, 433 U.S. at 308-10. This fact finding was 
error. Anderson v. Bessemer City, 470 U.S. 564 (1985).

2. Allowing Proof of Racial Imbalance to Establish 
Disparate Impact is Inimical to the Desirable Purposes of Title 
VII, and Provides an Unreasonable and Unworkable Standard 
in Practice. ^

Most responsible employers attempt to utilize employ­
ment practices that provide equal opportunity for women 
and minorities. Because respondents’ theory simplistically 
assumes that the highest nonwhite percentage in any job 
category (or in the overall work force) is the standard against 
which an employer’s hiring in all categories will be measured, 
even responsible employers would at least consider covert 
policies that could reduce job opportunities for protected 
classes e.g., establishing self-imposed ceilings on the hiring of 
women and minorities, both at the hiring stage and through 
layoffs that bring the work force into “balance.”

Respondents’ theory discourages affirmative action pro­
grams because successful recruitment of a large percentage 
of minorities in one category will be penalized where it is not 
achieved in all other categories — even if women and minorities 
are already proportionately represented in relation to the 
labor market.

Finally, respondents’ theory imposes an unworkable stan­
dard on employers. Because the focus is on the racial balance 
of persons hired, rather than on the percentage of persons who
32 The Ninth Circuit recognized this latter rule on appeal. Pet. App.

VI:14, but did not adhere to it.

84



22

are available, the employer never knows to what standard he 
will be held until after hiring is completed. Since there is no 
solace for the employer in achieving the available labor supply 
percentage (or even matching the racial composition of his 
applicant flow), the employer never knows in day-to-day prac­
tice what the standard will be or how to meet it — unless a 
self-imposed strict one:one racial hiring ratio is set up — that 
is, quota hiring. This is directly at odds with the purpose of 
the statute. Watson, supra, 101 L.Ed. 2d at 843-44 (plurality); 
A lbem arle Paper Co. v. Moody, 422 U.S. 405, 449 (1975) 
(Blackmun, J., concurring in judgment).

The Ninth Circuit’s application of respondents’ imbalance 
theory demonstrates the unfairness and the absence of common 
sense of the theory. These petitioners hired a very large percent­
age of nonwhites in the cannery worker category principally 
because they adhered to their obligations under a union con­
tract. The Ninth Circuit then demanded that petitioners justify 
why that level has not been reached in all of their other (at issue) 
jobs, regardless of the fact that nonwhites are nowhere near 
that portion of the available labor supply for those jobs.

3. R esp o n d en ts ' “Separate H iring  C hannels" A rg u ­
m en t is a R ed  Herring: Where R esponden ts H ave Failed to 
Show T hat the P etitioners' Practices for F illing Jobs N o t at 
Issue  E ith e r  In ten tiona lly  or in E ffe c t E xcluded  N onw hites  
from Jo b s  A t  Issue, Those Practices A re  Irrelevant.

The principal cause of nonwhite overrepresentation in 
cannery worker jobs was the dispatching practices of Local 37. 
(Pet. App. 1:110; 1:35-37, FF 103,105-108.) Petitioners did not 
intentionally hire cannery workers through Local 37 because 
of the race of the union members or the racial composition of 
the crews it dispatched. See  Pet. App. 1:33, FF 93; 1:119 (no 
discriminatory animus). More importantly, respondents failed 
to establish that hiring cannery workers through Local 37 had 
the effect of disproportionately excluding nonwhites from the 
at-issue jobs.35 There is no evidence showing that nonwhites, 33

85

33 Tb the contrary, the court found that nonwhites were not signifi­
cantly under utilized in those at-issue jobs, that nonwhites were

(footnote continued on next page)



23

once they were hired as cannery workers, were "locked in" or 
were precluded from applying for or being considered for the 
at-issue jobs. It is the cannery workers themselves who 
decided to become cannery workers by going to Local 37 in 
the first place; many other nonwhites initially sought work 
with the companies, instead of Local 37, and were hired.34 35 * * 
Thus, not only is the racial composition of the cannery worker 
crew legally insignificant, the practices used to fill those jobs 
are irrelevant. See Pet. App. 1:105, n. 1 (union nepotism in 
filling cannery worker jobs has “little, if any, bearing upon 
at-issue jobs”)

That these petitioners looked to Local 37, a narrow slice 
of the general labor market, for nonresident cannery workers 
is reasonable. Local 37 was the exclusive bargaining repre­
sentative for the jobs and therefore dispatched such workers 
to the canneries.

The fact that this "source" produced an overrepresentation 
of nonwhites in the cannery worker jobs does not alter the fact 
that nonwhites are not underrepresented in the jobs from which 
they claim exclusion. But for the historical anomaly that this 
union is run by and dispatches primarily Filipinos, respondents 
would not even have a case: if the racial composition of the 
crews Local 37 dispatched had matched the 10% nonwhite 
labor market, even respondents’ "imbalance" theory would 
fail.38 Respondents should not be allowed to use this institu­

not "deterred” from applying for at-issue jobs, that whites and 
non-whites alike were free to apply for at-issue jobs and that 
similarly situated applicants were treated equally. (Pet. App. 1:42, 
43, FF 123, 94.)

34 Ex. A-403, Tbl. 22 (790 nonwhite new seasonal hires in at-issue jobs 
in petitioners' Alaska facilities combined), Tbl. 23 (433 nonwhite 
new seasonal hires in at-issue jobs at the five class canneries 
combined) (E.R. 13, 14).

35 The effect of Local 37 on nonresident (of Alaska) hiring is graph­
ically demonstrated in Ex. A-499, Tbl. 1 (E.R. 19): Examining 
hiring of nonresident employees shows that with Local 37 mem­
bers excluded, whites held 90.2% of the cannery worker and laborer
jobs and 91.6% of the at-issue jobs at the five class facilities

(footnote continued on next page)
86



24

tional distortion of the labor market to establish an artificially 
high standard for nonwhite employment. See Carroll supra, 
708 F.2d at 191 (artificially high minority applicant flow not 
allowed as comparison standard). It is as irrelevant a standard 
as the race of students was in a case involving discrimination 
against black teachers. Hazelwood, supra, 433 U.S. at 308.36

B. Respondents Failed to Prove a Practice of "Nepotism" 
Existed, Their Statistics Purporting to Show Its Exis­
tence Were Properly Rejected, and The Ninth Circuit 
Committed Error in Finding Otherwise.

The only alleged hiring practice that respondents 
attempted to offer separate proof of causation or impact was 
the hiring of relatives. Petitioners concede that relatives were 
hired in some jobs, but deny that respondents ever established 
that a practice of "nepotism” existed. Nepotism is defined as 
"favoritism shown to. . . relatives as by giving them positions 
because of their relationship rather than on their merits.” 
Webster's Third New International Dictionary of the English 
Language Unabridged, p. 518. It has also been defined as the 
"use of family relationship qualifications for employment. . . 
opportunities.” B. Schlei and P. Grossman, Employment 
Discrimination Law (2d ed. 1983), p. 573.37

combined. See also Tfestimony of Dr. Rees, J.A. 294-299; Ex. 
A-498, Tbl. 4 (E.R. 17). At Ekuk, which had no Local 37 contract, 
not a single Filipino was employed as a cannery worker during the 
case period. R.T. 2892.

38 A more pertinent application of the impact model to the cannery 
worker hiring practices would be to compare the labor market for 
cannery workers (less than 1% Filipino; 10% nonwhite, E.R. 8-9) 
with the actual percentage hired (20-50%, E.R. 10-14). If the use 
of this practice has a disparate impact at all, it is on whites seeking 
cannery worker jobs — not on nonwhites seeking at-issue jobs.

37 See Gibson v. Local 40, 543 F.2d 1259, 1268 (9th Cir. 1976) 
("preference” given to relatives); United States v. Ironworkers 
Local 1, 438 F.2d 679, 683 (7th Cir.) cert, denied, 404 U.S. 830 (1971) 
(discovery into "nepotistic practices” allowed because giving 
"preference to relatives” of union members can violate Title VII).

87



25

In the cases where the plaintiff has prevailed in chal­
lenging nepotism, the issue was not whether the practice 
existed, but whether the practice had a disparate impact on the 
class or was justified. See, e.g., Bonilla  v. O akland Scavenger 
Ca, 697 F.2d 1297, 1303 (9th Cir. 1982); Gibson, supra, n. 37; 
A sb esto s  Workers, Local 53 v. Vogler, 407 F.2d 1047, 1053-54 
(5th Cir. 1969). lb  prevail, respondents must establish both 
the existence of nepotism and its impact on the protected class. 
See id.; E E O C  v. S h ee t M etal Workers, Local 122, 463 F. Supp. 
388, 422 (D.Md. 1978) (inference of discrimination through 
nepotism negated by number of blacks entering apprenticeship 
program under affirmative action plan); U nited S ta tes v. 
Jacksonville Terminal Co., 316 F. Supp. 567, 592, n. 36 (M.D. 
Fla. 1970), aff'd in relevant part, rev d  and rem anded on o ther  
grounds, 451 F.2d 418 (5th Cir. 1971), cert, denied  406 U.S. 906
(1972) (no showing policy of nepotism invoked).

Respondents’ proof that nepotism existed consisted of 
evidence that relatives were employees. No policy of prefer­
ential treatment was shown to exist and, importantly, re­
spondents failed to establish the crucial element of causation: 
not a single instance (let alone a pattern) of a relative being 
hired in an at-issue job because o f  that relationship was proven. 
The District Court considered respondents’ evidence, but 
declined to draw the inference respondents urged. Instead, it 
was found that there was “no ‘preference’ for relatives” (Pet. 
App. 1:105); that employees were “chosen because of their 
qualifications” (Id.) after being “evaluated according to job 
related criteria” (Id,  1:122); and “that numerous white persons 
who ‘knew’ someone were not hired due to inexperience” (Pet. 
App. 1:122-23).” These findings of fact were not clearly 
erroneous and the Ninth Circuit should not have ignored them 38

38 Given these findings, the District Court’s reference to the “inci­
dence of nepotism" being "present” (Pet. App. 1:103, 1:105) must 
be read as simply references to the existence of the fact that 
relatives were hired, not to nepotism as a term of art.

88



26

in finding that a practice of discriminatory nepotism did exist 
(Pet. App. VI:19-21).39 Anderson v. Bessemer City, 470 U.S. 
564 (1985).

In addition, the statistics on which respondents relied to 
establish both that nepotism existed and that it had a dis­
parate impact were severely distorted by gross overcounting 
due to unproven assumptions and obvious methodological 
errors.40 These flaws justified rejection of the statistics. 
Teamsters, supra, 431 U.S. at 340, n. 20. The Ninth Circuit 
simply assumed that respondents’ statistics were accurate. 
Pet. App. VI:21. The 349 “nepotistic hires” referred to is based 
on the evidence found to be flawed by the District Court. 
Pet. App. 1:105.

99 The Ninth Circuit’s finding is particularly confusing because the 
court accepted the District Court’s findings relating to the hiring 
of relatives (see Pet. App. VI.-20-21) and the same panel had found 
in their first opinion that nepotism did not exist. See 768 F.2d 
at 1126, 1133 (Pet. App. 111:22-23, 56).

40 Flaws in methodology were pointed out in cross-examination. J.A. 
407-413. Among these flaws were: (1) every hire that is counted 
in the tables assumes that respondents have otherwise established 
that the person was hired because of the relationship, rather than 
for some other reason, such as skill. The trial court found other­
wise. (2) Respondents’ statistics failed "to differentiate those 
persons who became related through marriage after starting work 
in the canneries.” Pet. App. 1:105. (3) Respondents counted as two 
nepotistic hires both persons who were related at a cannery. J.A. 
410. Obviously, one of them had to be hired first and should not 
be counted at all. This factor alone means that respondents’ tables 
overstate the number of hires by approximately 50%. (4) 
Respondents continue to count the same employee year after year 
as being a new “nepotistic hire” so long as he was employed, 
regardless of when he was first employed or why he was first hired. 
Illustrative of the flawed methodology is the fact respondents 
counted three men a total of seventeen times in the machinist 
department at Bumble Bee (of a total of 28). See Ex. 603 (E.R. 
57-59 "Machinists only”) (E. Puffinberger, Juola, Snyder).

89



27

Moreover, in deciding that the hiring of relatives (even if proven 
to be “nepotism”) had no disparate impact on minorities, the 
District Court was entitled to consider the fact that minorities 
were not underrepresented in the at-issue jobs.41

C. Respondents Failed to Prove Unlawful Discrimination 
Under the Disparate Impact Theory in Housing, 
Messing, or So-Called “Racial Labeling”

Respondents claim that the disparate impact of the 
housing and messing practices was that class members were 
racially segregated and deprived of job opportunities. The 
District Court found that petitioners did not house or feed 
employees based on raca The “segregation” was not racial—it 
was based on factors such as job crew and date of arrival. 
Most employees lived in integrated housing. Whatever 
imbalance did exist in the bunk houses and mess halls, existed 
primarily because of the racial composition of the cannery 
worker crews. As was true of hiring, if Local 37 had not 
dispatched an overrepresentation of nonwhites, respondents 
would not have a claim of “segregation" (i.e., imbalance) in 
housing or messing.

Tb the extent there were differences in food in different 
mess halls, it was attributable to personal taste and the ability 
of the cooks, not race. Separate messing during the season42
41 See Scott v. Pacific Maritime Ass'n, 695 F.2d 1199, 1207-08 (9th 

Cir. 1983); Presseisen v. Swarthmore College, 442 F. Supp. 593, 
625-26 (E.D. Pa. 1977) (“old boy network" for filling faculty 
positions not discriminatory in absence of evidence of under­
utilization of women), aff'd, 582 F.2d 1275 (3d Cir. 1978).

42 All employees at all canneries ate in only one mess hall during the 
pre-season and the post-season, i.e., before and after the cannery 
workers arrived. It was only during the season that the second 
mess hall was opened up at the canneries with a Local 37 cannery 
worker crew (i.e., other than CWF-Ekuk) when the large influx of 
cannery workers arrived just before canning started. E.g., J.A.

(footnote continued on next page)

90



28

and a different menu for the Local 37 crew was demanded by 
union leaders, reflected in the union contract, and desired by 
a large number of the Local 37 cannery workers — requests 
and demands to which management acceeded.

The Ninth Circuit made a finding that the “impact is 
clear” of housing and messing practices (Pet. App. VI:36) 
because nonwhites were deprived of job information at the 
canneries. Pet. App. VL36-37.43 However, the evidence showed 
and the District Court found that job opportunities during 
the season were “rare”; that whatever openings were filled then 
were filled from outside the cannery work force; and that union 
contracts providing for payment of guaranteed wages dis­
couraged mid-season transfers because the company might 
have to pay double guarantees. Morever, there was not time 
for mid-season training of inexperienced, unskilled personnel. 
Thus, job opening information would have had little value to 
any employee.44 The complained-of practices had no effect once 
the season was over, at which time all employees were free to 
apply for work at the company offices. Pet. App. 1:33, FF 94.

Most Local 37 cannery workers did not object to a separate 
mess hall or to the food served therein, but employees who did

422-24. Even then, if the crew is small enough, the cannery 
workers will be fed with the other employees in a single mess hall. 
See, e.g., R.T. 2773, 2803. (Red Salmon 1977: because of limited 
canning operation, all employees, including cannery workers, fed 
in a single mess hall): R.T. 2316 (South Naknek 1980). Ekuk, which 
had no contract with Local 37, had one mess hall. (R.T. 2441.

43 In so finding, the Ninth Circuit cites not to the evidence, but 
to another case tried by plaintiffs' counsel involving a different 
company and different facts. Pet. App. VI:37.

44 Plaintiffs also made no showing that whatever little job opening 
information was available during the season was discussed only 
in bunkhouses or during meals, as opposed to during working 
hours, during mugups, or during off-hour recreation periods when 
job crews intermingled freely.

91



29

could change mess halls if they gave notice.45 Thus, class 
members could “opt out” of any alleged impact. Cf. Garcia 
v. Gloor, 618 F.2d 264 (5th Cir. 1980), cert, denied, 449 U.S. 
1113 (1981) (English-only rule did not discriminate where 
bilingual plaintiff could avoid it by speaking English on job). 
Tb the extent they did not do so, it was either personal prefer­
ence or “peer pressure.” See e.g., J.A. 620; 463-64, 114.

Even if there was a prima facie case to rebut, petitioners 
demonstrated the business justification for each of the prac­
tices because they “significantly served” the petitioners’ 
"legitimate business goals” of efficient and economic use of 
its scarce resources in housing, and accommodating the prefer­
ences of a significant number of class members and the de­
mands of their union representatives in messing. New York 
City Transit Auth. v. Beazer, 440 U.S. 568, 587 n.31 (1979); 
Watson, supra, 101 L. Ed. 2d at 827 (O’Connor, J. plurality); 
Contreras v. Los Angeles. 656 F.2d 1267, 1275-80 (9th Cir.
1981), cert, denied, 455 U.S. 1021 (1982). Respondents offered 
no alternatives. See J.A. 455-46 (discussing practical effect 
of respondents’ housing contentions).

The District Court also found that race labeling was used 
by whites and nonwhites alike, but that it was not evidence 
of intentional discrimination as urged by respondents. Pet. 
App. 1:123. Respondents have offered no evidence that race 
labeling, whether done by whites or by class members, had a 
significant disparate effect on nonwhite job opportunities. 
If such labels were overheard by or used by class members, 
respondents have made no showing that this excluded them 
from job opportunities. Tb the extent respondents claim it 
resulted in “deterrence” of nonwhite applicants, the District 
Court found otherwise. Pet. App. 1:123. It did not result in 
a significant underrepresentation of nonwhite employees in 
the at-issue jobs. * 1
45 See R.T. 2393. 1 1-2; 2394-95, 11 6-10; and 2413, Ins. 15-20 (see 

excerpts in J.A. 432-33); R.T. 2542,1 11 (J.A. 435-36); R.T. 2708,
1 11; R.T. 2713, Ins. 1-30; R.T. 3190, 11 7. 9 (J.A. 587-88); Dep. of 
Leonardo (4/5/78), p. 37, Ins. 10-14.

92



30

II. Allowing Respondents to Challenge the Cumulative Effect 
Of An Entire Range of Non-Racially Motivated Employ­
ment Practices Based Merely On A Showing The Peti­
tioners’ Work Force Reflects An Uneven Racial Balance 
Is An Improper Application Of The Disparate Impact 
Model, Unfairly Allocates The Burdens of Proof And 
Encourages Conduct At Odds With The Purposes of 
Title VII.

A. Respondents Are Required to Prove the Causal Effect 
of Each Practice They Choose to Challenge Under the 
Impact Model.

The impact model was not designed for this type of shot­
gun, undifferentiated attack on a large number of diverse 
employment practices. In Pouncy v. Prudential Ins. Co. of 
America, 668 F.2d 795, 800 (5th Cir. 1982), the court held 
that the disparate impact model is not “the appropriate vehicle 
from which to launch a wide-ranging attack on the cumulative 
effect of a company’s employment practices.48 Accord Carroll 
v. Sears, Roebuck & Co., 708 F.2d 183, 189 (5th Cir. 1983); 
Robinson v. Polaroid Corp., 732 F.2d 1010, 1014,1016 (1st Cir. 
1984); A.F.S.CM.E. v. State of Wash., 770 F.2d 1401, 1405-06 
(9th Cir. 1985) (Kennedy, J.) (impact analysis limited to chal­
lenge of “a specific, clearly delineated employment practice 
applied at a single point in the selection process”; “decision 
to base compensation on the competitive market, rather than 
on a theory of comparable worth, involves an assessment 
too multifaceted to be appropriate for the disparate impact 
analysis"). The Court of Appeals unpersuasively sought to 
distinguish a similar case (Spaulding v. Univ. of Wash., 
740 F.2d 686 (9th Cir. 1984), because respondents here had 
“identified” the practices. Pet. App. V:38. n. 6.
48 In its first panel opinion, the Ninth Circuit described plaintiffs’ 

case here: “By and large, however, [plaintiffs) have not challenged 
a specific facially neutral practice. Rather (plaintiffs) have 
mounted a broad-scale attack against the gamut of defendants’ 
subjective employment practices." Pet. App. 111:48.

93



31

In Pouncy, as here, the employer had "an uneven racial 
balance” in the work force in which nonwhites were "over­
represented in the lower levels.” 668 F.2d at 800, 801. The 
plaintiff "singled out” three employment practices as being 
discriminatory, but because he relied only on cumulative hiring 
results, he could not show "that independent of other factors, 
the employment practices he challenge[d). . . caused the racial 
imbalance in Prudential’s work force.” Id. at 801. Petitioners 
submit that the Pouncy view is correct, particularly when 
applied to the facts of this case.

Respondents here chose 16 different practices that they 
assert had a discriminatory effect in job allocation. See n. 14, 
supra. With but one exception,47 they point to the same and 
only set of cumulative comparative statistics as evidence of 
the disparate impact of each and all of these practices. But 
by so doing, respondents necessarily must concede that they 
cannot prove causation by any one of the challenged practices. 
Without proof of causation, however, respondents’ impact 
claims must fail at the threshold. Pouncy, supra, 668 F.2d at 
800-802; Robinson, supra; Carroll, supra, 708 F.2d at 189.

This causation requirement is implicit in the decisions of 
this Court: in each of the successful impact cases, the plain­
tiffs established the discriminatory effect separately for each 
practica See, e.g., Griggs v. Duke Power Co., 401 U.S. 424
(1971) (high school diploma and aptitude test); Dothard v. 
Rawlinson, 433 U.S. 321 (1977) (height/weight requirements); 
Connecticut v. Thai, 457 U.S. 440 (1982) (test). This past 
Tferm, Justice O’Connor, speaking for a plurality of the Court, 
recognized this fundamental threshold burden on plaintiff:

[T]he plaintiff is in our view responsible for isolating and
identifying the specific employment practices that are

47 Nepotism — see discussion, supra, pp. 24-26.

94



32

allegedly responsible for any observed statistical dis­
parities. Cf. Connecticut v. Teal, 457 U.S. 440 (1982).
Once the employment practice at issue has been 
identified, causation must be proved; that is, the plaintiff 
must offer statistical evidence of a kind and degree suffi­
cient to show that the practice in question has caused the 
exclusion of applicants for jobs or promotions because of 
their membership in a protested group. (Emphasis added.)

Watson, supra, 101 L. Ed. 2d at 845.
Respondents urge a “collective” approach because their 

proof as to the existence or effect of individual practices failed. 
For instance, respondents complain that the petitioners fill 
at-issue jobs by “word-of-mouth recruiting.”4® The implicit 
assumption is that this practice must result in dispropor­
tionate numbers of whites being hired. That assumption can 
be tested by comparing the racial results of employers’ hiring 
with the racial composition of the relevant labor market. 
E.g., Clark, supra, 673 F.2d at 927-929.48 49 Here, that very 
comparison showed nonwhites were not significantly under­
represented in the jobs at issue. Pet. App. 1:42-43, FF 123; 
Ex. A-278, Tbl. 4 (for each cannery or combination of canneries) 
(E.R. 2-7).

With respect to the practice of hiring nonresident cannery 
workers through Local 37, as discussed, supra, pp. 22-23, there 
was no evidence that this practice disproportionately excluded 
nonwhites from the at-issue jobs.

Some of the practices “named” as causing racial dis­
parities in hiring did not even exist, were unproven or were
48 Word-of-mouth recruiting was only one of several methods of 

finding employees. Petitioners proved and the District Court 
found that petitioners accepted walk-in applicants and also looked 
to unions with appropriate jurisdiction as a source of employees.

49 Plaintiffs attempted to so prove impact with their own labor 
market theory. It postulated nonwhites were 50% of the available 
labor supply and assumed that nearly all of the at-issue jobs 
required no skills, but it was rejected.

95



33

irrelevant. The trial court found the petitioners did not dis­
criminate in terminations (individually or by ceasing to recruit 
in remote villages), did not discriminate in pay in any jobs, did 
not grant preference to relatives (“nepotism”), did not dis­
criminate in fraternization restrictions or in assigning “menial 
tasks,” and, however “informal” the petitioners’ promotion 
procedures were, it did not matter because petitioners did not 
promote from within anyway.50

Finally, respondents’ case failed on several practices 
because they were found to meet the “job relatedness” or 
"business necessity” test.51 See infra, p. 34.

Since respondents were unable to establish their case as 
to any practice separately, their burden should not be lessened 
on the mere allegation that the practices “collectively” caused 
an impact, particularly where that “impact” is nothing more 
than lack of racial balance.

50 The remaining challenged practices, e.g., “labeling,” were, again, 
not separately analyzed by plaintiffs for their effect on hiring. 
Presumably, their claim was that these practices had the effect 
of “deterring ’ or “chilling” nonwhites from applying. However, 
the District Court found that there was no deterrence. Pet. App. 
1:123.

51 E.g., English language requirement: rehire preference.

96



34

B. Dispensing With the Causation Requirement for 
Plaintiffs Places an Unfair Burden on Employers, and 
Encourages Conduct at Odds With Title VIPs 
Purpose.

If the plaintiff fails (or refuses) to show the causal con­
nection between any practice and any showing of disparate 
impact, the employer does not know which practice (or all) 
he must justify under the “business necessity” or “job­
relatedness” defense.5* This is unfair in and of itself, but as 
applied by the Ninth Circuit, the employer must justify all 
of the practices named by the plaintiff.

Here, the District Court found that employees in the at- 
issue jobs were “hired according to job-related criteria,” and 
that the English language requirement, the rehire preference, 
housing practices, and messing practices were all business 
necessities. That is, the District Court found that the peti­
tioners met their burden on at least five of the 16 practices 
named — and the Ninth Circuit affirmed on at least two 
(language requirement and rehire preference). (And, as dis­
cussed above, several others were shown not to exist.) If the 
plaintiffs insist that all of the practices “combined” to cause 
the impact but failed to show the causal connection for any 
one, then proof of business necessity of one of those practices 
should satisfy the employer’s burden, absent evidence by 
plaintiffs that the remaining practices had a significant effect. 
Such proof on five of those practices is present here. Tb require 
an employer to prove the necessity of all practices would simply 
encourage a plaintiff to name as many "practices" as he could 
in the reasonable expectation that the employer could show 
the business necessity of less than all. Pouncy, supra, 668 
F.2d at 801. **

** Nor would the trial court know which practice(s) to "change at 
the injunctive relief stage if the employer fails in his burden.

97



35

Under respondents’ theory, the burden of justification 
could still be shifted to the employer if there is a disparity, 
but it is caused by a practice not identified by plaintiff. For 
instance, here the employers filled at-issue jobs from walk-in 
applicants and by referrals from other unions, as well, but 
respondents did not challenge these practices, nor did they 
attempt to prove their relative significance.

It might be argued that the employer should be required 
to prove causation where the plaintiff is unable to do so. E.g., 
Griffin v. Carlin, 755 F.2d 1516, 1526-28 (11th Cir. 1985). First, 
this completely reverses the causation requirement explicitly 
stated in Watson, supra, 101 L. Ed. 2d at 845 (O’Connor, J., 
plurality), and implicit in this Court’s other impact opinions. 
Second, it is grounded on the belief that plaintiffs might not 
be able to obtain evidence of causation of the practices they 
challenge through normal civil discovery. If these processes are 
sufficient to allow plaintiffs the means to prove discriminatory 
motive,63 they should also suffice to prove discriminatory 
impact.*4

Third, the practical effect of reversing this burden of proof 
is both staggering and ominous. While the employer would 
know the reasons he has used certain practices, it is entirely 
unlikely that he does or could keep track of the statistical 
effect of each possible practice that might be litigated on each 
protected class and subclass that might raise a Title VII 
claim against the business. Obviously, this would be a 
mammoth effort because:

(It is) unrealistic to suppose that employers can eliminate,
or discovery and explain, the myriad of innocent causes * 54

"  Tbxas Dep t Community Affairs v. Burdine, 450 U.S. 248, 258 (1981).
54 Indeed, numerous plaintiffs have succeeded in establishing both 

impact and causation where they attack more than one practice. 
E.g., Griggs, supra, 401 U.S. 424 (diploma requirement and test); 
Clady v. County of Los Angeles. 770 F.2d 1421 (9th Cir. 1985) 
(written exam, education requirement, physical agility test), cert, 
denied, 475 U.S. 1109 (1986).

98



36

that may lead to statistical imbalances in the composition 
of their work forces.

Watson, supra, 101 L. Ed. 2d at 843 (O’Connor, J. plurality).

Many would find it more practical, particularly if plaintiff 
is allowed to proceed with a base showing of racial imbalance, 
simply to adopt an in-house policy of maintainig strict racial 
and gender balance in all job categories, i.e., quota hiring 
and layoffs, a specter this court warned about only last Tferm 
in Watson, supra, 101 L. Ed. 2d at 843-44 (Section 11—C) 
(plurality), 856 (Stevens, J. concurring in the judgment, citing 
Section II-C of plurality opinion favorably).

However, even if the petitioners in this case had such a 
burden, they have met it. The proof of disparate impact 
credited by the Ninth Circuit was respondents’ showing of 
imbalanca As to one "side” of the scales: the trial court found 
that the principal cause of the overrepresentation of nonwhites 
in the cannery worker jobs was the dispatching practices of 
Local 37, ILWU. As to the other side: the relatively low 
percentage of nonwhites in the at-issue jobs is attributable to 
hiring laterally from the relevant labor market that happens 
to be approximately 10% nonwhite and to the “rehire” practice
i.e., rehiring persons returning in the same job that they held 
the preceding season.*® Plainly, all of these "practices” were 
justified and respondents offered no practical alternatives.55 56 * 58

55 See Pet. App. 1:33, FF 95; Ex. A-320(a), pp. 3-5: 85% of all at issue
jobs are filled by either rehires (40%) or new hires from outside
petitioners' workforce (45%).

58 When one examines the "alternatives” to eliminate the imbalance 
of which plaintiffs complain, the fallacy of plaintiffs’ “hiring 
channels” argument becomes readily apparent: an obvious, cost- 
effective way is to stop hiring from Local 37 and begin hiring all 
cannery workers from the 10% nonwhite general labor market.

99



37

III. The Ninth Circuit Improperly Shifted The Burden Of 
Proof To Petitioners.

On remand from the en banc court, the Ninth Circuit 
panel proceeded to fashion a new allocation of the burdens of 
proof in an impact case, drastically lowering respondents’ and 
raising petitioners’

A. Respondents Did Not Meet the Initial Requirements 
to Establish An Impact Case.

As in any civil lawsuit, the plaintiff must bear the ultimate 
burden of persuasion; this is equally applicable to the impact 
or treatment models. New York City Transit Authority v. 
Beazer, 440 U.S. 568, 587, n. 31 (1979) (impact); Tkxas Dep't 
of Community Affairs v. Burdine, 450 U.S. 248, 253 (1981) 
(treatment); Fed. R. Evid. 301. Accord Watson, supra, 101 L. 
Ed. 2d at 847 (plurality).

Before any burden can be shifted to the employer, the 
plaintiff must establish a prima facie impact claim. These ele­
ments include (1) a significant statistical disparity (2) caused 
by an employment practice. E.g., Watson, supra, 101 L. Ed. 
2d at 851 (Blackmun, J. concurring in judgment). In fact, the 
Ninth Circuit recognized that the respondents’ burden in the 
impact case is more onerous than in a treatment case. Pet. 
App. V:37.ST The Ninth Circuit paid lip service to these re­
quirements but did not apply them. Pet. App. V:37; VI-3, 19.

The Ninth Circuit said that respondents’ comparative 
statistics, showing only racial stratification by job category, 
were ‘sufficient to raise an inference that some practice or 
combination of practices has caused the distribution of em­
ployees by race and to place the burden on the employer to 
justify the business necessity of the practices identified by 
plaintiffs.” Pet. App. V:18 (emphasis supplied). 57
57 The burden of proof on the employer is commensurate with the 

greater burden on the plaintiff to prove impact and establish the 
causal connection." The Ninth and Fifth Circuits have held that

(footnote continued on next page)
100



38

As noted above,58 respondents' comparative statistics, 
coupled with a litany of practices, is not adequate to establish 
disparate impact under the decisions of this Court. The 
ultimate question is whether petitioners did engage in racial 
discrimination. Respondents’ mere proof of prima facie treat­
ment case (described as “marginal” by the District Court for 
the skilled jobs) establishes only an inference of discrimination. 
See infra, pp.41-42. The Ninth Circuit erroneously concluded 
this showing sufficed to prove impact. Pet. App. VI:4-5 (proof 
of prima facie case identical under both theories). Under the 
impact model, the plaintiff must establish more than mere 
inference, he must establish that the practice has an improper
effect. Watson v. Fort Worth Bank & Trust, 487 U.S-------
101 L. Ed. 2d 827, 851 (1988) (Blackmun, J., concurring in 
judgment).

B. The Ninth Circuit's Decision Conflicts With the Order 
of Proof Requirements of Burdine.

The Ninth Circuit holds that petitioners' treatment case 
“explanation” supplies the missing elements of respondents’ 
case on causation, and also makes unnecessary the consid­
eration of petitioners’ labor market evidence and attack on 
respondents’ statistics. It recognizes that under the order of 
proof for a treatment case the employer is only required to 
meet respondents’ prima facie case with the articulation of a 
nondiscriminatory reason for the selection process. Texas 
Dep't of Community Affairs v. Burdine, 450 U.S. 248, 254
(1981). Pet. App. 111:16. The court also recognizes that

for a disparate impact, plaintiff must not merely prove circum­
stances raising an inference but must prove the discriminatory 
impact at issue. Moore v. Hughes Helicopters, Inc., 708 F.2d 475, 
482 (9th Cir. 1983); Johnson v. Uncle Ben's, Inc.. 657 F.2d 750, 753 
(5th Cir. 1981), cert, denied, 459 U.S. 967 (1982).

58 Section I.A.; II.

101



39

petitioners do not have to accept respondents’ statistics 
and may introduce statistics of their own. Pet. App. VI:5. 
However, the court states that if the employer defends by 
explaining the reason for the disparity, articulation is insuf­
ficient; the employer must then prove the business necessity 
of the named practices. Pet. App. VI:5.

Petitioners did articulate, and prove to the trial court's 
satisfaction, a number of nondiscriminatory reasons for the 
disparity: institutional factors caused stratification, the em­
ployers did not promote from within, transferring personnel 
between departments during the season required payment of 
two guarantees, there was insufficient time to train inex­
perienced help for most jobs, skills of a cannery worker are 
not a substitute for the skill and experience requirements 
of the skilled non-cannery jobs, and that the relevant labor 
market is 90% white.

The Court of Appeals states — incorrectly — that peti­
tioners "conceded” causation. Pet. App. VI:24. This appar­
ently is that court’s view of the explanation offered by peti­
tioners to meet the treatment claims. The Court of Appeals 
erroneously cites Albemarle Paper Co. v. Moody, 422 U.S. 405, 
425 (1975), in support of the proposition that explanation 
of the imbalance shifts the burden to the employer to show 
business necessity. Pet. App. VI:5. Albemarle holds that once 
the impact of a practice is established, the employer has such 
a burden.

A similar citation to Albemarle may be found in Segar v. 
Smith, 738 F.2d 1249 (D.C. Cir. 1984), cert, denied sub nom 
Meese v. Segar, 471 U.S. 1115 (1985). In Segar, plaintiffs 
advanced a treatment case. The Court of Appeals there specu­
lated that the degree of proof required of an employer in 
defending a class action case — as opposed to one of individual 
discrimination such as Burdine — might require more than 
just articulation of a reason to succeed. In this process the 
court states, an employer "will in all likelihood” point to a

102



40

specific job qualification as an explanation for the disparity. 
738 F.2d 1249, 1271. The statement is dicta since the court 
affirmed a finding of disparate treatment and went on to say 
that if the employer had advanced the requirement of an 
additional year’s experience as the reason for the disparity 
that discrete requirement “would have been" subject to an 
impact analysis. 738 F.2d 1249, 1288. A similar statement 
appears in Griffin v. Carlin, 755 F.2d 1516, 1528 (11th Cir.
1985), where the court stated that if on remand the plaintiff 
succeeded in establishing a treatment case, and if the employer 
defended by reliance on a supervisory register and a test, the 
employer had to validate those procedures. Simply stated, 
Segar and Griffin merely hold that if an employer defends a 
treatment case by explaining that the disparity is caused by 
a test, he may have to defend the test.

No circuit court appears to have actually applied this 
requirement nor have any circuits followed Segar or Griffin 
on this point.59

The concept advanced by the Ninth Circuit here is much 
broader than Segar or Griffin and is directly at odds with the 
holding of Burdine. There, the Fifth Circuit required the 
employer to prove — not merely articulate — a nondiscrimina- 
tory reason for the employer’s conduct. 450 U.S. at 252. This 
Court held that to be error, that the employer need only 
come forward with evidence sufficient to allow an inference 
of nondiscriminatory conduct. The policy reasons stated are 
sound: it is plaintiffs' case; defendants’ explanation must 
be clear enough to allow an attack on pretext grounds; the 
employer has the incentive to persuade; and liberal discovery 
rules are supplemented by the EEOC investigatory files.

59 The Third Circuit in Green v. USX Corp., 843 F.2d 1511 (3d Cir. 
1988), pet. for writ of cert, filed (No. 88-141), has approved the use 
of a multiple practice impact analysis in reliance on Segar and 
Griffin.

103



41

C. Respondents’ Evidence Was Not Adequate Under 
Either Impact or Treatment Order of Proof Require­
ments.

In Burdine, this Court described the establishment of a 
prima facie case as evidence which — if believed — and if 
the employer is silent — requires the entry of judgment for 
plaintiff. 450 U.S. 248, 254. This prima facie case is used 
in the sense of a rebuttable presumption; but the rebuttal is 
made by offering evidence which need not persuade the court 
of nondiscrimination but merely raises an inference of such 
conduct. Id. at 254.80 This form of presumption of discrimina­
tion, though fragile, was the one adopted in Fed. R. Evid. 301.81 * 61

80 This Court recognized that a prima facie case in common law may 
either refer to the level of evidence sufficient to allow a case to go 
to a jury, or a legally mandatory rebuttable presumption. The 
court used the term in the latter sense. The presumption dis­
appears as such when countervailing evidence is produced — even 
though the countervailing evidence is not believed. At that point 
in the case, the trier of fact may still consider both sides of the 
evidenca 450 U.S. 254, citing 9 Wigmore, Evidenca § 2491 (3d 
ed. 1940) and Fed. R. Evid. 301. In adopting Rule 301, this Court 
accepted the Thayer or “bursting bubble" view of presumptions 
rather than the Morgan view, which would give greater effect to 
a presumption than the mere burden of putting in evidence which 
may be disbelieved by the trier of fact. Wigmore, § 2493C (1981).

61 Rule 301. In all civil actions and proceedings not otherwise 
provided for by Act of Congress or by these rules, a presumption 
imposes upon the party against whom it is directed the burden 
of going forward with evidence to rebut or meet the presumption, 
but does not shift to such party the burden of proof in the sense 
of the risk of non-persuasion which remains throughout the trial 
upon the party on whom it was originally cast.

104



42

The use of the term prima facie case in the impact case 
may be analyzed on the same basis. If the plaintiff comes 
forward with statistics and other evidence showing that a 
specific practice has disproportionately excluded a protected 
group, he has made a prima facie case which will entitle him 
to entry of judgment if the employer remains silent. If the 
employer comes forward with his own statistics showing (but 
not necessarily proving) no disproportionate exclusion or that 
the practice complained of may not have caused the disparity, 
or if on cross-examination he shows flaws in plaintiffs’ statis­
tics, or impeaches or discredits plaintiffs’ witnesses, he has 
met his burden of production and the trier of fact may believe 
either side’s witnesses. Under a strict reading of Griggs, if 
the employer remains silent on the issue of disparate impact, 
that issue is established and he then must come forward with 
what amounts to an affirmative defense of business necessity. 
See Wigmore, § 2487 (1981).62 Such a strict reading has 
recently been questioned by a plurality of this Court. Watson, 
supra, 101 L. Ed. 2d at 847.

Here, however, petitioners did not stand silent before 
respondents’ evidence That evidence was vigorously attacked 
a9 to reliability and credibility.63 The respondents’ labor 
market, nepotism tables, and comparative statistics were all 
shown to be flawed. Their contentions of fraternization 
restrictions were flatly disproved by evidence of fraternization. 
Their claims of individual discrimination were demonstrated 
to be without merit because of lack of application, untimely 
application, lack of qualification, or the jobs were full at the

62 Citing Speasu v. Merchants Bank & TYust Co., 188 N.C. 524, 529, 
125 S.E. 398, 401 (1924), “The burden of proof continues to rest 
upon the party who, either as plaintiff or as defendant affirma­
tively alleges facts necessary to enable him to prevail in the cause"

63 The Court commited plain error in concluding that petitioners did 
not argue the practices had “no impact." Pet. App. VI:30. This 
error was pointed out in the Petition for Rehearing. Pet. App. VII.

105



45

Other circuits are in accord. Ste. Marie v. Eastern R. Assoc., 
650 F.2d 395, 400-401 (2d Cir. 1981); Coser v. Moore, 739 
F.2d 746, 752 (2d Cir. 1984); EEOC v. Fed. Reserve Bank of 
Richmond, supra, 698 F.2d at 658-660.

It is not surprising that respondents chose not to account 
for even the most basic qualifications of the “proxy” popula­
tion of potential employees. Petitioners did so with their labor 
market analysis and it established that qualified nonwhite 
availability was closer to 10% than to the 50% argued by 
respondents. See J.A. 255-56,1 7; 258-60,11 12-13. The Ninth 
Circuit demanded that petitioners prove the qualified nonwhite 
component in the labor market (Pet. App. VI:17, 26), but 
ignored the evidence doing just that. This evidence accounted 
for the fact that the different “job families” at the canneries 
required different skills and experience (e.g., machinists vs. 
cooks vs. carpenters). As adjusted, the nonwhite availability 
percentages range from about 2% (administration) to 20% 
(culinary), depending on the job family, and “centered” on 
10% for the at-issue jobs combined. Exhibit A-278, labor pool 
tables, Tbls. 4(a)-4(b), row “Nonwhite” for each job family 
column (E.R. 8-9). These adjusted availability figures were 
compared to actual hiring in the job families at issue: 
nonwhites were not significantly underrepresented. See J.A. 
266-280. Use of the unadjusted availability figures does not 
change the conclusion. See supra, n.29.

The panel also placed the burden on petitioners to prove 
why they did not hire from different sources for at-issue jobs,
e.g., promote from within, target Local 37 as a source of 
machinists, or scour the remote areas of Alaska for persons 
to fill at-issue jobs.66 Pet. App. VI:30. In doing so, the Ninth 
Circuit is plainly substituting its judgment for management 
as to the best way to operate the business. The Ninth Circuit 
is merely attempting to require petitioners to maximize the

66 A practice the District Court found unreasonable. Pet. App. 1:32. 
FF 90.

108



46

number of minority workers hired. This is a flat violation of 
the admonitions of this Court in Furnco Constr. Corp. v. 
Waters, 438 U.S. 567, 577-78 (1978); Burdine, supra, 450 U.S. 
at 259; and Watson, supra, 101 L. Ed. at 848.

Moreover, since nonwhites are not underrepresented, the 
Ninth Circuit’s demand is inconsistent with this Court's
opinion in Johnson v. Transp. Agency, 480 U.S------- 94 L.Ed.
2d 615 (1987) (inappropriate to adopt voluntary affirmative 
action plan to boost minority hiring in jobs where there is no 
underrepresentation of minority workers).

E. Clarification of the Order of Proof.

It is apparent that the Ninth Circuit misunderstood the 
proper allocation of proof and sailed into uncharted seas. 
When this Court defined the order of proof in Burdine, it was 
clarifying the discrete order of proof to raise a prima facie case 
under McDonnell Douglas. The Court should offer guidelines 
with respect to the order of proof in impact cases. Petitioners 
submit this can be accomplished within the principles an­
nounced by both the plurality and concurring opinions in 
Watson.

A plaintiff may offer various forms of evidence in his case 
in chief to show discrimination. In the impact case, he will 
offer statistics and evidence of causation, and he may proceed 
simultaneously on a treatment theory and offer McDonnell 
Douglas evidence and anecdotes. The employer, then, has an 
opportunity to come forward. The employer should be able 
to meet his burden of production with evidence showing that 
no inference of discrimination should be drawn. If plaintiff 
relies on a mere imbalance in job classifications as his impact 
case, the employer should be able to show — as in this case 
— that he chooses to hire his noncannery workers from lateral 
sources. He may also show that there are practical, non- 
discriminatory reasons for not promoting or transferring from

109



47

within. The employer thus meets the inferences put forward 
by plaintiff with inferences showing lack of discrimination. 
He may also offer evidence of the business necessity of a 
discrete practice. The plaintiff can still attempt to show 
pretext. When all the evidence is before the trial court, it is 
weighed and the facts are found.

Prior to Watson, as the plurality noted, the impact analysis 
had not been extended into the context of subjective selection 
practices. TVaditionally, that analysis had been applied to rigid 
objective criteria which automatically disqualified a portion 
of the protected group, Dothard v. Rawlinson, supra, 433 U.S. 
at 338 (Rehnquist. J. concurring), or as Judge Sneed stated 
in the en banc proceeding, criteria which make the “plaintiff’s 
true qualifications irrelevant.” Pet. App. V:59. Such criteria 
are arguably subject to standardized testing and necessity, job 
relatedness and manifest relationship may be determined.

However, the exercise of sound business judgment is far 
less subject to testing or validation. Discretion by its very 
nature is never rigid. Those who survive in business are 
probably far better able to predict success than psychologists, 
economists, professors, and courts. Since it is his business at 
risk, an employer must be allowed the freedom to make legiti­
mate choices. The Watson plurality observed that the em­
ployer may find it easier in the context of subjective decision 
making to produce evidence of a manifest relationship. 101 
L. Ed. 2d at 848. Indeed, the rigid formula of Griggs itself 
should be re-examined in this context. In many cases the 
formula will be difficult to apply, particularly if the plaintiffs 
case is marginal or is a shotgun attack on all practices. In that 
situation, a showing of a legitimate business reason — rather 
than necessity — should be adequate. At that point, the 
plaintiff’s case is still in the inference stage and any 
countervailing evidence should be adequate rebuttal of the 
inference.

110



48

CONCLUSION

This case has been pending for nearly 15 years. It should 
end here. Respondents’ evidence was carefully considered by 
the trial court and found insufficient to prove class-wide or 
individual disparate treatment. That evidence is no stronger 
under the impact analysis. The Ninth Circuit resurrected 
respondents’ case by inappropriately applying the impact 
analysis. In so doing, it ignored binding legal precedent, 
erroneously reallocated the burdens of proof, and filled re­
spondents’ evidentiary gaps with its own fact-finding. The end 
result is an unwise decision that drastically reduces the quality 
and quantity of evidence expected of plaintiffs and imposes 
unfair and unrealistic burdens on employers. As such, it 
represents a major intrusion into the operation of American 
businesses that all but compels employers to take actions at 
odds with the salutory purposes of Title VII of the Civil Rights 
Act of 1964. The decision should be reversed with directions 
to enter judgment for petitioners on all claims

Respectfully submitted,

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ikkelborg , B roz. 
W ells & F ryer

, , _ . Attorneys for Petitioners* Counsel of Record

111





No. 87-1387

IN THE

Supreme Court of the United States

O ctober Term, IPHH

WARDS COVE PACKING COMPANY, INC. 
CASTLE & COOKE, INC ,

Petitioners,

v.

PRANK ATONIO, vt n l , 

Respondents.

BRIEF OF RESPONDENTS

Abraham A. Arditi 
Northwest Labor and 
Employment Levy Qflice 
900 Huge Pudding 
Seattle, Wuahington 98101 
1-200 G2.M 590
Con use I of Record 
for Respondents

Bob be .Jean Bridge 
Garvey, Schubert & Barer 
Waterfront Place Budding 
Seattle, WA 98104 
1 20G 10 1:1939
Counsel for Respondents

113



INDEX
Page

Statement of the Case..........................................................1
1. Introduction............................................................. 1
2. Racial Stratification In Jobs.................................... 3

A. Job Departments............................................... 3
B. The Statistics.................................................... 4

3. Race Labelling of Jo b s........................................... 6
4. Segregated Hiring Channels.................................... 7
5. Nepotism In Upper-Level Jobs..............................10
6. Lack of Objective or

Discernible Qualifications .....................................11
7. Re-Hiring Past Employees

In Their Old Departments.....................................14
8. Individual Instances of Discrimination................... 15
9. The Labor Supply.................................................15

10. Housing Segregation............................................. 19
11. Messing Segregation............................................. 20

Summary of Argument...................................................... 21
Argument............................   23

1. Statistics on Job Segregation Or Practices 
Which Foster It Establish Disparate Impact 
Regardless of What Labor Market Comparisons 
Show......................................................................23
A. The Language of Title VII Makes Job 

Segregation and Practices Which Promote
It Illegal........................................................... 23

B. Even the Labor Market Statistics WCP 
and BBS Offered Establish a Prima Facie
Case for Many Jobs................  25

C. Labor Market Comparisons Cannot Rebut 
or Justify Statistical Showings of
Job Segregation...............................................26

D. The Labor Supply Findings Were Induced
by Errors of Law..................  30

E. Alleged Skill Requirements Do Not Detract
From Work Force Statistics Here.................... 32

i

114



11

2. The Employees Did Not Have to Offer Statistics 
On Qualified Non-Whites, Since the Employers 
Never Identified Criteria Actually Applied,
They Lacked Objective Qualifications and the 
Qualifications They Did Use Had a Disparate 
Impact.................................................................. 32

3. Housing and Messing Segregation and Race-
Labelling Have a Disparate Impact on Non- 
Whites .................................................................. 35

4. Nepotism Has a Disparate Impact on Non-
Whites H ere......................................................... 37

5. The Employees Established Causation..................38
6. WCP and BBS Have Not Met Their Heavy

Burden of Proving Business Necessity....................43
7. The Court Should Also Affirm on Alternate

Grounds of Disparate Treatment..........................49
Conclusion....................................................................... 50
Appendices:

Appendix A................................................................ A-l
Appendix B................................................................ B-l
Appendix ................................................................... C-l
Appendix D ...............................................................D-l

Page

115



TABLE OF AUTHORITIES 
Cases

Albemarle Paper Co. v. Moody,
422 U.S. 405 (1975).......................  24, 29, 33, 43, 46, passim
Arizona Governing Committee v. Norris,
463 U.S. 1073 (1983).......................................................... 30,36
Atonio v. Wards Cove Packing Co., 703 F.2d 329
(9th Cir. 1983)...............................................................................1
Bazemore v. Friday,
478 U.S. 385 (1986)...............................................28, 36, 40, 49
Bigelow v. RKO Pictures, Inc., 327 U.S. 251 (1946)............ 43
Burrus v. United Telephone Company of Kansas, Inc.,
683 F.2d 339 (10th Cir. 1982), cert denied,
459 U.S. 1071 (1982)................................................................ 34
Bushey v. New York State Civil Service Commission,
733 F.2d 220 (2nd Cir. 1984), cert denied,
469 U.S. 1117 (1985)................................................................ 35
California Brewers Ass’n v. Bryant, 444 U.S. 598 (1980)... 48
Capaci v. Katz and Besthoff, Inc., 711 F.2d 647
(5th Cir. 1983), cert denied, 466 U.S. 927 (1984).............. 38
Carpenter v. Nefco-Fidalgo Packing Co.,
No. C74-407R (W.D. Wash. May 20, 1982).............. 2, 29, 33
Carpenter v. Steven F. Austin State University,
706 F.2d 608 (5th Cir. 1983).............................................25, 27
Castro v. Beecher, 459 F.2d 725 (1st Cir. 1972)..................  44
Caviale v. State of Wisconsin, 744 F.2d 1289
(7th Cir. 1984)............................................................................ 34
Chance v. Board of Examiners, 458 F.2d 1167
(2nd Cir. 1 9 7 2 )...................................................................44, 45
Chandler v. Roudebush, 425 U.S. 840 (1976)...................... 24
Chrisler v. Complete Auto Transit, Inc.,
645 F.2d 1251 (6th Cir. 1981)..................................................33
Colby v. J.C. Penney Co., 811 F.2d 1119
(7th Cir. 1987)............................................................................ 36

iii

116



iv

Connecticut v. Teal,
457 U.S. 440 (1982)......................... 21, 22, 23, 27, 28, passim
Consumer Product Safety Comm’n v. GTE Sylvania, Inc.,
447 U.S. 102 (1980)................................................................. 24
Croker v. Boeing Co. (Vertol Div.), 662 F.2d 975
(3rd Cir. 1981)........................................................................... 44
Curtis v. Loether, 415 U.S. 189 (1974)............................. 22, 36
Davis v. Califano, 613 F.2d 957 (D.C. Cir. 1979)................ 34
De Medina v. Reinhart, 686 F.2d 997 (D.C. Cir. 1 982 ) . . . .  34
Domingo v. New England Fish Co., 445 F. Supp 421 (W.D. 
Wash. 1977), reversed on other issues, 727 F.2d 1429 (9th Cir. 
1984), modified, 742 F.2d 520.......  29, 30, 31, 33, 36, passim
Domingo v. New England Fish Co.,
727 F.2d 1429 (9th Cir. 1984),
modified, 742 F.2d 520 (1984)........ 24-25, 29, 33, 34, passim
Dothard v. Rawlinson,
433 U.S. 321 (1 9 7 7 ).............................  25, 27, 30, 35, passim
EEOC v. J.C. Penney Co., 843 F.2d 249
(6th Cir. 1988)........................................................................... 36
EEOC v. Radiator Specialty Co., 610 F.2d 178
(4th Cir. 1979)............................................................................ 33
EEOC v. Rath Packing Co., 787 F.2d 318 
(8th Cir. 1986), cert denied,
107 S. Ct. 307 (1986).......................................................... 33,49
EEOC v. Shell Oil Co., 466 U.S. 54 (1984).......................... 41
EEOC v. S t Louis-San Francisco Ry. Co., 743 F.2d 739
(10th Cir. 1984).......................................................................... 35
Fadhal v. City and County of San Francisco, 741 F.2d 1163 
(9th Cir. 1984)............................................................................ 35
Franks v. Bowman Transportation Co.,
424 U.S. 747 (1976).................................................................. 33
Fumco Construction Corp. v. Waters,
438 U.S. 567 (1978)...........................................................28, 47
General Building Contractors Association, Inc. v.
Pennsylvania, 458 U.S. 375 (1982)..........................................31
Giles v. Ireland, 742 F.2d 1366 (11th Cir. 1984)................  25

117



V

Grant v. Bethlehem Steel Corp., 635 F.2d 1007 (2nd Cir.
1980), cert denied, 452 U.S. 940 (1981).......... 26, 36, 37, 48
Green v. USX Corp., 843 F.2d 1511 (3rd Cir. 1988), 
petition for cert filed, 57 U.S.L.W. 3123 (U.S. July 23, 1988)
(No. 88-141)............................................................................... 41
Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1 9 8 5 )........ 41, 42
Griggs v. Duke Power Co.,
401 U.S. 424 (1971)........................ 23, 24, 25, 27, 33, passim
Guardians Ass'n v. Civil Service Commission of the
City of New York, 463 U.S. 582 (1983).................................45
Hazelwood School District v. United States,
433 U.S. 299 (1977)..........................  1, 27, 29, 30, 32, passim
James v. Stockham Valves and Fittings Co., 559 F.2d 310 
(5th Cir. 1977), cert denied, 434 U.S. 1034 (1978) . . . .  25, 27
Johnson v. Uncle Ben's, Inc., 657 F.2d 750
(5th Cir. 1981), cert denied, 459 U.S. 967 (1982).............. 44
Keyes v. School District No. 1, 413 U.S. 189 (1 9 7 3 ).......... 45
Laffey v. Northwest Airlines, Inc., 567 F.2d 429 (D.C. Cir.
1976), cert denied, 434 U.S. 1086 (1977).............................34
Latinos Unidos De Chelsea v. Secretary of Housing,
799 F.2d 774 (1st Cir. 1986)...................................................41
Legate v. Maloney, 334 F.2d 704 (1st Cir. 1964),
cert denied, 379 U.S. 973 (1965)...........................................37
Lewis v. Bloomsburg Mills, Inc., 773 F.2d 561
(4th Cir. 1985)............................................................... 41,44, 46
Lilly v. Harris-Teeter Supermarket, 720 F.2d 326
(4th Cir. 1983), cert denied, 466 U.S. 951 (1984)..............  50
Lynn v. Regents of the University of California, 656 F.2d 
1337 (9th Cir. 1981), cert denied, 459 U.S. 823 (1982) . . .  34
Mackey v. National Football League, 543 F.2d 606 (8th Cir.
1976), cert dismissed, 434 U.S. 801 (1 9 7 7 )........................ 31
Markey v. Tenneco Oil Co., 635 F.2d 497
(5th Cir. 1981).....................................................................29,30
Mayor of Philadelphia v. Educational Equality League,
415 U.S. 605 (1974).................................................................35

118



vi

McDonnell Douglas Corp. v. Green,
411 U.S. 792 (1973)............................................... 24, 29, 33, 44
Meritor Savings Bank v. Vinson, 477 U.S. 57 (1986)..........  36
Moore v. Hughes Helicopters, Inc.,
708 F.2d 475 (9th Cir. 1983).....................................33, 44, 46
Mozee v. Jeffboat, Inc., 746 F.2d 365
(7th Cir. 1984)........................................................................... 34
Ml Healthy City Board of Education v. Doyle,
429 U.S. 274 (1977)................................................................. 42
Nashville Gas Co. v. Satty,
434 U.S. 136 (1977)............................................. 23, 36, 45, 47
New York Transit Authority v. Beazer,
440 U.S. 568 (1979)....................................... 27, 28, 35, 43, 47
NLRB v. Transportation Management Corp.,
462 U.S. 393 (1983)...........................................................45, 46
Patsy v. Florida Board of Regents, 457 U.S. 496 (1982) . . .  44
Paxton v. United National Bank, 688 F.2d 552 (8th Cir.
1982), cert denied, 460 U.S. 1083 (1983)............................ 27
Payne v. Travenol Laboratories, Inc., 673 F.2d 798 (5th Cir.
1982), cert denied, 459 U.S. 1038 (1982)............................  27
Pullman Standard v. Swint 456 U.S. 273 (1982)................ 30
Rowe v. General Motors Corp., 457 F.2d 348
(5th Cir. 1972).....................................................................32,33
Segarv. Smith, 738 F.2d 1249 (D.C. Cir. 1984),
cert denied, 471 U.S. 1115 (1985)............................ 33, 34, 46
Square D Co. v. Niagara Frontier Tariff Bureau, Inc.,
476 U.S. 409 (1986).................................................................44
Steelworkers v. Weber, 433 U.S. 193 (1979)........................ 27
Teamsters v. United States,
431 U.S. 324 (1977)........................ 24, 25, 26, 27, 28, passim
Texas Department of Community Affairs v. Burdine,
450 U.S. 248 (1981)...............................................35, 44, 45, 49
Trans World Airlines, Inc. v. Thurston,
469 U.S. I l l  (1985)...................................... 23, 30, 45, 46, 49

119



vii

Trout v. Lehman, 702 F.2d 1094 (D.C. Cir. 1983), 
vacated on other grounds, 465 U.S. 1056 (1984).............. 33
United States Postal Service Board of Governors
v. Aikens, 460 U.S. 711 (1983)................................... 22, 40
United States v. Nixon, 418 U.S. 683 (1974)..................... 42
Vuyanich v. Republic National Bank, 521 F. Supp. 656 
(N.D. Tex. 1981), vacated and remanded on other grounds, 
723 F.2d 1195 (5th Cir. 1984), cert denied,
469 U.S. 1073 (1984)...................................................... 46
Wambheim v. J.C. Penney Co., 705 F.2d 1492 (9th Cir.
1983), cert denied, 467 U.S. 1255 (1984).........................36
Wang v. Hoffman, 694 F.2d 1146 (9th Cir. 1982)....... 34, 35
Washington v. Davis, 426 U.S. 229 (1976)..................23, 47
Watson v. Fort Worth Bank and Trust,
108 S.Ct 2777 (1988)................... 25-26, 34, 39, 40, passim
Western Air Lines, Inc. v. Criswell, 472 U.S. 400
(1985)................................................................... 45, 46, 47
Williams v. Owens-Illinois, Inc., 665 F.2d 918
(9th Cir. 1982), cert denied, 459 U.S. 971 (1982)....... 29, 30

Statutes
42 U.S.C. §1981........................................................ 1,50
42 U.S.C. §2000e-2(a)(2)..............  21, 22, 23, 35-36, passim
42 U.S.C. § 2000e-2(j)............................................... 21,26
42 U.S.C. § 2000e-8(c) and (d)................................... 42, 43

Rules
Fed. R. Civ. Pro. 8(c)........................................................44
Fed. R. Evid. 301............................................................ 45

120



Texts
J. Chadboum, 9 Wigmore on Evidence §2486 (1981).......  45
E. Cleary, McCormick on Evidence §337 (3rd Ed. 1984)... 45 
D. Louisell and C. Mueller, 1 Federal Evidence,
§66 (1977)......................................................................  45
J. Moore 2A Federal Practice tJ 8.27(4) (1987)...........44, 45
J. Weinstein and M. Berger, 1 Weinstein’s Evidence,
(1988)......................................................   44
C. Wright and A. Miller, 9 Federal Practice and Procedure 
§2588 (1971)...............................................................31,37

Regulations
29CFR §1602 ........................................................... 29,42
29 CFR §1607 .......................................................... 29,42
29 CFR §1607 (1978)..................................................... 42
29 CFR §1608.................................................................. 40

Legislative Materials
118 Cong. Rec. 7166 (1972)............................................. 44
43 SR 38, 312 (1978)........................................................42
H.R. Rep. No. 92-238 (1971)...................................... 41,48
H.R. Rep. No. 92-415, p. 5 (1971)..............................41, 44

viii

121



1

STATEMENT OF THE CASE 
1. INTRODUCTION

This class action under Title VII, 42 U.S.C. § 2000e et seq., 
and 42 U.S.C. § 1981 challenges a pattern of racial segregation 
in jobs, housing and messing at three Alaska salmon canneries.1 
The employers are Wards Cove Packing Co., Inc. ( WCP ), 
which owns Wards Cove and Red Salmon canneries, and Castle 
& Cooke, Inc. (“BBS”), which owns Bumble Bee cannery. (App. 
Cert. 1:4-5.*) Because it involves migrant, seasonal work, the 
case has unique features.

First, the work force* * 3 of the Alaska salmon canning industry 
is—as is true of other migrant, seasonal industries—far more 
heavily non-white than the areas from which it is drawn. (J.A. 
90, 93-95, 103-4, 369, 372-73; see also Tr. 336-37, 344, 423, 
434-35, 483, 607.) For the eight decades spanning 1906-78, it 
has been 47-70% non-white. (App. Cert. 1:42.) While non­
whites in the industry in recent years have been largely of 
Filipino or Alaska Native descent, workers of Chinese, Japan­
ese and Mexican descent preceded them, but left the industry,

'The case originally encompassed two other canneries— namely, Ekuk and 
Alitak, which are run by W C P  and BBS as part of their Columbia Wards 
Fisheries (“C W F ”) joint venture. (App. Cert 1:4-6.) Title VII claims against 
C W F  were dismissed on procedural grounds. A to n io  v. W a r d s  C o v e  P a c k in g  
C o ., 703 F.2d 329 (9th Cir. 1983). Title VII claims against W C P  and BBS for 
their role in the C W F  venture were dismissed on the ground they were outside 
the scope of the E E O C  charges (App. Cert 1:95-96,111:14, VIII:2), although 
the charges allege “each discriminates throughout its Alaska operations in 
Alitak... and other facilities.” (Ex. 1-3,5-10; R.P.0.132.) The charges were 
later amended to clarify they cover W C P  and B B S  as joint venturers. (Ex. 31- 
35; R.P.O. 132, 134.) The district court dismissed 42 U.S.C. § 1981 claims 
and the court of appeals affirmed. (App. Cert. 1:96-97,1:129-30,111:15-43, 
VI: 16.) This Court declined to grant certiorari on issues which affect claims 
involving the C W F  canneries. (No. 87-1388.)

*“ App. Cert” refers to the appendix to the petition for certiorari, “J.A." to 
the joint appendix, “E.R.” to the excerpt of record, “R.P.O.” to the revised 
pretrial order and “Tr.” to the transcript of proceedings.

3“Work force” refers to those employed by an employer or an industry. 
(See J.A. 90-91,369-70.) “Labor force” or “labor supply" refers to individuals 
employed in and rejected applicants for work in an industry. (See ib id .)  
"Labor market” refers to areas from which workers are or could be hired. See 
H a z e lw o o d  S c h o o l D is tr ic t  v. U n ite d  S ta te s , 433 U.S. 299, 308, 310-12 (1977).

122



2

in part because of changes in immigration laws. (Ex. 625; Tr. 
345-46, 433, 771, 775; App. Cert. 1:42.)

Second, because of the migrant, seasonal nature of the 
work, WCP and BBS provide bunkhouses as well as meals to 
employees (App. Cert 1:17), so patterns of segregation extend 
beyond jobs into several layers of an employee’s life (see p. 19- 
21, infra). Whether because of the extent of the segregation or for 
other reasons, there is “pervasive” race labelling (App. Cert. 
VI:33) of jobs, bunkhouses and messhalls, with terms such as 
“Native Crew,” “Filipino cannery worker,” “Phillipine jsicj 
Bunkhouse,” “Native Gallej Jy Cook” and “Filipino Mess” in 
use. (App. Cert 1:76-77.) Even the salmon butchering machine 
has a name with racial overtones, the “Iron Chink.” (App. Cert. 
Vl:33; see also id. at 1:22.)

Third, because the industry is seasonal, workers often have 
other pursuits during the rest of the year, so their jobs with 
WCP or BBS do not necessarily reflect the full measure of their 
skills. Of the ten named plaintiffs, seven had some college when 
they worked in the canneries. (J.A. 38, 52; Tr. 951-52, 1036, 
1050, 2214-15; Dep. Viemes-1975 3-4.4) One was a structural 
engineer when he sought but was denied an upper-level job. 
(App. Cert. 1:86.) Others later became architects (Tr. 951, 
2214), mechanics (Tr. 869-70, 872; 2061(a)), a captain in the 
Air Force (Tr. 2215) and a graduate student in public admin­
istration (Tr. 76). Several were students when they worked for 
WCP or BBS, but they held menial jobs, while white students 
frequently held choice jobs. (J.A. 78,114,118; Tr. 1010,1320- 
21,1373,2534-35,2838-39,2926-27,3315.) Whites were hired 
as deckhands as early as age 14,17 and 18 (J.A. 196-97,200; Tr. 
2757-58), as fishing boat crew members as early as age 15 (Dep. 
Aiello 30), on the beach gang as early as age 16 (Tr. 2652-53) 
and as machinist helpers as early as age 16 (J.A. 133-34). Since 
they live in coastal regions, residents of the Alaska Native vil­
lages where WCP and BBS recruit often have vast boating and 
fishing experience, which more than qualifies them for almost 
any tender or fishing job. (See J.A. 414,418-19; cf. id. at362-63.)

4A  number of witnesses testified by deposition, which were admitted in 
evidence. (Tr. 2291.)

123



3
Some non-white employees have been workine in the canneries 
for decades without promotion (Tr. 953, 967), but could have 
acquired skills for other jobs.

2. RACIAL STRATIFICATION IN JOBS
The court of appeals accepted statistics showing racial 

stratification in jobs as proof of disparate impact, regardless of 
the racial mix of the labor supply. (App. Cert. VI:14-17.) Be­
cause the subjective criteria on which WCP and BBS rely have 
a disparate impact on non-whites, the court of appeals did not 
require the employees to offer statistics differentiated by quali­
fications. (Id. at VI: 17, VI:24-27.)

While the degree varies, the administrative, machinist, fish­
erman, tender, carpenter, beach gang, clerical, quality control 
and miscellaneous departments are all white or heavily white. 
(Ex. 588-90 (E.R. 35-37); Tr. 2231, 2261.) By contrast, the 
largest department—cannery worker—is heavily non-white. 
(Ibid) At one cannery, the laborer department is also heavily 
non-white. (Ex. 589 (E.R. 36); Tr. 2231,2261.) For the Court’s 
convenience, work force statistics offered by each side are 
contained in Appendices A and B.®

A. Job Departments
Cannery machinists operate and maintain machinery, but 

despite the similarity in name, are not true machinists as the 
term is used in the Lower 48. (J.A. 399, 541-42.) They are 
apparently called machinists because they are represented by 
the machinists union. (Ibid.) Their supervisors characterized 
them—in interviews with the WCP and BBS expert on qual­
ifications—as machine operators almost as frequently as they 
described them as craft workers. (Tr. 2955.) The machinist 
crew is a small one, supervised by two skilled foremen—the 
cannery foreman and first machinist—who oversee all major *

*The main difference between each side's statistics in Appendices A and B 
is in the way vacancies are counted. The employees counted year-round 
employees once at the initial point of hire in the job in question and seasonal 
employees each season they were hired or re-hired. (Tr. 2234-36.) The e m ­
ployers omitted year-round employees, counted seasonal employees only in 
the year they were first hired in a job and eliminated seasonal employees who 
worked some of the winter months at Lake Union Terminals, a C W F  sub­
sidiary. (J.A. 261-62.)

124



4

and many minor repairs,6 a division of duties which reduces the 
need for skills among the machinists themselves. Machine over­
hauls are done in the winter in Seattle rather than at the cannery 
by the machinists. (J.A. 709; Dep. Snyder 34; Dep. Mullis 32; 
Dep. Jorgensen 28; Dep. Rohrer 11-12.) While not always 
available, manufacturer’s representatives assist with some 
machine repairs and set-ups. (Tr. 239, 3062-63; Dep. Snyder 
34-36; Dep. Rohrer 10-11; Dep. Jorgensen 27-29; Dep. Mullis 
32-34.)

The quality control job is often filled by a college student, 
who receives on-the-job orientation in conducting the necessary 
tests. (See J.A. 78-79.)

Fishing boats in Bristol Bay—the only locale where WCP 
and BBS have employee fishermen (J.A. 179)—are small vessels 
which are limited by Alaska law to 32’ in length. They are 
staffed by two people, a captain and a puller. (J.A. 180-81; Tr. 
902(c).)

Tenders bring the fish from the fishing boats to the cannery 
grounds. (Tr. 1144.) They have crews usually of four (ibid.), so 
the tender cook prepares family-size meals (Tr. 124,2384-85), 
although sometimes in Bristol Bay the cook also feeds fisher­
men (J.A. 21). Lacking brine refrigeration, dry tenders gener­
ally stay less than a day from shore, since salmon must be 
processed within 48 hours of catch. (Tr. 1144; App. Cert. 1:21.) 
Because they can chill the catch, brine tenders make longer 
voyages. (Ibid; Tr. 1144.) The major repair work on tenders is 
done at the CWF shipyard in Seattle before the season rather 
than in Alaska by the tender crews. (Tr. 2385.) Once at the 
cannery, port engineers help tender engineers with some repairs. 
(J.A. 124; Tr. 123; Dep. Milholland 12; Dep. Rohrer 24; Dep. 
Jorgensen 25.) There are no licensing requirements for tender 
jobs. (Dep. Leonardo-1978 14; Dep. J. Brindle 18-19.)

Beach gang involves largely laborer work (cf. Tr. 1514; App. 
Cert. 1:66-67) for which unskilled personnel have been hired. 
(e.g. Tr. 1546; Dep. Sifferman-1980 30, 33-34, 41).

B. The Statistics
Each side’s statistics show between six and seven upper- *

*J.A. 119, 542-43; Tr. 239, 708-09, 3062-64, 3267; Dep. Snyder 15, 23; 
Dep. Rohrer 4-7; Dep. Jorgensen 18-19; Dep. Mulli9 14-15,22-23,26-27,29; 
Dep. Landry 24-25.



5

level departments at Bumble Bee during 1971-80 were at least 
90% white, although the cannery worker department was be­
tween 52% and 59% non-white. (Ex. 588 (E.R. 35); Tr. 2231, 
2261; Ex. A-278 Table 4 SN (E.R. 4); Tr. 2646-47; App. A-l, B- 
1.) During this period there were 741 hires counting re-hires 
and 335 hires not counting re-hires in departments which were 
at least 90% white. {Ibid.)

Each side’s statistics show five upper-level departments at 
Red Salmon during 1971-80 were at least 94% white, although 
the cannery worker department was between 64% and 70% 
non-white. (Ex. 589 (E.R. 36); Tr. 2231,2261; Ex. A-278 Table 
4 RS (E.R. 3); Tr. 2646-47; App. A-2, B-2.) During this period, 
there were 384 hires counting re-hires and 152 hires not counting 
re-hires in departments which were at least 94% white. {Ibid)

Each side’s statistics show between four and six upper-level 
departments at Wards Cove during 1971-80 were at least 93% 
white, although the cannery worker department was between 
31% and 37% non-white. (Ex. 590 (E.R. 37); Tr. 2231, 2261; 
Ex. A-278 Table 4 WC (E.R. 2); Tr. 2646-47; App. A-3, B-3.) 
During this period, there were 612 hires counting re-hires and 
227 not counting re-hires in departments which were at least 
93% white. {Ibid.)

Even departmental figures do not tell the whole story. Fully 
61 of 95 job titles filled more than once at Bumble Bee were at 
least 90% white or 90% non-white during 1971-80. (Ex. 598 
(E.R. 38-44); Tr. 2231,2261.) During the same period, fully 62 
of 93 job titles filled more than once at Red Salmon were at least 
90% white or 90% non-white. (Ex. 599 (E.R. 45-51); Tr. 2231, 
2261.) Similarly, at Wards Cove during 1971-80, fully 54 of 72 
job titles filled more than once were at least 90% white or 90% 
non-white. (Ex. 600 (E.R. 52-56); Tr. 2231, 2261.)

The wage disparities between the upper-level and lower- 
level jobs are extreme, with upper-level jobs often paying three 
or four times as much as lower-level jobs for a season only about 
a month longer. (Ex. 598-600 (E.R. 38-56); Tr. 2231, 2261; 
App. Cert. VI: 17.)

The recent job segregation reflects a long standing pattern 
both at WCP and BBS canneries and industry-wide. From 
1949, when WCP purchased the cannery, until 1972, the ma­
chinists, tendermen, storekeepers and clerical workers at Red

126



6

Salmon were 100% white, while cannery workers were heavily 
non-white. (J.A. 151-52, 154.) For the same quarter century, 
some Alaska Natives but no Filipinos, Chinese, Japanese or 
blacks worked on the beach gang or as company fishermen. 
{Ibid.) Similarly, industry-wide statistics show tender jobs were 
between 90% and 99% white each year during 1907-39 and 1941- 
55, although the industry as a whole was between 47% and 70% 
non-white. (Ex. 637 (E.R. 21-34); Tr. 771, 776; App. Cert. 1:42.)

3. RACE LABELLING OF JOBS
Race labelling at the highest levels of management enforces 

the racial identifiability of jobs and departments. (See App. 
Cert. VI:33; see also id. at 1:76-79.) While a more extensive 
recap is contained in Appendix C, the employees summarize 
some of it here.

Far from condemning race labelling, Alec Brindle—WCP’s 
current president—testified cannery workers are called “the 
Native crew” for “mere ease or habit of identification,” since 
“|o)ne would normally assume, if you recruited from a Native or 
Eskimo village, the people who came from there... would often 
be referred to in that manner . . . (J.A. 156-57, 182-83.)

A.W. Brindle, who until 1977 was president of WCP and 
superintendent of Red Salmon, referred from 1970 on to res­
ident cannery workers® as “Eskimo labor,” “these Eskimos,” 
“Eskimo males,” "Young native boys,” "those natives,” “Eski­
mos,” “those Natives” and "the Eskimos” (App. Cert. 1:24, 
1:28; Ex. 245,254,397,452,721,749-50; R.P.0.132,145,154; 
Tr. 2279); non-resident cannery workers as “the Filipinos” (Ex. 
484, 497; Tr. 2279); other employees as “the four natives that 
work with Vem” (Ex. 376; R.P.0.132,153); salmon butchering 
machines as “chinks” and the operator as the "chink man” (Ex. 
289; R.P.0.132,147); the flight carrying non-resident cannery 
workers as the “Filipino Charter Flight” (Ex. 502; Tr. 2279); 
and Local 37, ILWU as the “Filipino Union” (Ex. 328,508; Tr. 
311, 313, 2279). He also wrote,

7The “transporters" in Exhibit 63 are tendermen who bring fish from the 
fishing grounds to the cannery. (S e e Tr. 560.)

®"Resident" cannery workers are those who normally reside in Alaska and 
“non-resident” cannery workers those who normally reside in the Lower 48. 
(App. Cert. 1:29-30.)

127



7

ITJhese Eskimos are completely impossible. We 
have had nothing but trouble and we probably had less
trouble than the majority__ There is no question in my
mind that the Eskimo labor is going to be less desirable as 
time goes on and actually it will be a detriment. The 
trouble comes pretty much from these younger ones that 
have gone to college.

(App. Cert. 1:79; Ex. 452; Tr. 2279.)
Warner Leonardo, superintendent of Bumble Bee, referred 

from 1970 on to non-resident cannery workers as the “Filipino 
cannery crew,” “21 Filipino” and the “Filipinos” (J.A. 216; Ex. 
294, 407, 414; R.P.O. 132, 148, 155); classified employees as 
“Women cannery workers,” “Filipino cannery workers,” “Na­
tive cannery workers,” “Japanese,” “Filipinos,” “Natives” and 
“Native Galley Cook” (Ex. 327, 342-50; R.P.O. 132, 150-51); 
and called cannery worker aign-on pay “Filipino sign-on pay” 
(Ex. 414; R.P.O. 132,155). Badge assignments at Bumble Bee 
include, “09-525 thru 09-574 Filipinos” and “09-575 thru 09- 
659 Natives”. (App. Cert. 1:77-78.)

Even laundry bags and mail slots are marked with racial 
designations like “Oriental Bunkhouse.” (App. Cert. 1:78.)

4. SEGREGATED HIRING CHANNELS
The district court did not directly address use of essentially 

segregated hiring channels, but the court of appeals found the 
practice had an obvious disparate impact for which—on the 
findings made—there was no business necessity. (App. Cert. 
VI: 27-32.)

WCP and BBS solicit for low-paying cannery worker jobs in 
Alaska Native villages, such as Tuluksak, Kwethluk and Napas- 
iak, which are 96-99% Alaska Native.9 Typically, a bush pilot or 
village leader lines up the workers at the direction of the home 
office or the cannery superintendent. The practice cuts village 
residents off from the more desirable jobs, which are filled in 
the Lower 48 through different channels. (See p. 9-10, infra.) 
WCP and BBS hire laborers from heavily Alaska Native areas 
immediately around the canneries. (App. Cert. 1:32,1:38.) *

*App. Cert. Vl:28; Ex. 480; Tr. 2026; Ex. A-382;Tr. 1390-31,1433; J.A. 
4-5; Tr. 637-40,1125,2527-28; Dep. Leonardo-1975 41-42; Dep. Leonardo- 
1978 7-8; s e e  a ls o Tr. 2439; Dep. Ekem-1978 10-11; Dep. W.F. Brindle-1978 
17-18.

128



8

WCP and BBS also recruit for low-paying cannery worker 
jobs through cannery worker foremen of Asian descent and Local 
37, ILWU, a union whose membership is largely Filipino. (App. 
Cert. 1:36, VI:38; J.A. 3-6, 223, 645-46; Tr. 2923-25, 3120-21, 
3136.) The parties stipulated,

The majority of non-resident cannery workers are lined up 
by the cannery worker foremen after management has 
estimated the number that will be needed.

(J.A. 3.) The foremen are management, for the Local 37 labor 
contract provides they “ .. .shall be selected by the Company... 
(and bej representative[sj of and responsible to the Company. ” 
(Ex. A-l through A-ll, Local 37, ILWU; Tr. 2345-46; see also 
id. at 3136) (emphasis added). The racial impact of this recruit­
ment is evident in a letter from a foreman to the Alitak super­
intendent.10

Yes, the entire crew will be Asians, unless Local #37 slips 
a ‘stray’ in there. However, anytime you want some more 
whites or blacks, just let me know as I can recruit some 
good ones, I believe.

(Ex. 394; Tr. 3121, 3140-41.)
The upshot is every cannery worker hired under the con­

tract was initially selected by the company, for the union has no 
formal role in selecting employees. Under the contract, first 
preference goes to past company employees at the same can­
nery, second preference to past company employees at a dif­
ferent cannery and third preference to,

Persons satisfactory to the Company, including but not 
limited to Union members or men recruited for employ­
ment by the Union.

(Ex. A-l through A-ll, Local 37, ILWU (e.g. E.R. 1); Tr. 2345- 
46.) (Emphasis added.) A number of cannery workers testified 
they were hired directly by the foreman. (J.A. 53,817-18,926.) 
Local 37 does not have an exclusive hiring hall, for its contract 
simply provides,

Persons selected and employed by the Company shall 
register with the Union, or their names shall be furnished

,0Evidence of policies at Ekuk and Alitak is relevant, for whde they are no 
longer covered by the case, their policies were set by W C P  and BBS. The 
district court found W C P  and B B S  "operated the (CWF joint) venture jointly 
and equally" and set its “hiring policies, firing policies, promotion policies 
and employee regulations.” (App. Cert. 1:5,1:26.)

129



9

to the Union by the Company prior to leaving port of 
embarkation.

(Ex. A-l through A-ll, Local 37, ILWU; Tr. 2345-46.) Some 
nonresident cannery workers—particularly women, who are 
nearly all white—are hired at company offices with no contact 
with the union. (R.P.O. 17; J.A. 4-6, 459-62; Tr. 646-47, 695, 
939-40, 2593, 2600, 2935, 2939.)

Hiring channels for cannery workers are clearly isolated, for 
as WCP’s president testified, “None of the cannery worker 
foremen . . .  is vested with authority to hire for any position 
outside the cannery worker department” or "to even discuss 
those jobs on behalf of management.” (J.A. 156-57, 163.) Nor 
are the bush pilots who recruit in Alaska Native villages. (Tr. 
2527-28.)

Word-of-mouth recruitment is the norm in filling upper- 
level jobs, a fact freely conceded by nearly all management 
witnesses." Nearly all the people who recruit this way for 
upper-level jobs are white. (See Ex. 598-60 (E.R. 38-56); Tr. 
2231,2261.)

The Bumble Bee cannery superintendent acknowledged a 
"preference (in all upper-level jobs) for people who have been 
recruited over people who have applied on a walk-in basis.” 
(J.A. 216, 229-30; see also id. at 222-24.) WCP’s president 
testified—as in essence did Bumble Bee’s cannery superin­
tendent—fish boat captains have “complete latitude in hiring” 
crew members. (J.A. 156, 180-81; see also id. 460; Tr. 1514, 
3163.) One cannery superintendent wrote an applicant,

It is pretty much the universal practice that each captain
selects crew members__ It is not simple, of course, to find
captains who are looking for crew men as they usually have 
relations or friends they think of first when an opening comes 
up.

(Ex. 464; Tr. 2537-38; see also Ex. 465; Tr. 2279.) (Emphasis 
added.) Tender captains—who are given discretion in hiring 
because they live in close quarters with their crews—often 
select friends and relatives. (Tr. 631, 1141, 1374, 2909; Dep. 
Leonardo-1978 13; Dep. Ekem-1978 15; Ex. 603-605 (E.R.

"App. Cert Vl:38; R.P.O. 16-18; J.A. 13-15,222-24,229-30; Tr. 627-37, 
1146, 2772; Dep. Gilbert-1975 98; Dep. Snyder 4-5; Dep. Lessley 7-8; Dep. 
Leonardo-1978 13; Dep. A.W. Brindle-1975 27-29, 77; Dep. Landry 4, 12 .

130



10

65-101), 608-10 (E.R 102-104); Tr. 2231, 2237, 2244-46, 
2261.) WCP’s president acknowledged, “the machinists’ fore­
man would generally select his own crew, just like the skipper of 
a cannery tender selects his crew.” (Dep. A.W. Brindle-1975 
28.)

WCP and BBS do not publicize vacancies in upper-level 
jobs. (App. Cert. 1:28-29.) Nor do they generally require written 
applications for upper-level jobs. (J.A. 225; Tr. 1316-17,2917- 
19; Dep. Leonardo-197 5 26; Dep. Aiello 32; Dep. A.W. Brindle- 
1975 77; see also Dep. Bozanich 24.)

The court of appeals commented the disparate impact of 
segregated hiring channels is “obvious.” (App. Cert. VI:28.) 
During 1971-80, 90% of non-whites at Bumble Bee, 80% of 
non-whites at Red Salmon and 90% of non-whites at Wards 
Cove were hired in cannery worker or laborer jobs, while only 
40% of whites at Bumble Bee, 22% of whites at Red Salmon and 
58% of whites at Wards Cove were. (Ex. 588-90 (E.R. 38-56); 
Tr. 2231, 2261.) The district court found these imbalances 
were caused by: (1) tapping Local 37, a union with a "predom- 
inently (sicj Filipino” membership; and (2) hiring from villages 
near the canneries, where "Alaska Natives comprise a high per­
centage of the . . .  local labor market.” (App. Cert. 1:32,1:36- 
38.)

5. NEPOTISM IN UPPER-LEVEL JOBS
The district court made confusing findings on nepotism, 

citing the "pervasive incidence of nepotism,” "the nepotism 
which is present in the at-issue jobs” and "the strength of the 
nepotism evidence,” but writing "the nepotism . . . does not 
exist because of a ‘preference’ for relatives.” (App. Cert. 1:103, 
1:105,1:114.) The court of appeals reversed, noting, "If nepo­
tism exists, it is by definition a practice of giving preference to 
relatives.” (Id. at VI:21.)

Fully 345 of 347 nepotistic hires during 1970-75 at WCP 
and BBS in eight upper-level departments were of whites,11 
including 129 hires in tender jobs, 93 in fishing jobs, 67 in 
machinist jobs, 24 in clerical jobs, 15 in beach gang jobs, 9 in 
carpenter jobs, 6 in quality control jobs and 4 in administrative *

^The Court of Appeals noted 349 nepotistic hires in four departments at 
five canneries originally covered by the case. (See App. Cert. VI:21.)

131



11

jobs. (Ex. 608-10 (E.R. 102-104); Tr. 2231, 2261.) WCP’s 
president tried to justify hiring relatives as “a reasonable bus­
iness practice,” saying “it is the incentive to get the relative 
back to Alaska or is part of an economic package which makes 
working for our company more attractive.” (J.A. 184.)

8. LACK OF OBJECTIVE OR 
DISCERNIBLE QUALIFICATIONS

The district court commented on the “general lack of objec­
tive qualifications” (App. Cert. 1:106; see also id. at IV-.23), 
finding in essence there were no fixed criteria since “(qjualifi- 
cations for any individual position depend to a certain extent on 
. . . the age and condition of equipment, skill level of other 
incumbents and supervisors, and other such factors” {id. at 
1:46). The expert WCP and BBS called on qualifications con­
curred, saying “jobs are often structured around the skills of the 
people who are available to fill them, rather than the other way 
around.” (Tr. 2941, 3000.)

Bumble Bee’s cannery superintendent acknowledged there
were no established qualifications, so he “just rel(ied] on [his] own 
judgment and the judgment of the foreman who (wasj hiring.” 
(Dep. Leonardo-1978 2, 46-47; see also Tr. 2617, 2642.) A 
home office employee who recruited in nearly all upper-level 
jobs acknowledged “there were no set qualifications a person 
had to meet” (J.A. 105-11.) WCP’s president testified quali­
fications for tender captain, engineer and—to some degree— 
deckhand jobs are so fluid they vary from boat to boat. (J.A. 
210-12; see also Dep. E. Sifferman-1980 9.) A fish boss testified 
"|t]here were no qualifications” for fishing boat crews, since 
each captain decided what—if any—criteria to impose. (J.A. 
582; see App. Cert. 1:72.) Electing to “take calculated risks,” 
Bumble Bee’s cannery superintendent testified he waives what 
might be considered minimum qualifications. (J.A. 463-67.)

WCP and BBS were unable throughout fourteen years of 
litigation to identify the qualifications they actually applied. 
While the district court made findings on qualifications which 
could be “reasonably required” (App. Cert. 1:58), it is clear the 
qualifications were never actually imposed. The court of appeals 
observed the district court “did not.. .  find that these specific 
criteria were actually applied,” “(tjhere is anecdotal evidence

132



12

which suggests that these criteria were not applied” and the 
district court "must make findings as to the job-relatedness of 
criteria actually applied.” (Id. at VI:23, VI:25, VI:27.)

On May 28, 1974, the employees served interrogatories 
asking, "what qualifications (WCP and BBS| required for [each] 
job... including... what prior work experience if any and what 
special training if any were required . . .  (First Interrogs. to 
Defs. 17.) The answers do not give qualifications as they were 
applied. One cannery superintendent acknowledged they were 
his “ideal for qualifications,” rather than "qualifications as they 
were actually imposed at Alitak from 1970 onward.” (J.A. 463, 
468-69.)

WCP and BBS then offered an entirely different set of 
qualifications at trial, namely, one which their expert—Larry 
DeFrance—believed could be “reasonably required” (J.A. 471, 
499). The district court adopted DeFrance’s hypothetical qual­
ifications verbatim (App. Cert. 1:58-71; J.A. 499-508), even 
though they had never actually been applied.13

THE COURT: All right Mr. DeFrance, in this case, I 
believe you have already testified that the Defendants 
have not adopted, to your knowledge, the minimum qual- 
ifications that you recommended; is that correct?
THE WITNESS: That’s correct. I don’t know that they 
have ever been adopted.

(J.A. 574-75; see also id. at 545.) WCP and BBS openly stip­
ulated some employees in upper-level jobs could not meet 
DeFrance’s suggested qualifications.14 (Tr. 3076-79.) Even so, 
many are indefinable, for they are phrased as "ability to’5 per­

,3The district court also found it takes "extensive experience” or "sub­
stantial prior skill and experience” to perform a number of jobs (App. Cert. 
L56-57), but did not say what the experience or skill was or what qualifica­
tions were actually applied. The employees acknowledge some— but not all—  
of these jobs require special skills, but they can generally be acquired in entry- 
level jobs at the cannery. (S e e p. 13-14, in fra .) The skills necessary for other 
jobs are by no means apparent The shop machinist job has been filled by a 
white who took a night school course, but never served an apprenticeship or 
worked as a shop machinist (Dep. Rohrer 22-23.) The port engineer job—  
which entails repairing tenders and fishing boats— can be filled by one whose 
only mechanical experience is working on his or her own car. (Id . at 25; s e e  a ls o  
Dep. Snyder 43; Dep. Mullis 37.)

,4Lacldng information on the skills or background of their own employees,

133



13

form instead of the standards by which such ability is measured, 
while others are highly subjective. These features of DeFrance’s 
hypothetical qualifications are summarized in Appendix D. 
Faced with this, the employees offered anecdotal evidence of 
employees hired on far more modest qualifications than 
DeFrance’s.18

WCP and BBS also called lay witnesses on qualifications, 
but since they almost all lacked hiring authority, they could not 
say what qualifications were actually imposed. (Tr. 1339,2357, 
2548, 2569, 2617, 2632, 2642, 2742, 2842, 2848, 2884, 3172, 
3267,3272.) Of the qualifications lay witnesses cited, many are 
in any case entirely subjective. They also are summarized in 
Appendix D.

Whites advanced from entry-level jobs to the more difficult 
jobs at the cannery.16 A cannery superintendent’s nephew rose 
from machinist helper—an unskilled job (App. Cert. 1:107- 
08)—at age 18 to seamer machinist at age 19 to salmon butch­
ering machinist at age 20 to first machinist at age 21, all while he 
attended college during the winter months. (J.A. 114-22; Tr. 
705-10, 770.) White relatives of management progressed to 
tender captain from tender deckhand, another unskilled job 
(App. Cert 1:107-08), some starting as early as age 14 (Tr.

W C P  and BBS retained a firm to interview some of them in 1980. (See Tr. 
1144 e t  s e q .) From the interviews, DeFrance could say only 80%— 155 of 
193— of employees in upper-level jobs would have survived even a first cut or 
prescreening based on his qualifications when first hired from 1971 on. (J.A. 
525.) DeFrance did not require they meet the hypothetical qualifications 
when first hired before 1971 as long as they later acquired the relevant 
experience. (J.A. 672-73.)

,5One white dry tender engineer who was related to a C W F  home office 
employee "had no mechanical experience or training other than performing 
preventative maintenance on his car, and no experience working on a boat” 
(App. Cert VT.26; s e e  a ls o J.A. 20-21,22-24; Dep. Millholland 3.) Other white 
tender engineers had a similar lack of background. (J.A. 60-62; s e e  a ls o  id . at 
123-24,131-35.) Two whites— one a cannery worker and the other a stock- 
room clerk— were promoted to machinist jobs without any such experience. 
(J.A. 25-29, 30-31, 34-37; s e e  a ls o Dep. Landry 15-17.) Other machinists 
could not meet DeFrance's proposed qualifications when first hired. (J.A. 
144-45, 600; Tr. 2534-35.)

,6Even under DeFrance’s hypothetical qualifications, upper-level jobs—  
such as machinist— can be learned largely or exclusively through experience 
in positions the district court found were unskilled. (J.A. 499,505; App. Cert. 
L107-08.)

134



14

1319-20; see also J.A. 131-36; see also Dep. Sifferman-1978 
20). The parties stipulated a fishing boat captain’s job can be 
learned through prior experience as a fishing boat partner. (J.A. 7.)

The effects of the regime of subjectivity are apparent from 
the same statistics as word-of-mouth recruitment. WCP and 
BBS do not require written applications {see p. 10, supra), so 
there are no systematic applicant flow rfecords. Bumble Bee 
and Red Salmon destroyed applications through 1977 and 
Wards Cove through 1979. (Tr. 1143,1518,2718,2773.) Even 
when they retained applications, WCP and BBS rarely race- 
identified applicants, so applicant flow figures are unreliable. 
(Tr. 1403; see p. 18 n.20, infra.) Given this, one cannot separate 
the impact of subjective qualifications from the impact of the 
recruitment process.”

7. RE-HIRING PAST EMPLOYEES 
IN THEIR OLD DEPARTMENTS

WCP and BBS have a practice of re-hiring past incumbents 
in their old departments. (App. Cert. 1:29, VI:32.) For union 
jobs, the practice is memorialized in re-hire preferences in 
collective bargaining agreements. {Ibid.) The district court held 
the practice did not have a disparate impact, reasoning it could 
not perpetuate past discrimination unless such discrimination 
were first proved. {Id. at 120-21.) But it also made a hypo­
thetical or alternate finding of business necessity. {Id. at 122.) 
The court of appeals held the practice in fact had a disparate 
impact since “(wjhen jobs are racially stratified, giving rehire 
preference to former employees tends to perpetuate existing 
stratification.” {Id. at VI:32.) But it affirmed the finding of 
business necessity {id. at 33), despite the absence of any sup­
porting evidence.

” While never imposed, DeFrance’s requirements— even as a proxy for 
actual criteria— have a disparate impact on non-whites. W C P  and BBS hiring 
area statistics which incorporate his views on skills show the percent non­
white who meet the qualifications is lower than the percent non-white in the 
general labor force in the hiring areas in every at-issue job family except 
fishermen and culinary. (Ex. A-278 Tables 1 and 3 WC, RS and SN; Tr. 1929- 
32, 2231, 2281.) Similarly, DeFrance's review of 1980 employee interviews 
shows the pass rate of non-white employees under his hypothetical qual­
ifications is less than 80% of the pass rate for whites. (Tr. 1985-88.)

135



15

8. INDIVIDUAL INSTANCES OF DISCRIMINATION
Twenty-two non-white9 testified they applied unsuccessfully 

for or were deterred from seeking upper-level jobs. {See App. 
Cert. 1:84-94.) Many had special skills or rose rapidly with other 
employers, but were confined to menial jobs at the canneries. 
{See p. 2, supra; see also J.A. 64; Tr. 806.) The district court 
did not find any unqualified for the jobs they sought or were de­
terred from seeking. (App. Cert. 1:84-94, VI:30.)

Of the 22,12 applied orally, in writing or both (App. Cert. 
1:84-90), but the district court found they applied too early, too 
late or to the wrong person {id. at 1:86,1:88-89,1:115-17). When 
one asked a machinist foreman for work as a machinist, he was 
asked in turn, “What’s wrong with being on the Filipino crew?” 
(J.A. 52, 56-57.) Others were deterred by foremen who told 
them they “had to know someone” to be hired as a machinist 
(J.A. 77), advised them “not to make waves” by seeking promo­
tions (J.A. 7 6), refused to tell them how to seek upper-level jobs 
(J.A. 85-86) and said Filipinos "were not supposed to have” 
upper-level jobs (Tr. 832). Yet others were deterred by segre­
gation in jobs, housing and messing. (App. Cert. 1:92-93; Tr. 
282, 294, 872-73, 953, 967-68, 1037, 1051-52.) The district 
court gave little weight to evidence of deterrence, since it be­
lieved “the test for... discrimination is whether a defendant in 
fact discriminates, and not whether class members subjectively 
believe a defendant discriminates.” (App. Cert’ 1:117.) The 
court of appeals held findings on individual claims were pre­
mature until liability issues were resolved. {Id. at VI:41.) Be­
yond noting the deterrent effect of race-labelling, segregation 
in housing and segregation in messing {id. at 33), it held infor­
mal, discriminatory hiring practices “should serve to excuse the 
cannery workers from the necessity of establishing the time­
liness of their applications and automatically elevate oral in­
quiries to the status of applications” {id. at 41-42).

9. THE LABOR SUPPLY
The district court found the percentage of non-whites work­

ing in the industry during 1906-39 and 1941-55 “ha9 histor­
ically been from about 47% to 70%,” “(tjoward the end of this 
period it stbilized (sicj at about 47% to 50%” and a sample of 
about half the industry showed it to be 48% non-white during

136



16

1970-78. (App. Cert. 1:42.)
Since the Census is dominated by people unwilling to take 

migrant, seasonal work, Drs. Robert Flanagan and Shirley Smith 
—a labor economist and demographer, respectively—found 
industry statistics to be the best available measure of the racial 
mix of the labor supply here. (J.A. 90,369,373-78; Tr. 370-73, 
571-77.) From these statistics, they concluded the labor supply 
is about 47% non-white. (J.A. 378-79; Tr. 353,370.) This figure 
matches closely: (1) the 47% non-white in all Alaska fruit, 
vegetable and seafood processing industries as shown by the 
1970 Census (Ex. 626; Tr. 347-48, 776); and (2) the 42% non­
white in the work forces during 1971-80 of the five canneries 
covered by this case (Ex. 588-92; Tr. 2231, 2261).

Using the 47% non-white figure, Dr. Flanagan found a pat­
tern of race segregation, with non-whites significantly under­
represented in upper-level jobs by margins of two or more 
standard deviations. (J.A. 379-90; Ex. 634-36 (E.R. 122,124- 
25); Tr. 2278.) Skills adjustments were hampered by: (1) the 
absence of statistics by race on people qualified for the more 
skilledjobsinthe industry (J.A. 100-02; see also id. at 373-78); 
and (2) the lack of objective qualifications at WCP and BBS 
which would enable one to use such statistics even if they 
existed (id. at 384). Dr. Flanagan compensated for these diffi­
culties by making the extremely conservative assumption WCP 
and BBS hired every available non-white qualified for truly 
skilled jobs, basing his computations largely instead on jobs 
without significant skill barriers. (Id. at 384-86.)

By contrast, Dr. Albert Rees—a labor economist for WCP 
and BBS—concluded the labor supply is about 10% non-white, 
only about a fifth of the actual percent non-white in the industry 
since the turn of the century. (J.A. 250,292.) The district court 
accepted this view, holding it rebutted the prima facie case of 
disparate treatment based on work force statistics. (App. Cert. 
1:41-43,1:118-19,1:124.) The court of appeals did not rule on 
the labor supply question, which it believed was relevant to 
disparate treatment but not disparate impact claims. (Id. at 
VL15-19.)

Faced with a work force which is about 42% non-white, 
WCP and BBS could justify job segregation only by arguing 
they hired too many non-whites in the menial jobs rather than

137



17

too few in the choice jobs. The centerpiece of their approach 
was a Census-based study, which bears two unusual features.

First, the study shows an absence of non-whites at statis­
tically significant levels in several departments, which led Dr. 
Rees to “conclude that [at WCPJ non-whites are significantly 
under-represented in hiring in the tender job family” (J.A. 269- 
70); “ftjhe under-representation of nonwhites" in the tender job 
family at Red Salmon is "statistically significant at the 5% level” 
(J.A. 271); and “non-whites were significantly under-represented 
in hiring in the fisherman and machinist job family at South 
Naknek [Le., Bumble BeeJ" (J.A. 274; see also id. at 289-90). 
(Emphasis added.)

Second, the study’s non-white availability figure is so low it 
suggests a striking under-representation of whites in the menial 
jobs.11 (J.A. 347; see also id. at 356-57.) The suggestion of 
discrimination against whites arises from an over-inclusive 
measure of the labor supply.1* Of the two types of availability

,8When the figures in the “comps dev” row of the defense labor market 
tables reach minus 1.96 standard deviations, there is statistically significant 
under-representation of whites, which raises the inference of discrimination 
against whites. (J.A. 341; Tr. 1851.) Exhibit A-278 Table 4— the one Dr. Rees 
prefers (s e e J.A. 267)— shows overall discrimination against whites at B u m ­
ble Bee, Red Salmon and Wards Cove at levels ranging from minus 23.139 to 
minus 10.269 standard deviations. (Ex. A-278 Table 4 WC, R S  and SN; Tr. 
2647.) They show under-representation of whites at the same facilities in 
cannery worker and laborer jobs at levels of minus 28.187 to minus 11.537 
standard deviations. ( I b id .) Similarly, they show under-representation of 
non-whites at Bumble Bee and Red Salmon in at-issue jobs at levels of minus 
4.585 and minus 3.672 standard deviations respectively and under-represen­
tation of whites in at-issue jobs at Wards Cove but not at a statistically 
significant level. ( I b i d )

l9Dr. Rees believed “to the extent” whites are under-represented “it is the 
result of the influence of Local 37.” (J.A 298; s e e  a ls o  i d  at 294-95.) But his 
study shows whites are markedly under-represented at Ekuk, which has no 
Local 37 workers ( i d  at 342,344-45,347-48) and at several canneries in jobs 
outside Local 37’s jurisdiction ( i d  at 348-49). Similarly, historical statistics 
show the percentage of non-whites actually declined slightly after Local 37 
came into being. (J.A. 96-98.) With whites concentrated in the choice jobs at 
W C P  and BBS (J.A 355-56), whites dominating management at W C P  and 
BBS ( i d  at 356-57) and discrimination in the society as a whole far more 
pervasive against non-whites than whites ( i d  at 354-55), Dr. Rees con­
ceded— as did his statistician— the suggestion of discrimination against 
whites might be a sign his non-white availability figure was too low. (Tr. 1851; 
J.A 237, 246-48; s e e  a ls o  i d  at 350-52.)

138



18

figures Dr. Rees offered, one makes no effort to adjust for 
migrant, seasonal work, leading him to repudiate it at trial. (Tr. 
1934; see also J.A. 2576,305-07.) The other purports to adjust 
for availability for seasonal but not migrant work. (J.A. 307-
OS.) Since it is the migrant rather than the seasonal element of 
work in the industry which largely accounts for the high per­
centage of non-whites (J.A. 103-4; see also Tr. 403-08), these 
figures assume away the central issue in defining the labor 
supply.

Dr. Rees includes people in Alaska fruit, vegetable and 
seafood processing industries (J.A. 257, 311; Tr. 1598), who 
are 47% non-white. {See p. 16, supra.) But he also includes: (1) 
university professors with no interest in work in this industry 
(J.A. 310-11, 316-18; see also id. at 319-20; Tr. 385); (2) con­
struction workers, whose season peaks at the same time as the 
salmon season, leaving them largely unavailable for work in the 
industry (J.A. 310-14; Tr. 382-83); (3) unionized construction 
workers who might have to forfeit seniority to work in a salmon 
cannery (J.A. 313-16); (4) construction contractors, whose in­
vestment in an on-going business makes them unlikely can­
didates for migrant, seasonal work {see Tr. 1774); (5) year- 
round workers in industries with a seasonal component, who 
have no interest in migrant, seasonal work (J.A. 320-21); and (6) 
all unemployed, regardless of whether they would accept mi­
grant, seasonal work {see J.A. 310). With these increasingly 
remote categories, the non-white availability figure drops from 
47% to roughly 10%.

Even WCP and BBS statistics show fully 26% of race- 
identified applicants in 1978-80 were non-white (Ex. A-133 
Table 1; Tr. 1433, 1438), despite the fact they severly under­
estimate non-white availability.20

20The statistics wSre compiled from written applications made directly to 
company offices. (SeeTr. 1518,2718,2773.) Yet non-whites are hired in the 
low-paying jobs through oral solicitation by cannery worker foremen, by bush 
pilots and at Local 37. (S e e p. 7-9, s u p r a .) Of 278 applicants for cannery worker 
jobs, only two were identified as Filipino or Alaska Native (Ex. A-133 Table 
3), the non-white groups which contribute most heavily to cannery worker 
jobs. Beyond this, many race identifications were made on the basis of those 
ultimately hired. (Tr. 1403.) Since they generally occupy the very depart­
ments from which non-whites are excluded, this undercounts non-whites. 
Finally, employment practices at W C P  and BBS deter non-whites from 
applying for upper-level jobs { se e p.15 . s u p r a ) , so once again the applications 
underestimate non-white availability.

139



19

10. HOUSING SEGREGATION
WCP and BBS house employees at the canneries almost 

completely along racial lines. (Ex. 615-17 (E.R. 105-19); Tr. 
2231, 2261.) The district court found "housing where non­
whites predominate has generally been poorer than housing 
whites predominate (sic).” (App. Cert. 1:82.) But it held the 
disparate impact on non-whites was justified by a desire to 
avoid winterizing large bunkhouses by opening small ones first 
for the heavily white crews who arrived early. (Id. at 1:126-27.) 
The court of appeals reversed, holding this rationalization 
"without more” would not "sustain a finding of business neces­
sity.” (Id at VI:37.)

One home office employee wrote an applicant for a cannery 
worker job,

We are not in a position to take many young fellows to our 
Bristol Bay canneries as they do not have the background 
for our type of employees. Our cannery labor is either 
Eskimo or Filipino and we do not have the facilities to mix 
others with these groups.

(App. Cert 1:81-82,11:1-2; J.A. 105-9.) (Emphasis added.)
Like jobs, bunkhouses are labelled by race. A.W. Brindle, 

until 1977 president of WCP, referred from 1970 on to cannery 
worker bunkhouses as "the Filipino house,” "the Eskimo quar­
ters” and “the Filipino and Eskimo areas.” (Ex. 328,361,366; 
Tr. 2026; R.P.0.132,152.) Warner Leonardo, the Bumble Bee 
cannery superintendent, referred to them from 1970 on as “the 
native bunkhouse” and the "Filipino house.” (Ex. 340; R.P.O. 
132,160.) The recent housing patterns are part of a long tradi­
tion of racial segregation in housing. (J.A. 152-54; Tr. 681, 
2387; Dep. Leonardo-1978 36; Dep. J. Brindle 24; see also Tr. 
1348-51.) Housing follows job lines to some degree but by no 
means exclusively, so assignments cannot be justified solely by 
time of arrival of different crews.11 Housing does not follow

11 Filipino culinary workers are housed with Filipino cannery workers. (J.A. 
39, 228; Tr. 872, 2061-62; Ex. 83-87 p a s s im ; R.P.O. 132, 136-37.) Filipino 
cannery workers are invariably housed apart from Alaska Native cannery 
workers. (J.A. 43,88,127-28;Tr. 197,834; Ex. 83-85 p a s s im ; R.P.O. 132,136- 
37.) Both groups, which are almost all male, are housed apart from white 
cannery workers, who are nearly all female. (J.A. 43,69,88; Tr. 37,41,2202- 
203; Ex. 83-85 p a s s im ; R.P.O. 132, 136-37.) These women are housed with 
males on other white crews, at least at two canneries. (J.A. 43, 69; (cont)

140



20

union lines consistently either.”
11. MESSING SEGREGATION

Each cannery has two messhalls, of which one is identifiably 
non-white and the other identifiably white. (E.g. J.A. 45-46, 
112-13,128-29,141-42; Tr. 36,162,196.) The non-white mess- 
hall is invariably located in or near the non-white cannery 
worker bunkhouses. {Ibid.) The district court acknowledged 
the disparate impact of the messing practices on non-whites, 
but held they were justified by a union contract which provides 
for a separate culinary crew for Local 37 workers. (App. Cert. 
1:126-28.) Since the district court also recognized a union con­
tract will not immunize an employer under Title VII {id. at 128- 
29), the court of appeals wrote it was “unsure what the (district 
court’s) conclusion was as to” the disparate impact claim of 
separate messing {id. at VI:38).

Like jobs and bunkhouses, messhalls are often designated 
by race. A.W. Brindle, until 1977 WCP’s president, referred 
from 1970 on to the “Filipino mess house” and the “white 
messhouse.” (Ex. 359, 426; R.P.O. 132, 152, 156.) Company 
records refer to the "Filipino mess hall,” “native cook” and 
“native galley cook.” (Ex. 300, 347-50, 382, 504; R.P.O. 132, 
148, 151,153; Tr. 2279.) Messing practices should be viewed 
against the backdrop of a long pattern of segregation. (J.A. 153- * ll

ll(cont) Tr. 37, 2202-203; Ex. 84-85 p a s s im ; R.P.O. 132, 136-37.) On 
occasion, white males performing the same jobs as non-white males are 
housed separately. (J.A. 69,628-30; Tr. 941,2202-03.) Culinary workers for 
the predominantly non-white messhalls are almost all housed apart from 
culinary workers for the predominantly white messhalls. (S e e J.A. 228; Tr. 
2067; Ex. 83-85 p a s s im ; R.P.O. 132, 136-37.) Different white crews are 
commonly housed together. (Tr. 239,655-57; Ex. 83-85 p a s s im ; R.P.0.132, 
136-37.)

llMale Local 37 members are nearly all non-white, but are housed apart 
from female members who are nearly all white. (See p. 19 n.21, s u p r a .) In turn, 
female Local 37 workers are housed with white male members of other 
unions. (See p. 19 n.21, s u p r a .) Beachmen, fishermen and certain culinary 
workers are all members of the Alaska Fishermen’s Union. (Ex. A-l through 
A-ll, Alaska Fishermen’s Union; Tr. 2646-47.) Yet they are frequently 
housed apart from one another. (S e e , Ex. 83-85 p a ss im ', R.P.0.132,136-37.) 
Conversely, whites who have different union affiliations or no union affiliation 
at all are often housed together. (See, e .g ., Tr. 239.)

141



21

54; Tr. 668; see also Tr. 2589.) Non-white male Local 37 work­
ers are fed separately from white female Local 37 workers. (E.g. 
J.A. 45-46, 83-84, 112-13, 141-42, 154; Tr. 78-79.) Instead, 
they are often fed with Alaska Native cannery workers, who are 
not represented by Local 37. (E.g., J.A. 45-46, 88-89; cf. App. 
Cert. 1:36.)

SUMMARY OF ARGUMENT
Under § 703(a)(2), it is illegal to “limit, segregate, or clas­

sify” employees by race, regardless of what labor market com­
parisons show. 42 U.S.C. §2000e-2(a)(2). (Emphasis added.) 
Title VII’s only mention of labor market comparisons is in 
§703(j), which discourages labor market defenses because it 
prohibits “preferential treatment. . .  on account of an imbal­
ance” between the race of those hired and those “in the avail­
able work force in any community, State, section or other area.” 
42 U.S.C. §2000e(2)(j). The labor market showing WCP and 
BBS urge is a form of “bottom line defense,” since it focuses on 
the number of jobs filled, rather than on limitations of job 
opportunities. But the Court has squarely rejected “bottom 
line” defenses to disparate impact claims. Connecticut v. Teal, 
457 U.S. 440 (1982). Job segregation statistics serve Title VU’s 
prophylactic aim, since they afford certainty, simplicity and 
ease of use, while labor market comparisons often involve un­
certainties and arcane variables. The district court’s adoption 
of the labor market defense here in any case (1) incorrectly 
assumes the legitimacy of racially segregated hiring channels; 
(2) stems from a misreading of the union contracts and (3) 
overrides eight decades of 47-70% non-white employment by 
finding the labor supply is only 10% non-white.

Skill issues do not detract from job segregation statistics 
here, because jobs are often unskilled or low-skilled or require 
only skills which can be acquired at the entry level. Nor were the 
employees required to offer statistics on non-whites qualified 
for even ostensibly skilled jobs, since: (1) WCP and BBS failed 
to identify qualifications actually applied, providing instead 
qualifications prepared for litigation; (2) WCP and BBS only 
have subjective criteria, which themselves often mask preju­
dices or stereotypes; (3) Employees need not show they meet 
qualifications which have a discriminatory effect, unless the

142



22

business necessity of the qualifications is first established.
Housing and messing segregation and race-labelling are 

prohibited by § 703(a)(2), since—as “dignitary” wrongs (Curtis 
v. Loether, 415 U.S. 189, 196 n.10 (1974))—they “adversely 
affect (one’s) status as an employee.” 42 U.S.C. §2000e-2(a)(2). 
Room and board are fringe benefits whose allocation may be 
challenged on a disparate impact theory. Because they deter 
non-whites from seeking upper-level jobs, housing and messing 
segregation and race-labelling all “tend to deprive . . . (non- 
whites) of employment opportunities.” 42 U.S.C. §2000e- 
2(a)(2). Citing only those of the district court’s contradictory 
findings on nepotism which favor them, WCP and BBS main­
tain they give no preference to relatives, but even the testi­
mony of WCP’s president establishes a preference.

Despite claims by WCP and BBS, issues of causality are 
largely absent from this case. The employees offered separate 
proof of the disparate impact of several practices. WCP and 
BBS conceded causality in briefs in the court of appeals and 
this Court. They offered evidence and proposed findings estab­
lishing causality in the district court. They omitted causality as 
a ground for their motion to dismiss for failure to make a prima 
facie case. Since the district court denied the motion, the suffi­
ciency of the prima facie case is in any event beyond challenge. 
United States Postal Service Board of Governors v. Aikens, 460 
U.S. 711 (1983). The failure of WCP and BBS to articulate their 
hiring criteria or to maintain systematic records prevented the 
employees from compiling separate statistics on the impact of 
subjective qualifications. Even so, employees need not always 
prove specific practices cause racial imbalances. Congress in­
tended to prohibit “ ‘complex and pervasive’ ” discrimination 
{Teal, 457 U.S. 440,447 n.8), which does not always lend itself 
to easy correlations between cause and effect. The Uniform 
Guidelines on Employee Selection Procedures require em­
ployers to maintain “records or other information” on the ad­
verse impact of each facet of the overall selection process. 29 
CFR § 1607.15A(2)(a). Requiring employees to show causality 
in every case would make employers the beneficiaries of their 
own record-keeping violations.

Unlike the “articulation” of a legitimate reason for osten­
sibly disparate treatment, business necessity is an affirmative

143



23

defense on which the employer bears the burden of persuasion, 
for it allows the employer to prevail by proving facts unrelated 
to the prima facie case. It entails showing a practice is essential 
to job safety and efficiency, a standard designed to limit defer­
ence to an employer’s belief in the reasonableness of its own 
practices. Because they set independent standards which serve 
the same purpose, the Uniform Guidelines provide an alternate 
standard for business necessity—namely, job relatedness. 
When it amended Title VII in 1972, Congress ratified this view 
of the business necessity defense.

Disparate treatment provides an alternate basis for affirm­
ing. Segregated hiring channels which funnel employees to 
race-labelled jobs are facially discriminatory, so the shifting 
burden analysis is inappropriate. See TransWorld Airlines, Inc. 
v. Thurston, 469 U.S. I l l  (1985). But even if it applied, the 
justifications WCP and BBS offer for racial disparities in treat­
ment—namely, job qualifications prepared for litigation and a 
misreading of the Local 37 contract—are clear pretext.

ARGUMENT
1. STATISTICS ON JOB SEGREGATION OR 

PRACTICES WHICH FOSTER IT ESTABLISH 
DISPARATE IMPACT REGARDLESS OF WHAT 

LABOR MARKET COMPARISONS SHOW
A. The Language of Title VII Makes Job Segregation 

and Practices Which Promote it Illegal
Under § 703(a)(2), it is illegal for an employer to “limit, 

segregate, or classify his employees or applicants” in any way 
which “would tend to deprive any individual of employment 
opportunities or otherwise adversely affect his status” as an 
employee because of race.” 42 U.S.C. §2000e-2(a)(2). The 
statute makes job segregation and practices which promote it 
illegal perse, subject only to the affirmative defense of business 
necessity. (See p. 44 * 45, infra.) 13

13 Claims of disparate impact arise under § 703(a)(2). W a ts o n  v. F o r t  
W o r th  B a n k  a n d  T Y ust, 108 S.Ct 2777,2783-84 (1988); C o n n e c tic u t  v. T e a l,  
467 U.S. 440, 445-46 (1982); S a t t y  v. N a s h v i l le  G a s  C o .. 434 U.S. 136, 144 
( 1 9 7 7 ) ;  G r ig g s  v. D u k e  P o w e r  C o ., 401 U.S. 424,426n.l (1971). The Court has 
not yet decided whether they also arise under §703(a)(1). Satty, 434 U.S. 
136, 144.

144



24

Segregated hiring channels "limit, segregate, or classify” 
employees or applicants by the way they are recruited. Since 
the abilities of individuals recruited through different channels 
are never compared, non-whites cannot compete effectively on 
the basis of job qualifications for upper-level jobs. The absence 
of fixed, objective job qualifications reinforces the effect of 
separate hiring channels by giving white foremen free rein in 
selecting their acquaintances. Nepotism "limit(s), segregate(s), 
or classif[ies|” employees or applicants on the basis of family 
ties, which gives whites an edge in obtaining upper-level jobs. 
Similarly, a policy of re-hiring past incumbents in their old 
departments “limit[s|, segregatejs], or classif(ies)” employees 
or applicants by the jobs they held with WCP or BBS in past 
seasons. Because jobs are racially stratified, this limits oppor­
tunities for non-whites.

"The language of Title VH makes plain the purpose of 
Congress” to "eliminate those discriminatory practices and 
devices which have fostered racially stratified job environ­
ments. McDonnell Douglas Corp. v. Green, 411 U.S. 792, 800
(1973) (emphasis added); see also Teamsters v. United States, 
431 U.S. 324, 348 (1977). Since the words of the statute are 
clear, they should be applied as read. See Consumer Product 
Safety Comm’nv. GTE Sylvania, Inc., All U.S. 102,108 (1980); 
Chandler v. Roudebush, 425 U.S. 840, 848 (1976). The Court 
has repeatedly applied the disparate impact analysis to prac­
tices which promote job segregation. Griggs v. Duke Power Co., 
401 U.S. 424, 427, 432 (1971) (transfer criteria operated as 
"built in headwinds” in plant where "Negroes were employed 
only in the labor department” while “only whites were em­
ployed” in four others); Albemarle Paper Co. v. Moody, 422 U.S. 
405, 409 (1975) (tests inhibited transfers in plant which still 
carried effects of "racial(ly) identifiable]” lines of progression); 
see also Teamsters, 431 U.S. 324, 344 (dictum).

Contrary to assertions by WCP and BBS, work force sta­
tistics are evidence of discrimination in hiring as well as promo­
tions. E.g., Teamsters, 431 U.S. 324,329,342 n.23 (jobsegrega­
tion statistics accepted where discrimination "in hiring... line- 
drivers” is alleged, since they show "(tjhose Negroes and Span- 
ish-sumamed persons who had been hired... were given lower 
paying, less desireable jobs”); accord Domingo v. New England

145



25

England Fish Co., 727 F.2d 1429, 1436 (9th Cir. 1984), mod­
ified, 742 F.2d 520 (1984); Carpenter u. Steven F. Austin State 
University, 706 F.2d 608,618,622-25 (5th Cir., 1983); James v. 
Stockham Valves and Fittings Co., 559 F.2d 310, 321-28 (5th 
Cir. 1977), cert denied, 434 U.S. 1034 (1977). Crediting work 
force statistics only when an employer announces a policy of 
promoting from within makes the employer the arbiter of its 
own discrimination, for it enables the employer to avoid liability 
by simply declining to announce the policy. Giving work force 
statistics weight only when the employer promotes from within 
permits an employer to freely perpetuate job segregation by a 
systematic failure to promote, which is itself discriminatory.24 
See Giles v. Ireland, 742 F.2d 1366,1381 (11th Cir. 1984) (“The 
failure to promote would appear to operate to ‘freeze’ blacks in 
the lowest . . . categories . . .”); Griggs, 401 U.S. 423, 424 
(practice which "freeze(s| the status quo of prior discrimina- 
t[ion|” illegal unless justified by business necessity); see Team­
sters, 431 U.S. 324, 349-50 (same) (dictum).

B. Even The Labor Market Statistics WCP and BBS 
Offer Establish a Prima Facie Case For Many Jobs
While WCP and BBS broadly challenge the sufficiency of 

the prima facie case, their own statistics establish a significant 
exclusion of non-whites from several jobs. For these jobs, their 
labor market and skills contentions are irrelevant.

From the “table(s) that in [his] judgement best enable( ) 
one to test the allegations of racial discrimination” (J.A. 267), 
Dr. Rees acknowledged a statistically significant absence of 
non-whites in: (1) tender jobs at WCP; (2) tender jobs at Red 
Salmon; and (3) machinist and fisherman jobs at Bumble Bee 
(See p. 17, supra.) Certainly, an employer "is free to adduce 
countervailing evidence” if it "discerns fallacies or deficiencies 
in the data” offered by employees to show disparate impact. 
Dothard v. Rawlinson, 433 U.S. 321,331 (1977); see also Wat­

24W C P  and BBS discourage mid-season promotions of Local 37 workers, 
because they entail paying the same person two season guarantees. (Tr. 1104, 
1134,1352-53; s e e App. Cert. 1:39.) Once the season is over, re-hire prefer­
ences and word-of-mouth recruitment inhibit promotions for the next season. 
(S e e p. 7-9, 11, s u p r a .) But even the relatively few promotions awarded go 
disproportionately to whites. (S e e Ex. 613-14; Tr. 2231, 2261.)

146



26

son v. Fort Worth Bank and Trust, 108 S. Ct. 2777,2789 (1988) 
(O’Connor, J.). But the rule has only academic relevance for 
jobs like these in which even the employer concedes a signi­
ficant exclusion of non-whites.

Beyond this, Dr. Rees concedes a statistically significant 
absence of non-whites in other jobs when re-hires are counted, 
even though his skill and labor market contentions remain 
intact.” They are: (1) fisherman, machinist, tender and carpen­
ter jobs at Bumble Bee; (2) tender and machinist jobs at Wards 
Cove; (3) tender and fisherman jobs at Red Salmon; and (4) 
tender and fisherman jobs at WCP as a whole. (J. A. 332-36; Ex. 
A-280 Table 4 WC, RS, SN and WC-RS; Tr. 2646-47.) Unless 
re-hire preferences are justified by business necessity, re-hires 
must be counted, for “treating as unassailable” a right of past 
incumbents to return in their old jobs in largely white depart­
ments “perpetuate(s) impermissibly the result of earlier dis­
crimination.” Grant v. Bethlehem Steel Corp., 635 F.2d 1007, 
1018 (2nd Cir. 1980), cert denied, 452 U.S. 940 (1981); see 
Teamsters, 431 U.S. 324, 349-50, 372-76. For reasons given 
below, the re-hire preferences are not justified by business 
necessity. (See p. 48, infra.)

C. Labor Market Comparisons Cannot Rebut or
Justify Statistical Showings of Job Segregation
Title VITs only language on labor market comparisons ap­

pears in § 703 (j)» which prohibits “preferential treatment... on 
account of an imbalance” between the race of those hired and 
those “in the available work force in any community, State, 
section or other area.” 42 U.S.C. §2000e-2 (j). This provision 
discourages uncritical labor market defenses.

WCP and BBS argue the racial imbalances in their work 
force are acceptable, because hiring area comparisons show 
they employ too many non-whites in lower-level jobs rather 
than too few in upper-level jobs. But this confuses the end with 
the means, for while hiring area comparisons are evidence of *

,8The tables Dr. Rees prefers count employees only In the first season they 
held a given job, rather than each season they filled it (See p. 3 , n. 5 s u p r a .)  
But by narrowing the statistical case, they decrease the likelihood any in­
stance of under-representation will be statistically significant (See Tr. 
2121 fk)-(l).)

147



27

violations (Hazelwood School District v. United States, 433 U.S. 
299, 307 (1977)), they do not define the violations. “Title VII 
imposes no requirement that a work force mirror the general 
population.” Teamsters, 431 U.S. 324,340 n.20. Courts, in any 
case, have rejected precisely the reasoning WCP and BBS urge. 
Carpenter, 706 F.2d 608, 622 (labor market statistics showing 
non-whites “over-represented” in lower-level jobs do not rebut 
job segregation statistics in hiring discrimination case).26

Section 703(a)(2) “speaks, not in terms of jobs and promo­
tion, but in terms of limitations and classifications.”

This Court has never read § 703(a)(2) as requiring the 
focus to be placed instead on the overall number of minor­
ity or female applicants actually hired or promoted. 

Connecticut v. Teal, 457 U.S. 440,448,450 (1982) (emphasis in 
original). Where practices conspicuously limit opportunities 
for non-whites, a labor market defense would reverse the focus 
back from opportunities to jobs, which in turn would encourage 
employers to adopt not remedial or affirmative goals but exclu­
sionary quotas, which have been historically disfavored. Cf. 
Steelworkers v. Weber, 433 U.S. 193, 208 (1979).

Labor market showings are a form of “bottom line” defense, 
which are ineffective against disparate impact claims.27

Although we noted in passing (in Dothard, 433 U.S. 32lj 
that women constituted 36.89 percent of the labor force 
and only 12.9 percent of correctional counselor positions, 
our focus was not on this “bottom line.”

26S e e  a ls o  J a m e s , 559 F.2d 310, 341 (same In hiring and promotion dis­
crimination case); P a x to n  u. U n i te d  N a t io n a l  B a n k , 688 F.2d 552,563-64 (8th 
Clr. 1982), c e r t  d e n ie d , 460 U.S. 1083 (1983) (Same in promotion discrim­
ination case); P a y n e  v. T r a v e n o l  L a b o r a to r ie s ,  In c ., 673 F.2d 798,823-24 (5th 
Clr. 1982), c e r t  d e n ie d , 459 U.S. 1038 (1982) (hiring area statistics do not 
rebut applicant flow statistic in hiring discrimination case).

27 The Court has never insisted on labor market statistics to establish 
disparate impact, but has instead relied on scores for written exams given by 
the employer (T e a l, 457 U.S. 440, 443 and n.4), pass rates for standardized 
exams compiled in other cases (G rig g s , 401 U.S. 424,430 n.6), national height 
and weight statistics (D o th a r d , 433 U.S. 321, 330) and statewide education 
statistics (G rig g s , 401 U.S. 424, 430 n.6). S e e  a ls o  B e a z e r , 440 U.S. 568, 585 
(where reliable, applicant flow statistics are preferable to hiring area statis­
tics); cf. D o th a r d , 422 U.S. 321, 330 (same).

148



28

*  *  *

The suggestion that disparate impact should be measured 
only at the bottom line ignores the fact that Title VII 
guarantees these individual respondents the opportunity 
to compete equally with white workers on the basis of job- 
related criteria.

Teal, 457 U.S. 440, 450-51 (emphasis in original).
Three times before Teal, the Court rejected labor market 

defenses. While it found the statistics flawed in Beazer, the 
Court held a disparate impact approach was not precluded by 
the fact “the percentage of blacks and Hispanics in (the em­
ployer’s) work force is well over twice that of the percentage in 
the work force in the New York Metropolitan area.” New York 
Transit Authority v. Beazer, 440 U.S. 568, 584 n.25 (1979). 
Facing criticisms of labor market statistics on disparate treat­
ment claims in Teamsters, the Court held,

At best, these attacks go only to the accuracy of the com­
parison between the composition of the company’s work 
force at various terminals and the general population of 
the surrounding communities. They detract little from the 
Government's further showing that Negroes and Spanish- 
sumamed Americans who were hired were overwhelmingly 
excluded from line-driver jobs.

Teamsters, 431 U.S. 324, 342 n. 23 (emphasis added). Sim­
ilarly, confronting disparate treatment claims in Fumco, it 
wrote an employer must “provide an equal opportunity for each 
applicant regardless of race, without regard to whether mem­
bers of the applicant’s race are already proportionately repre­
sented in the work force.” Fumco Construction Corp. v. Waters, 
438 U.S. 567, 579 (1978) (emphasis in original).

Hiring area statistics cannot rebut job segregation statis­
tics, because they do not answer the violation alleged—namely, 
among those hired choice jobs are allocated unfairly. Clearly, 
an employer may not limit non-whites to the same share of its 
payroll as they comprise in the hiring area, for if they are more 
heavily represented in its work force than the hiring area, this 
condemns them to lower wages than whites for reasons un­
related to merit, which is itself discriminatory. Cf. Bazemore v. 
Friday, 478 U.S. 385 (1986). For the same reason, an employer 
may not limit non-whites to the same percentage of choice jobs 
as non-whites comprise in the hiring area, for when—as WCP 
and BBS claim here—they are “over-represented” in an em­

149



29

ployer’8 menial jobs, this ensures a pattern of racial segregation 
for reasons unrelated to merit, which once again is discriminatory.

Relying on work force statistics facilitates “ ‘self examin­
ation]’ ” and “ *self-evaluat|ionJ,’ ” which enable employers to 
voluntarily “ ‘eliminate... the last vestiges of discrimination’ ” 
(Albemarle, 422 U.S. 405, 417-18), for they afford certainty, 
simplicity and ease of use. The Uniform Guidelines and other 
EEOC regulations require employers to record and report the 
race of employees in different job categories (29 CFR § 1602.7, 
§ 1602.13, § 1607.4A-B, § 1607.15A(l)-(2)), so employers al­
ready have the information necessary to compile work force 
statistics. By contrast, labor market questions involve uncer­
tainties, such as the effect of weighting schemes (see Markey v. 
Tenneco Oil Co., 635 F.2d 497,499 (5th Cir. 1981)), the effect of 
an employer’s recruitment practices (see Williams v. Owens- 
Illinois, Inc., 665 F.2d 918,927 (9th Cir. 1982), cert denied, 459 
U.S. 971 (1982)) and perhaps distortions of the labor market 
(Hazelwood, 433 U.S. 297,313 n.20). WCP and BBS prevailed 
here on virtually the same labor market statistics courts re­
jected In two companion cases involving the same industry. 
Domingo v. New England Fish Co., 445 F. Supp. 421, 431-33 
(W.D. Wash. 1977), reversed on otherissues, 727 F.2d 1429 (9th 
Cir. 1984), modified, 742 F.2d 520 (1984); Carpenter v. Nefco- 
FidalgoPacking Co., No. C74-407R (W.D. Wash. May 20,1982) 
(order on liability). Far from yielding certainty, reliance on 
hiring area statistics here “ 'produce(dJ different results for 
breaches of duty in situations that cannot be differentiated in 
policy.’ ” Albemarle, 422 U.S. 405, 417.

WCP and BBS argue reliance on work force statistics would 
invite employers to reduce the number of non-whites in low 
paying jobs to eliminate a pattern of job segregation. (Brief of 
Pet. 21.) But this is simply to say employers will deny non­
whites all opportunities if they must afford them equal oppor­
tunities, which is no basis for limiting a statute whose aim is ‘‘to 
assure equality of employment opportunities” by eliminating 
practices ‘‘which have fostered” racial segregation in jobs. 
McDonnell Douglas, 411 U.S. 792, 800. WCP and BBS also 
maintain work force statistics are an unwieldy measure of dis­
crimination, since an employer will not know the percentage of 
non-whites it employs until after it has finished hiring. (Brief of

150



30

Pet 21-22.) But few employers have turnover so rapid this un­
certainty will be meaningful. Even WCP and BBS—seasonal em­
ployers who reconstitute their work forces every year (see App. 
Cert. 1:40)—have had a relatively constant percentage of non­
whites in their work force for decades. (See, e.g. J.A. 151-53.)

D. The Labor Supply Findings Were Induced 
by Errors of Law

The district court’s findings on the labor supply were in­
duced by three errors of law.28

First, one of two factors to which the district court attri­
buted the concentration of non-whites in menial jobs is hiring 
from near the canneries. (App. Cert. 1:37-39.) But since it relied 
on a line of cases which—like Hazelwood—define the relevant 
labor market as the community surrounding the work place (id. 
at 1:109), it could not logically discount the effects of the prac­
tice by saying it distorts the work force. Recruiting from heavily 
non-white areas only for lower level jobs can itself be discrim­
inatory. Domingo, 445 F. Supp. 421,433; see also Williams, 665
F. 2d 918,927; Markey, 635 F.2d 497,500-01. By assuming the 
legitimacy of the practice without proof of business necessity, 
the district court prevented the employees from even chal­
lenging it.

Second, the other factor to which the district court 
attributed the concentration of non-whites in low-paying jobs is 
Local 37 dispatching. (App. Cert. 1:36.) But under its labor 
contract, Local 37 enjoys no control over selecting non-resi­
dent cannery workers. (See p. 8-9, supra.) Nor does it have an 
exclusive hiring hall.2* (Ibid.) The hiring provisions in the Local * VII

18The Court may affirm on this basis, even though the court of appeals did 
not reach it S e e  T r a n s W o r ld  A ir l in e s , Inc. v. T h u rs to n , 469 U.S. Ill, 119n. 14 
(1985). Findings of fact affected by errors of law are infirm. S e e  P u l lm a n  
S ta n d a r d  v. S w in t , 456 U.S. 273, 292 (1982). A  remand to re determine the 
labor supply is unnecessary, since the errors in the district court’s findings 
signal the inability of W C P  and B B S  to discredit the job segregation statis­
tics. S e e  D o th a r d , 433 U.S. 321, 331.

”  If Local 37 has a role in hiring, it is only because W C P  and BBS informally 
delegate authority to it Yet an employer may not avoid liability under Title
VII by delegating management prerogatives to third parties. S e e  A r iz o n a  
G o v e r n in g  C o m m it te e  v. N o r r is , 463 U.S. 1073,1089-91 (1983). Whether an 
Informal grant of authority to a union is an "institutional constraint" (cont)

151



31

37 contract are almost identical to those in contracts covering 
upper-level jobs. (Ex. A-l through A-l 1, Local 37, ILWU, Alas­
ka Fishermen’s Union, Machinists Union, Carpenters Union; 
Tr. 2345-46.) Yet neither the employers nor the district court 
attributed any distortion of the work force to other unions. The 
district court’s erroneous reading of the Local 37 agreement 
may be freely reviewed on appeal. Mackey u. National Football 
League, 543 F.2d 606,612 (8th Cir. 1976), cert dismissed, 434 
U.S. 801 (1977); see also C. Wright and A. Miller, 9 Federal 
Practice and Procedure §2588 p. 750 (1971). Even so, of the 
five canneries originally covered by this case, only one—Ekuk 
—does not use Local 37. (App. Cert. 1:35,1:37-38.) While it has 
the lowest percentage of Filipinos (id at 37-38), it has the 
highest percentage of non-whites generally (Ex. 588-92; Tr. 
2231, 2261). Even if every worker from the Lower 48—where 
Local 37 has jurisdiction—were white, the WCP and BBS work 
force would still be 29% non-white, because of the number of 
Alaska Natives. (Tr. 366.)

Third, courts credit hiring area statistics only because, 
Absent explanation, it is ordinarily to be expected that 
nondiscriminatory hiring practices will in time result in a 
work force more or less representative of the racial and 
ethnic composition of the population in the community 
from which employees are hired.

Teamsters, 431 U.S. 324,340 n.20 (emphasis added). But since 
for eight decades this industry has been far more heavily non­
white than its hiring areas, the rationale for using such statistics 
is absent. Even so, WCP and BBS actively recruit for all jobs, so 
the issue is not whether their work force fairly reflects the areas 
from which they hire, but whether in recruiting they give whites 
and non-whites an equal chance at the desirable jobs. On this 
issue, work force statistics speak eloquently. Significantly, in an 
industry which has been 47-70% non-white for eight decades, 
upper-level jobs at WCP and BBS remain at least 90% white. 
See Domingo, 445 F. Supp. 421, 432. 29

29(cont) is distinct from the issue in G e n e r a l  B u ild in g  C o n tr a c to r s  A s s o c i ­
a tio n , Inc. u. P e n n s y lv a n ia , 458 U.S. 375 (1982)— namely, whether an e m ­
ployer who delegates authority to a union in collective bargaining makes the 
union its agent

152



32

E. Alleged Skill Requirements Do Not Detract From 
Work Force Statistics Here

WCP and BBS erroneously suggest work force statistics 
have no value, since their upper-level jobs require special skills 
not found generally among their workers.

First, this rationale cannot apply to those jobs which the dis­
trict court found were unskilled. (See App. Cert. 1:107-08.)

Second, this rationale cannot apply to jobs for which the 
required skills can be acquired through experience in entry- 
level jobs in white departments (see p. 13-14, supra), since ex­
perience requirements “cannot be automatically applied to 
freeze out” non-whites, when "for the years of its segregated 
policy” the employer did not “afford them an opportunity to 
acquire experience.” Rowe v. General Motors Corp., 457 F.2d 
348, 358 (5th Cir. 1972).

Third, this rationale cannot apply to jobs which—even as­
suming the accuracy or pertinence of the district court’s find­
ings on qualifications which could be “reasonably required ” 30 
(App. Cert 1:58)—entail only skills “many persons possess or can 
fairly readily acquire.” Hazelwood, 433 U.S. 299,308 n.13. The 
district court construed “skills” in a highly rarefied sense, for 
while this Court held statistics on qualified non-whites are 
unnecessary for cross-country truck driving jobs (see id. at 308 
n.13), it ruled they are required for truck driving on the beach 
(App. Cert. 1:108), because in a seasonal industry the particular 
skill of driving vehicles is “not readily acquirable” (ibid).

For reasons given below, the work force statistics showed 
disparate impact even for ostensibly skilled jobs.

2. THE EMPLOYEES DID NOT HAVE TO OFFER 
STATISTICS ON QUALIFIED NON-WHITES, SINCE 
THE EMPLOYERS NEVER IDENTIFIED CRITERIA 
ACTUALLY APPLIED, THEY LACKED OBJECTIVE

QUALIFICATIONS AND THE QUALIFICATIONS 
THEY DID USE HAD A DISPARATE IMPACT

For three reasons, the employees were not required to offer 
statistics on qualified non-whites for any jobs here.

30S e e App. Cert. 1:66-67,1:70-71,1:73 (quality control, beach gang, truck 
driver, office assistant and tender cook).

153



33

First, WCP and BBS failed to show what hiring criteria were 
actually applied. Under Title VII, ability is tested only under 
qualifications actually applied. Franks v. Bowman Transpor­
tation Co., 424 U.S. 747, 773 n. 32 (1976) (back pay defeated 
only under “non-discriminatory standards actually applied')’, 
Albemarle, 422 U.S. 405, 433 (use of subjective rankings in 
validation inappropriate since “no way to determine whether 
the criteria actually considered” were job related) (emphasis in 
original in each). The burden of proving a job requires special 
skills or experience rests with the employer.31 Only qualifica­
tions actually applied can be tested for fair application (see 
McDonnell Douglas, 411 U.S. 792, 804) or business necessity 
{see Albemarle, 422 U.S. 405,433). While—within these limits— 
Title VII makes qualifications the employer’s prerogative (see 
Rowe, 457 F.2d 348,358), the employer must at least disclose 
the criteria it applied, for “ ‘(ojne clear purpose of discrim­
ination law is to force employers to bring their employment 
processes into the open.’” Segar v. Smith, 738 F.2d 1249,1276 
(D.C. Cir. 1984), cert denied, 471 U.S. 1115 (1985). Only then 
can qualifications be the "controlling factor” Congress intended. 
Griggs, 401 U.S. 424, 436. When an employer fails to disclose 
criteria actually applied, an employee need not offer statistics 
on qualified non-whites. EEOC v. Rath Packing Co., 787 F.2d 
318, 328, 336 (8th Cir. 1986), cert denied, 107 S. Ct. 307
(1986); Domingo, 727 F.2d 1429, 1437 n.4; Trout v. Lehman, 
702 F.2d 1094, 1102 n.10 (D.C. Cir. 1983), vacated on other 
grounds, 465 U.S. 1056 (1984).

By relying on qualifications prepared for litigation, WCP 
and BBS adopted a strategy which was rejected in two compan­
ion cases. Domingo, 445 F. Supp. 421, 437-38; Carpenter, No. 
C74-407R (W.D. Wash. May 20, 1982) (order on liability). An 
employer may not impose more stringent qualifictions on non­
whites than whites either in practice (McDonnell Douglas, 411 
U.S. 792, 804) or in proof (Domingo, 445 F. Supp. 421, 438). 
Even standards which are "reasonable” will not defeat a Title

31E E O C  v. R a th  P a c k in g  C o ., 787 F.2d 318, 336 (8th Cir. 1986), c e r t  
d e n ie d , 107 S. Ct 307 (1986); D o m in g o , 727 F.2d 1429, 1437 n.4; M o o r e  v. 
H u g h e s  H e lic o p te r s ,  In c ., 708 F.2d 475,483 (9th Cir. 1983); C h r is le r v .  C o m ­
p le te  A u to  T r a n s i t ,  In c ., 645 F.2d 1251, 1259 n.5 (6th Cir. 1981); E E O C  v. 
R a d ia to r  S p e c ia l t y  C o ., 610 F.2d 178, 185 (4th Cir. 1979).

154



34

VII claim, if they were never imposed during the liability period. 
Laffey v. Northwest Airlines, Inc., 567 F.2d 429, 456-57 (D.C. 
Cir. 1976), cert denied, 434 U.S. 1086 (1977).

Second, an employee need not as part of a prima facie case 
offer statistics on non-whites who meet subjective qualifica­
tions. When it endorsed use of statistics on experienced teachers 
in Hazelwood, the Court did not insist on a further showing of 
non-whites who met the subjective qualfications of “ ‘most 
competent’ ” or “ ‘personality, disposition, appearance, poise, 
voice, articulation, and ability to deal with people.’ ” Hazel­
wood, 433 U.S. 299,302,308 n.13. Even where skilled jobs are 
at issue, courts do not insist on statistics on non-whites who 
meet subjective qualifications.” Since subjective criteria can 
mask “subconscious stereotypes and prejudices” (Watson, 108
S. Ct 2777, 2786), statistics on them just “measure . . .  the 
amount of discrimination operating through” them. Segar, 738 
F.2d 1249,1276. Simply saying—as WCP and BBS do—they 
look for a “qualified person,” “skill” or “experience” is no 
substitute for having objective qualifications, since “ ‘affirma­
tions of good faith in making individual selections are insuffi­
cient to dispel a prima facie case of systematic exclusion.’ ” 
Teamsters, 431 U.S. 324, 343 n. 24. Undefined “job related 
experience” is not considered objective when—as here—"(e)ach 
hiring decision [is] made by a cannery superintendent or a 
foreman on the basis of his personal judgment.” Domingo, 727 
F.2d 1429, 1433.

Third, an employee “cannot be required to prove that he 
was qualified... under a system he alleges to be discriminatory 
unless the legitimacy of the system is first established.” Wang *

" S e g a r , 738 F.2d 1249, 1274-75 (GS-7 to GS-12 positions); C a v ia le  v. 
S t a t e  o f  W isc o n s in , 744 F.2d 1289, 1294 (7th Cir. 1984) (regional director); 
M o z e e  v. J e f fb o a t ,  In c . 746 F.2d 365, 372-73 (7th Cir. 1984) (foreperson); 
W a n g  v. H o ffm a n , 694 F.2d 1146,1148 (9th Cir. 1982) (GS 12 positions); D e  
M e d in a  v. R e in h a r t, 686 F.2d 997,1007 (D.C. Cir. 1982) (technicians, writers 
and editors); B u r r u s  v. U n i te d  T e le p h o n e  C o m p a n y  o f  K a n s a s ,  In c ., 683 F.2d 
339, 342 (10th Cir. 1982), c e r t  d e n ie d , 459 U.S. 1071 (1982) (accounting 
supervisor); L y n n  v. R e g e n ts  o f  th e  U n iv e r s i ty  o f  C a lifo r n ia , 656 F.2d 1337, 
1344-45 (9th Cir. 1981), c e r t  d e n ie d , 459 U.S. 823 (1982) (university pro­
fessor); D a v is  v. C a lifa n o , 613 F.2d 957, 964 (D.C. Cir. 1979) (research 
chemist); s e e  a ls o  D o m in g o , 727 F.2d 1429, 1437 n.4.

155



35

v. Hoffman, 694 F.2d 1146,1148 (9th Cir. 1982).33//a2e/u;ood, 
a disparate treatment case, is distinguishable,34 * * since a prima 
facie case of disparate treatment is designed to raise the infer­
ence of illegal intent by eliminating “the two most common 
legitimate reasons on which an employer may rely” in rejecting 
applicants, one of which is “an absolute or relative lack of 
qualifications.” Teamsters, 431 U.S. 324, 358 n.44; see also 
Texas Department of Community Affairs v. Burdine, 450 U.S. 
248,253-54 (1981). By contrast, the aim of a prima facie case of 
disparate impact is to show non-whites fail “in a significantly 
discriminatory pattern” to meet the qualifications imposed. 
Dothard, 433 U.S. 321, 329. Because the employees showed 
disparate impact (see p. 39-40, infra), they were relieved of the 
need to offer statistics on qualified non-whites. See Beazer, 440 
U.S. 568, 685 (statistics on "otherwise qualified” non-whites 
required when some qualifications remain unchallenged) (em­
phasis added).

3. HOUSING AND MESSING SEGREGATION 
AND RACE-LABELLING HAVE A DISPARATE 

IMPACT ON NON-WHITES
Because they “limit, segregate, or classify” employees along 

racial lines, housing and messing segregation and race-labelling 
fall within the literal terms of §703(a)(2)’s prohibition. See 42 
U.S.C. §2000e-2(a)(2). Even if housing and messing are as­
signed by crew or time of arrival rather than race (see App. Cert. 
1:126), the racial impact is still clear (see Ex. 615-17 (E.R. 105- 
19); Tr. 2231, 2261).

WCP and BBS argue these practices do not “tend to de­
prive any individual of employment opportunities” (42 U.S.C. 
§ 2000e-2(a)(2)), so they survive a § 703(a)(2) challenge. (Brief 
of Pet. 28-29.) But § 703(a)(2) also covers practices which 
"adversely affect (an individual’s) status as an employee.” 42

33A c c o r d E E O C v .  S t  L o u is - S a n F r a n c is c o R y .  C o ., 743 F.2d739,742 (10th 
Cir. 1984);Fad/ia/o. C i ty  a n d  C o u n ty  o f  S a n  F ra n c isc o , 741 F.2d 1163,1165- 
66 (9th Cir. 1984); B u s h e y  v. N e w  Y o rk  S t a t e  C iv i l  S e r v ic e  C o m m is s io n , 7 3 3  
F.2d 220, 225 (2nd Cir. 1984), c e r t  d e n ie d , 469 U.S. 1117 (1985).

34 S e e  a ls o  M a y o r  o f  P h i la d e lp h ia  v. E d u c a t io n a l  E q u a l i t y  L e a g u e , 415 U.S.
605, 620-21 (1974) (statistics on qualified non-whites required only when
plaintiffs “do not challenge the qualifications for service”).

156



36

U.S.C. § 2000e-2(a)(2). Since racial segregation is a “dignitary” 
wrong (Curtis v. Loether, 415 U.S. 189, 196 n.10 (1974)), it 
“adversely affectjsj” one’s status as an employee. {See, e.g., J.A. 
405-06; Tr. 79,836-37.) “Title VII is not limited to ‘economic’ 
or ‘tangible’ discrimination,” but “affords employees the right 
to work in an environment free from discriminatory intimi­
dation, ridicule and insult.” Meritor Savings Bank v. Vinson, 
477 U.S. 57, 64-65 (1986) (construing § 703(a)(1)).

Second, in any case, room and board are fringe benefits. See 
e.g. Domingo, 727 F.2d 1429,1446. Claims of discrimination in 
fringe benefits may be raised under a disparate impact theory.* 38 
Non-whites lost fringe benefits because largely non-white bunk- 
houses were “generally poorer” (App. Cert. 1:82), less spacious 
(Ex. 620-22; Tr. 2231,2261) and often simply squalid {see, e.g., 
J.A. 39-40,44-45,127-28,143; Tr. 31,38,77-78,162,197-98, 
808,1039). Non-whites were so dissatisfied with the food they 
held food strikes.36 {See, e.g., J.A. 47; Tr. 200, 284.)

Third, segregation in housing and messing “isolate|s non­
whites] . . .  from the ‘web of information’ about higher-paying 
jobs.” Domingo, 445 F. Supp. 421,439. Similarly, it deters non­
whites—-as does race-labelling—from seeking upper-level jobs, 
because of the clear message it conveys. {See p. 15 supra.) 
These practices thus “tend to deprive . . .  (non-whites] of em­
ployment opportunities.” 42 U.S.C. §2000e-2(a)(2).

WCP and BBS argue racial imbalances in housing and 
messing are due to the abundance of non-whites Local 37 
dispatches. (Brief of Pet 27.) But the district court found on 
stipulated facts cannery superintendents were “ultimately res­

38N a s h v i l le  G a s  C o . v. S a t t y , 434 U.S. 136, 139-45 (1977) (under 
9703(a)(2)); E E O C  v. J .C .  P e n n e y  C o ., 843 F.2d 249,251-52 (6th Cir. 1988); 
C o lb y  v .J .C .  P e n n e y  C o ., 811 F.2d 1119,1126-27 (7 th Cir. 1987); W a m b h e im  
i>. J .C .  P e n n e y  C o ., 705 F.2d 1492,1494 (9th Cir. 1983), cert d e n ie d , 467 U.S. 
1255 (1984) (under §703(a)(1)).
38 W C P  and B B S  say differences in food are due to request* by Local 37, 

the ability of the cooks and personal tastes of older Filipino crew members. 
(Brief of Pet 27-28.) But even if this explained segregated messing for Alaska 
Natives or feeding largely white Local 37 female workers apart from their 
largely non-white male counterparts (s e e p. 21, s u p r a ) , W C P  and BBS could 
not escape liability for the racial differences by delegating decisions to third 
parties. A r iz o n a  G o v e r n in g  C o m m it te e  v. N o r r is , 463 U.S. 1073,1089 and n.21 
(1983); G ra n t, 635 F.2d 1016.

157



37

ponsible for assigning employees to bunkhouses (and| assign­
ing crews to dining areas.” (App. Cert. 1:37; R.P.O. 9; see also, 
e.g., J.A. 39,53-55,73,128-29; Tr. 833.) Beyond this, housing 
and messing assignments often have little to do with job or 
union affiliation. (See p. 19-20 n. 21-22, supra.) WCP and BBS 
maintain non-whites could “opt out” of discriminatory messing 
practices by taking occasional meals in largely white messhalls 
on appropriate notice. (Brief of Pet. 29.) But non-whites who 
tried to eat in largely white messhalls were sometimes rebuffed. 
(See, e.g., J.A. 73-74; Tr. 40; see also Tr. 668.) Even so, because 
messhalls were “assignjedj” (App. Cert 1:37), non-whites did not 
have an “entirely voluntary” choice. Bazemore, 478 U.S. 385, 408.

4. NEPOTISM HAS A DISPARATE IMPACT 
ON NON-WHITES HERE

The district court made contradictory findings on nepotism, 
citing on the one hand its “pervasive” nature, while saying on 
the other there was no “preference” for relatives. (See p. 10, 
supra.) Invoking only findings in their favor, WCP and BBS 
disclaim any nepotism, saying relatives were “chosen because 
of their qualifications” and not “due to inexperience.”37 (App. 
Cert. 1:105,1:122.) But even if the latter findings control, they 
are premised on too narrow a legal standard, for nepotism can 
involve preference in recruiting employees and publicizing job 
opportunities as well as in evaluating qualifications. Domingo, 
747 F.2d 1429,1436; Grant, 635 F.2d 1007,1012,1016-17. On 
this, WCP’s president acknowledged, “jTjhere is no doubt 
relatives have better information as to what jobs are available.” 
(J.A. 156,183-84.) Significantly, the district court found appli­
cations of non-whites often failed because they were untimely 
or made to the wrong person. (App. Cert. 1:115-17.)

During 1970-75, roughly 67% of administrative jobs, 43% of 
quality control jobs, 39% of clerical jobs, 37% of fisherman 
jobs, 29% of machinist jobs, 27% of tender jobs and 20% of 
beach gang jobs were filled by individuals who had a relative at *

*7 W C P  and BBS argue the findings they cite are not clearly erroneous. 
(Brief of Pet 25.) But the “clearly erroneous” rule does not apply to contra­
dictory findings. S e e  L e g a te  v. M a lo n e y , 334 F.2d 704,707-08 (lstCir. 1964), 
c e r t  d e n ie d , 379 U.S. 973 (1965); C. Wright and A. Miller, 9 F e d e r a l  P r a c t ic e  
a n d  P r o c e d u r e §2614 p. 812 (1971).

158



38

the same cannery in the same or a prior year.38 The justification 
WCP’s president offered—namely, hiring a candidate’s rela­
tives is a way of attracting him or her to the company (see p. 11, 
supra)—establishes relatives are hired for reasons other than 
merit. Finally, the criticisms WCP and BBS offer of the statis­
tics are without real basis. 39
B. THE EMPLOYEES ESTABLISHED CAUSATION

Of the three requirements the court of appeals articulated 
for a prima facie case of disparate impact, one is “show(ing| the 
causal relationship between the identified practices and the 
(disparate] impact.” (App. Cert. V: 19-20.) WCP and BBS con­
ceded this element, for the court of appeals observed “(TJhe 
challenged practices are agreed to cause [the] disparate im­
pact” (id. at V:29) and “|T|he companies concede the causal 
relationship between their hiring criteria and the number of 
non-whites in the at-issue jobs” {id. at VI:24-2540). Even in their

38 S e e Ex. 583-85,608-10 (E.R. 102-104); Tr. 2231,2261. Exhibits 608-10 
show nepotistic hires for 1970-75, while Exhibits 683-85 show total hires for 
1970-80. To compute the ratio of nepotistic to total hires in 1970-75, the e m ­
ployees assumed roughly 55% of the total 1970-80 hires were made in 1970-75.

39 W C P  and BBS argue the statistics fail to exclude persons who become 
related by marriage after they were hired. (Brief of Pet 26 n.40.) But they 
failed to meet their burden of showing this would affect the racial impact 
apparent from the charts. S e e  C a p a c i  v. K a t e  a n d B e s t h o f f  In c ., 711 F.2d 647, 
653-54 (5th Cir. 1983), c e r t  d e n ie d , 466 U.S. 927 (1984). Similarly, W C P  and 
BBS maintain the first of two relatives hired should not be counted. (Brief of 
Pet 26 n.40.) The nepotism charts take account of this criticism when the 
relatives work in different years. (Ex. 603-05 (E.R. 65-101); Tr. 2231,2261.) 
Even so, halving the nepotistic hires— to 172 white and 1 non-white— would 
not alter the clear pattern.

40 W C P  and B BS said recruiting for cannery workers in Alaska Native 
villages and through a largely Filipino local creates the abundance of non­
whites in menial jobs. (Brief of Appellees 8 and 29.) But this only means hiring 
through separate channels has a disparate impact on non-whites. W C P  and 
BBS argued counting re-hires aggravates the statistical picture, since half of 
the challenged hiring decisions are attributable to a practice of re-hiring 
incumbents in their old jobs. (Id . at 34.) But this is simply to say the practice 
has a “lock-in” effect in an already segregated job environment Similarly, 
W C P  and B B S  claim the racial imbalance in jobs results from the inability of 
non-whites to meet the undisputedly subjective qualifications they impose. 
(See id. at 27-28.) But this only means the criteria disqualify non-whites at a high­
er rat# than whites, an observation which virtually defines disparate impact

1 5 9



39

brief before this Court, WCP and BBS openly concede the 
causal links.41 Yet they argue the Court should disregard their 
admissions, forcing the employees to show what everyone agrees 
is true.

First, in any case, the employees offered separate proof of 
the racial impact of separate hiring channels,42 although it 
might have been superfluous, since the causal link “is quite 
clear.” Domingo, 727 F.2d 1429,1436 n.3. They offered separ­
ate statistics on the racial impact of nepotism (See p. 10-11, 
supra) and re-hire preferences (see p.14, supra). They could not 
offer separate statistics on the disparate impact of subjective 
qualifications, since: (1) WCP and BBS never identified the 
criteria they actually applied, leaving qualifications invisible 
apart from their application through word-of-mouth recruit­
ment; (2) WCP and BBS destroyed applications throughout 
nearly the entire case period (see p. 14, supra), so the effects of 
subjective qualifications and word-of-mouth recruitment could 
not be separated; and (3) Their personnel records are so sketchy 
WCP and BBS had to hire a firm to collect background infor­
mation on their employees through interviews in Alaska just 
before trial (see p. 13 n.14, supra), a circumstance which makes 
regression analysis impractical. Under these conditions, sep­
arate statistics were not required. See Watson, 108 S. Ct. 2777 
(effect of subjective criteria measured through application in 
interview process). Even so, once DeFrance identified his hypo­
thetical qualifications, the employees showed they had a dis­

4,W C P  and B B S  say “the relatively low percentage of non-whites in the at- 
issue jobs is attributable... (in parti to the ‘rehire’ practice." (Brief of Pet. 36.) 
They acknowledge use of separate hiring channels is a cause of job segre­
gation, saying Local 3 7 is a “ ‘source’ (which] produced an over-representation 
of non-whites in the cannery worker jobs.” (Id . at 23; see a ls o  i d  at 39.) 
Similarly, they appear to concede their asserted qualifications have a dis­
parate impact ( I d  at 45.)

42This includes statistics on the racial mix of Alaska Native villages (Ex. 
480; Tr. 2026), stipulations on the race of cannery workers dispatched by 
Local 37 (J.A. 3-6), statistics on the race of incumbents in each job (see p. 4-6, 
supra), admissions on the racial impact of the practices (see, e .g .. Ex. 394; Tr. 
3121,3140-41) and anecdotal evidence, such as the testimony of the machin­
ist foreman whose years of word-of-mouth recruitment turned up whites but 
no non-whites. (J.A. 14-18.)

160



40

parate impact.43 WCP and BBS say the employees claimed 
other practices—a total of 16—contributed to job segregation 
(Brief of Pet. 31), but this simply is not true.44

Second, WCP and BBS never asserted in their motion to 
dismiss the employees failed to prove causality. (Tr. 2294-98, 
2310-12.) The district court in any case denied the motion, 
saying “I feel that [plaintiffs] have established a prima facie 
case”. (Tr. 2313.) When an employer “fails to persuade the 
district court to dismiss (a Title VIII action for lack of a prima 
facie case,” the sufficiency of the prima facie case can no longer 
be challenged either in the trial court or an appellate court. 
United States Postal Service Board of Governors v. Aikens, 460 
U.S. 711,714-15 (1983) (disparate treatment); Bazemore, 478 
U.S. 385, 398 (same). Once the employees rested, WCP and 
BBS offered evidence showing separate hiring channels caused 
job segregation, re-hire preferences perpetuated the job segre­
gation and DeFrance’s hypothetical qualifications had a dis­
parate impact. (E.g. Tr. 1868-69, 1880-82; see p. 13-14 n. 14, 
17, supra.) Following trial, they even proposed findings on 
causality, which the district court adopted. (Defs. Prop. Find, 
and Concl. 4-6, 31.)

Whether or not in other cases it is “unrealistic to suppose 
that employers can . . . discover and explain the myriad of 
innocent causes that may lead to statistical imbalances in the 
composition of their work forces” (Watson, 108 S. Ct. 2777, 
2787 (O’Connor, J.)), it clearly is not when the employer has 
already done so. WCP and BBS defended disparate treatment

43S e e p. 13-14, n. 14,17, s u p r a . The employee Interviews showed disparate 
impact under the Four Fifth’s rule of the Uniform Guidelines. 29 C F R  
§ 1608.4.C; s e e  T e a t, 457 U.S. 440, 443 n.4.

44Besides the practices discussed here: (1) failure to post openings was 
treated as part of word-of-mouth recruitment; (2) lack of formal promotion 
procedures highlights the job segregation statistics (se e p. 25 and n. 24, 
s u p r a ) ; and the employees challenged (3) discriminatory terminations, (4) pay 
discrimination, (5) retalitory discharge, (6) no-fraternization rules and (7) 
assigning non-whites menial make-work tasks, but all as Independent vio­
lations, rather than practices which contribute to job segregation. The e m ­
ployees (8) never challenged the English language requirement, because 
W C P  and B B S  interrogatory answers showed all or nearly all class members 
met It (Ex. 73-76; R.P.0.132,136.) The district court observed "this issue is 
not squarely addressed by the parties." (App. Cert 1:102.)

161



41

claims here “by introducing) evidence showing that (specific) 
employment practice(s) in fact cause(d) the observed (statisti­
cal) disparity,” but in so doing they made the case “ripe for 
resolution using disparate impact analysis.”45 If they now com­
plain the courts believed their evidence or accepted their argu­
ments, it is an odd complaint indeed.

Since § 703(a)(2) prohibits not just practices which cause 
job segregation but the segregation itself (42 U.S.C. §2000e- 
2(a)(2)), separate proof of the causes underlying the statistics 
should in any case not be required. Beyond this, when it amended 
Title VII in 1972, Congress stated its intent to reach “complex 
and pervasive” discrimination, which “(e)xperts familiar with 
the subject generally describe . . .  in terms of ‘systems’ and 
‘effects’ . . .  ”4* Similarly, it recognized “ ‘(u)nrelenting broad- 
scale action against patterns or practices of discrimination’ was 
essential if the purposes of Title VH were to be achieved.” 
EEOC v. Shell Oil Co., 466 U.S. 54, 69 (1984). Complex dis­
crimination is not always amenable to easy correlations be­
tween cause and effect, so requiring them can defeat this aim. A 
selection process can be “so poorly defined that no specific 
criterion can be identified with certainty, let alone be con­
nected to the disparate impact.” Watson, 108 S. Ct. 2777,2797 
n.10 (Blackmun, J.). Even relatively well-defined practices 
overlap, as nepotism and word-of-mouth recruitment do here, 
making it hard to separate out each’s effects. Sometimes it is 
“the interaction of two or more components” of a selection **

* * S eg a r , 738 F.2d 1249,1270; a c c o r d  L a t in o s  U n id o s D e  C h e ls e a  v. S e c r e ­
ta r y  o f  H o u s in g , 799 F.2d 774,787 n.22 (1st Cir. 1986); L e w is , 773 F.2d 561, 
571 n. 16; G riff in  v. C a r lin , 755 F.2d 1516,1528 (11th Cir. 1985). The same 
reasoning applies when an employer counters broad statistics showing dis­
parate Impact by showing the disparities were caused by practices which are 
justified by business necessity. S e e , e .g ., G re e n  v. U S X  C o rp ., 843 F.2d 1511, 
1524-25 (3rd Cir. 1988), p e t i t io n  f o r  c e r t  f i le d , 57 U.S.L.W. 3123 (U.S. July 
23,1988) (No. 88-141).

M  T e a l 457 U.S. 440, 447 n.8, q u o tin g S. Rep. 92-415 p. 5 (1971); s e e  a ls o  
H.R. Rep. No. 92-238 p. 8 (1971). This legislative history is pertinent, since 
Congress amended 9703(a)(2) in 1972 to include the phrase “applicants for 
employment" and expanded its scope to cover local, state and federal e m ­
ployers. S e e  T e a l, 457 U.S. 440, 447 n.8; c o m p a r e  T e a m s te r s , 431 U.S. 324, 
354 n.39 (this legislative history of little value in construing sections un­
affected by 1972 amendments).

162



42

process which creates the disparate impact. Griffin v. Carlin, 
755 F.2d 1516, 1525 (Uth Cir. 1985). This is especially true 
when they interact simultaneously—as do word-of-mouth re­
cruitment and lack of objective criteria—rather than serially. 
The employee need not always shoulder the burden of proof on 
causation alone. ML Healthy City Board of Education v. Doyle, 
429 U.S. 274, 286-87 (1977) (employee need only show con­
stitutionally protected conduct was a " ‘substantial factor’ ” or a 
“ ‘motivating factor’ ” leaving employer to establish “by a 
preponderance of the evidence” it was not the cause of the 
discharge).

The Uniform Guidelines require each employer with oyer 
100 employees to “maintain and have available for inspection 
records or other information which will disclose the impact
which its tests and other selection procedures have---- ” 29
CFR § 1607.4A; see also 29 CFR § 1607.15A(2)(a). “Where a 
total selection process for a job has an adverse impact, the [em­
ployer] should maintain and have available records or other 
information showing which components have adverse impact ” 47 
29 CFR §1607.15A(2)(a) (emphasis added). Because the 
EEOC issued these regulations under an express mandate 
from Congress,* 4® they have the “force of law.” See United States 
v. Nixon, 418 U.S. 683,695 (1974). While WCP and BBS argue 
“it is entirely unlikely that (an employer) does or could keep 
track of the statistical effect" of its practices (Brief of Pet. 35) 
(emphasis in original), this is exactly what the law ̂ requires. 
When compliance “would result in undue hardship,” the em­

47There are abbreviated requirements for employers with fewer than 100 
employees (29 C F R  8 1607.15 A(l)), but they are inapplicable here. (See Ex. 
588-90 (E.R. 35-37); Tr. 2231, 2261.) A  more general regulation requires
employers to keep applications for at least six months. 29 CFR § 1602.14(a).
While it exempts seasonal jobs (29 C F R  § 1602.14(b)), the exemption does 
not affect the more specific obligation to keep records showing adverse 
impact The previous E E O C  Guidelines on Employee Selection Procedures 
Imposed record-keeping obligations like those in the Uniform Guidelines. 29 
C F R  § 1607.4(a)(1978).

4*43 F.R. 38, 312 (1978). The EEOC "shall, by regulation, require each 
employer... to maintain such records as are reasonably necessary to carry out 
the purposes of this title,” "consult|ingj with other interested federal agen­
cies” to "coordinate its requirements with those adopted by such agencies." 
42 U.S.C. 8 2000e-8(c) and (d).

163



43

ployer may apply to the EEOC or a district court for an exemp­
tion (42 U.S.C. § 2000e-8(c)), but there is no evidence WCP or 
BBS ever did so. Requiring the employee to prove causality 
when the employer’s record-keeping violations make it impos­
sible rewards the employer for its wrong-doing, when in fact “the 
wrongdoer (should) bear the risk of the uncertainty which his 
own wrong has created.” Bigelow v. RKOPictures, Inc., 327 U.S. 
251,265 (1946).

8. WCP AND BBS HAVE NOT MET THEIR HEAVY 
BURDEN OF PROVING BUSINESS NECESSITY
When an employee makes a showing of disparate impact, 

the burden shifts to the employer to prove business necessity. 
E.g. Teal, 457 U.S. 440, 446; Dothard, 433 U.S. 321, 329; 
Albemarle, 422 U.S. 405, 425.

The employer’s burden is one of persuasion rather than 
production, for the Court has twice held evidence which would 
qualify as an “articulation” in a treatment case fails as proof of 
business necessity in an impact case. First, in Griggs, the em­
ployer offered testimony from a vice president to the effect the 
challenged transfer “requirementjs] were instituted on the 
Company’s judgment that they generally would improve the 
overall quality of the work force,” but the Court held it insuf­
ficient to establish a “demonstrable relationship to successful 
performance.” Griggs, 401 U.S. 424, 431. Second, in Albe­
marle, the Court held an employer could not meet its burden 
simply by saying it validated an exam, since,

[N)o record of this validation was made. Plant officials 
could recall only the barest outlines of the alleged val­
idation. Job relatedness cannot be proved througn vague 
and unsubstantiated hearsay.

Albemarle, 422 U.S. 405, 428 n. 23 (emphasis added). Simi­
larly, the Court has repeatedly described the employer’s bur­
den as in essence one of persuasion.49

" T e a l , 457 U.S. 440,446 ("employer must... d e m o n s tr a te that 'any given 
requirement (has| a manifest relationship to the employment in question’ "); 
A lb e m a r le , 422 U.S. 405, 425 (employer has “burden of p r o v in g that Its tests 
are ‘job related’ ’’); D o th a r d , 433 U.S. 321,329 (employer must “prou/e/ that 
the challenged requirements are job related”); G rig g s , 401 U.S. 424, 432 
(employers has "the burden of s h o w in g that any given requirement (has| a 
manifest relationship to the employment in question”); s e e  B e a z e r , (cont)

164



44

When it amended Title VII in 1972, Congress “recognized 
and endorsed the disparate impact analysis employed by the 
Court in Griggs,” which places the burden of persuasion on 
business necessity squarely on the employer.*0 Similarly, “[i)n 
any area where the new law does not address itself’ Congress 
“assumed that the present law”—including Griggs (Teal, 457 
U.S. 440, 447 n.8)—“would continue to govern.” 118 Cong. 
Rec. 7166,7564 (1972). With this express ratification, altering 
the burdens would undermine the clear intent of Congress. See 
Square D Co. v. Niagara Frontier Tariff Bureau, Inc., 476 U.S. 
409, 419 (1986); Patsy v. Florida Board of Regents, 457 U.S. 
496,508-09 (1982). The courts of appeals have widely imposed 
the burden of persuasion of business necessity on employers.81

Unlike its disparate impact counterpart, a prima facie case 
of disparate treatment raises a classic presumption, “re- 
quir(ing) the existence” of one fact “to be assumed” from evi­
dence of another until rebutted. J. Weinstein and M. Berger, 1 
Weinstein's Evidence Para. 300(011 p. 300-1 (1988); see Bur- 
dine, 450 U.S. 248, 254. The employer’s “articulation” is a 
“negative” defense, “merely controvert(ing) plaintiffs prima 
facie case.” See J. Moore, 2 A Federal Practice Para. 8.27 (4j p. 8- 
193 (1987). By contrast, the employer’s burden of busines 
necessity, is an “affirmative defense,” since it "constitut(es) an 
avoidance” (Fed. R. Civ. Pro. 8(c)) or “raises matter outside the * 8

49(cont) 440 U.S. 568, 587 (prima facie case “rebutted by (employer's! 
d e m o n s tr a t io n  that its narcotics rule ... ‘is job related ") (emphasis in each 
added). The Court has cited with approval court of appeals decisions placing 
the burden of persuasion on employers (M c D o n n e l l  D o u g la s , 411 U.S. 792, 
802 n.14)— namely, C a s tr o  v. B e e c h e r , 459 F.2d 725, 732 (1st Clr. 1972) 
(employer “must come forward with c o n v in c in g  f a c t s establishing a fit be­
tween the qualifications and the job”);Chance u. B o a r d  o f  E x a m in e r s , 458 
F.2d 1167,1176 (2nd Cir. 1972) (employer bears "a h e a v y  b u r d e n of justifying 
its contested examinations”) (emphasis in each added).

*°Teat 457 U.S. 440,447 n.8; s e e  a ls o S. Rep. No. 92-415 p. 5 (1971); H.R. 
Rep. No. 92-238 P- 8 (1971).

8 ,E .g . L e w is  v. B lo o m s b u r g  M ills , In c ., 773 F.2d 561, 571 (4th Cir. 1985); 
M o o r e  v. H u g h e s  H e lic o p te r s ,  In c ., 708 F.2d 475,482 (9th Cir. 1983); J o h n s o n  
v. U n c le  B e n ’s, In c ., 657 F.2d 750, 753 n.3 (5th Cir. 1981), c e r t  d e n ie d , 459 
U.S. 967 (1982); b u t  s e e  C r o c k e r  v. B o e in g  C o. ( V e r to l  D iv  ) , 662 F.2d975,991 
(3rd Cir. 1981).

165



45

scope of plaintiffs prima facie case.”52 J. Moore, 2A Federal 
Practice j|8.27|4) p. 8-193 (1987); accord Guardians Assn v. 
Civil Service Commission of the City of New York, 463 U.S. 582, 
598 (1983) (White, J.) (employer “bear(sj the burden of proving some 
‘business necessity’ ” as “affirmative defense”) (Title VI case) 
(emphasis added). A prima facie case of disparate impact does 
not make the existence of business necessity more or less likely, 
so it does not create an inference for the employer to dispel.53 
The impact itself is the violation. Nashville Gas Co. v. Satty, 434 
U.S. 136, 144 (1977) (“|A| violation of § 703(a)(2) can be es­
tablished by proof of a discriminatory effect”). Given this, 
business necessity must be a defense which the employer af­
firmatively proves.

A party raising an affirmative defense usually bears the 
burden of persuasion on it (E. Cleary, McCormick on Evidence 
§337 p. 948-49 (3rd Ed. 1984); D. Louisell and C. Mueller, 1 
Federal Evidence, §66 p. 528 (1977)), a rule which the Court 
should apply here. Because the employer has superior access to 
the relevant proof, it is better able to bear this burden. Chance 
v. Board of Examiners, 458 F.2d 1167, 1176 (2nd Cir. 1972) 
(employer “has responsibility of designing... examinations” so 
it bears the “heavy burden of justifying” them); see also E. 
Cleary, McCormick on Evidence §337 p. 950 (3rd Ed. 1984); J. 
Chadboum, 9 Wigmore on Evidence §2486 p. 290 (1981). 
“Policy” and “fairness” dictate the same result. See Keyes v. 
School District No. 1, 413 U.S. 189, 209-10 (1973); E. Cleary, 
McCormick on Evidence §337 p. 952 (3rd Ed. 1984). Since a 
prima facie showing of disparate treatment In a non-statistical 
case is “not onerous” (Burdine, 450 U.S. 248, 253), the em­
ployer bears only the light “articulation” burden as rebuttal. 
Because a prima facie case of disparate impact usually involves 10

#JThe related B F O Q  showing of “reasonable necessity” is also an affirm­
ative defense. S e e  W e s te r n  A ir l in e s , Inc. v. C r is w e ll , 472 U.S. 400, 408-09 n.
10, 413-17 (1985) (ADEA case); T r a n s  W o r ld  A ir lin e s , Inc. v. T h u rs to n , 469 
U.S. Ill, 122 (1985) (same).

M Fed. R. Evid. 301 is irrelevant, since it "merely defines the term ‘per­
suasion,’ ” but “in no way restricts the authority of a court... to change the 
customary burdens of persuasion in a manner that otherwise would be per­
missible.” N L R B v .  T r a n s p o r ta t io n  M a n a g e m e n t  C o rp ., 462 U.S. 393,404 n.7 
(1983).

166



46

a showing of systematic effects (see Watson, 108 S.Ct. 2777, 
2789 n.3), the employer’s rebuttal burden increases accord­
ingly. 84 * * *

However suggestive, the plurality opinion in Watson does 
not compel a different result. Saying the ultimate burden of 
proof cannot be shifted to the employer (Watson, 108 S. Ct. 
2777,2790) (O’Connor, J.) does not relieve the employer of the 
burden of persuasion on an affirmative defense.88 * * * * Nor does per­
mitting an employee to “show that other tests or selection 
devices, without a similarly undesirable racial effect, would also 
serve the employer’s legitimate interest” {ibid.) suggest any­
thing contrary, for it simply gives the employee a chance to resist 
the affirmative defense by showing the challenged practice is not 
really necessary.58

Under Griggs, “(t)he touchstone jof an employer’s defense 
to a showing of disparate impact) is business necessity.” Griggs, 
401 U.S. 424,431. This defense triggers a "more probing judi­
cial review of, and less deference to, the seemingly reasonable 
acts” of employers than does the rebuttal to a showing of

84L e w is  v. B lo o m s b u r g M il l s ,  In c ., 773 F.2d 561,572 (4th Cir. 1985); M o o r e  
v. H u g h e s  H e lic o p te r s ,  In c ., 708 F.2d 475, 482 (9th Cir. 1983). V u y a n ic h i>. 
R e p u b lic  N a t io n a l  B a n k , 521 F. Supp.656,661 (N.D. Tex. 1981), v a c a te d  a n d  
r e m a n d e d  o n  o th e r  g r o u n d s , 723 F.2d 1195 (5th Cir. 1984), c e r t  d e n ie d , 469 
U.S. 1073 (1984). A  statistical showing of even disparate treatment forces 
from the employer a more exacting rebuttal than a mere "articulation.” S e g a r ,  
738 F. 2d 1249, 1268-70; s e e  a ls o  T e a m s te r s , 431 U.S. 324, 342-43 n. 24 
(affirmation of "best qualified” hiring insufficient to meet proof of "system­
atic exclusion”); cf. T Y an s W o r ld  A ir l in e s , Inc. v. T h u r s to n , 469 U.S. Ill, 121 
(1985) (burden does not shift when evidence of discrimination is direct).

88-See N L R B  v. T r a n s p o r ta t io n  M a n a g e m e n t  C o rp ., 462 U.S. 393, 400-01 
(1983) (statute placing on N L R B ’s General Counsel "burden of proving the 
elements of an unfair labor practice” is consistent with rule placing on the
employer “affirmative defense" of proving "by a preponderance of the evi­
dence” its actions would have been the same "regardless of (its) forbidden 
motivation”); s e e  a ls o  W e s te r n  A i r  L in e s , Inc. v. C r is w e ll , 472 U.S. 400,408-09 
n. 10 (1985) (placing burden of proving B F O Q  on employer as affirmative 
defense consistent with leaving burden of persuasion on disparate treatment 
on employee).

88 An employee wishing to pursue disparate treatment claims may also
show pretext at this stage (s e e  T e a l, 457 U.S. 440, 447; A lb e m a r le , 422 U.S.
405, 436), for “(e)ither (the disparate Impact or the disparate treatment)
theory may... be applied to a particular set of facts.” T e a m s te r s , 431 U.S. 324,
336 n.15.

167



47

disparate treatment. Washington u. Davis, 426 U.S. 229, 247 
(1976); see also Western Airlines, Inc. v. Criswell, 472 U.S. 400, 
422 (1985) (“under a ‘rational basis’ standard” a court “might 
well consider that its ‘inquiry is at an end’ with an expert 
witness’ articulation of any ‘plausible reaso(n|’ for the employ­
er’s decision”) (construing BFOQ defense of reasonable neces­
sity in ADEA case). Proving it entails showing "a discrimina­
tory practice” is “necessary to safe and efficient job perfor­
mance”. Dothard, 433 U.S. 321, 332 n.14; see also Satty, 434 
U.S. 136,143 (employer must show “company’s business neces­
sitates” the challenged policy).

Griggs accepts the alternative showing of “job relatedness” 
through validation under EEOC Guidelines, since the regula­
tions serve the same purpose of limiting deference to the em­
ployer’s belief in the reasonableness of its own practices. See 
Griggs, 401 U.S. 424, 433 n.9. Following Griggs, Albemarle 
“clarified” the “appropriate standard of proof for job related­
ness,” holding a “validation study [was] materially defective” 
when "|m|easured against the (then current EEOC) Guide­
lines,” which were “ 'entitled to great deference’ ” as “ ‘(tjhe 
administrative interpretation of the Act by the enforcing agency.’ ” 
Albemarle, 422 U.S. 405, 431 436; see also Teal, 457 U.S. 440, 
445, 446 (test must be “shown to be job related” through 
evidence it “|has( a manifest relationship to the employment in 
question”). Since Griggs, the Court has with reasonable con­
sistency required employers to show a practice with disparate 
impact is either (1) “necessary to safe and efficient job per­
formance” (Dothard, 433 U.S. 321, 332 n.14); or (2) “job re­
lated” under prevailing validation standards in EEOC Guide­
lines.97 When it amended Title VII in 1972, Congress ratified 
Griggs, citing the employer’s need to show “overriding business

97 Only B e a z e r might be read to depart from these requirements. B e a z e r ,  
440 U.S. 568, 587 n.31 (even absent validation, business necessity is shown 
where safety and efficiency are “significantly served by— even if they do not 
require” challenged practice). But to the degree it does, it also strays from the 
expressed will of Congress. W a s h in g to n was not a Title VII case. While it 
might have applied "standards similar to those obtaining under Title VII” 
(W a s h in g to n , 426 U.S. 229, 249), it apparently did not apply Title VII stan­
dards p e r  s i .

168



48

necessity” or an “overriding reason why [thej tests (with dis­
parate impact) were necessary.” H. Rep. No. 92-238 p. 21-22 
(1971).

The district court never expressly ruled on the business 
necessity of separate hiring channels. Nor would the observa­
tions it made support a finding of business necessity.58 Literally 
without a whisper of evidence, the district court said it “would 
be required”—if faced with a prima facie case—“to find bus­
iness necessity for. . .  rehire” preferences. (App. Cert. 1:121- 
33.) But when—as here—an employer “produce(s| no evidence 
correlating” a criterion with "good job performance” or other­
wise “fail(s) to offer evidence . . .  in specific justification of it,” 
there is no basis for such a finding. Dothard, 433 U.S. 321,331; 
Satty, 434 U.S. 136,143. Here, in fact, there is evidence show­
ing the rehire preferences actually undermine “best qualified” 
hiring.” The failure of WCP and BBS to identify criteria

“ The district court wrote “|i|t is not a reasonable business practice to 
scour . . . sparsely populated, remote regions jin Alaska] for skilled and 
experienced workers.” (App. Cert 1:32.) But it took the observation verbatim 
from testimony of WCP's president (Tr. 1125), who offered it "without 
meaningful study of (the practice’s) relationship to job-performance ability” 
(G r ig g s , 401 U.S. 424,431). The observation is not cast in business necessity 
terms. It does not explain the failure to recruit non-whites from the Lower 48 
for upper-level jobs or in Alaska Native villages for unskilled or low-skill jobs 
in largely white departments. Nor does it say why— without "scouring” re­
mote areas— it is impractical to give Alaska Natives already recruited for 
menial jobs a chance to bid on desirable jobs.
“ The preferences require W C P  and BBS to re-hire past Incumbents, even 

when better candidates surface. W C P  and B B S  hired whites who could not 
meet minimum qualifications the district court endorsed. ( S e e p. 13 n.15, 
s u p r a .) They also gave preference to relatives without regard to merit (S e e p. 
38, s u p r a .) Under these circumstances, re-hire preferences simply perpet­
uate past mistakes. G ra n t, 635 F.2d 1007, 1018-19.

Both courts below applied the disparate impact analysis, since the re-hire 
preferences do not comprise a seniority system. W C P  and BBS conceded in 
no fewer than twenty-one interrogatory answers they had no seniority system. 
(Ex. 113-132; R.P.O. 132, 138-40.) The essence of a seniority system is the 
“ allotment] to employees of ever improving employment rights and benefits 
as their relative lengths of pertinent employment increase.” C a lifo r n ia  B r e w ­
e r s  A s s ’n. v. B r y a n t , 444 U.S. 598, 606 (1980). But the re-hire preferences 
here are not based on length of service, only the fact of service, for they give 
any two employees who worked in a job the preceeding season precisely the 
same right to return— even though one worked a single day and the other 
worked twenty years. (S e e Ex. A-l through A-11; Tr. 2345-46.) ^



49

actually applied precludes a finding their qualifications were 
justified by a business necessity. Rath, 787 F.2d 318, 328. 
DeFrance openly admitted he did not validate even his hypo­
thetical qualifications under EEOC Guidelines. (J.A. 470.) No 
business necessity justification was offered for nepotism. (See 
p. 38, supra.) Nor did WCP and BBS offer a particularized 
showing which would justify a finding of business necessity for 
their housing practices.*0 See Domingo u. Nefco, 445 F. Supp. 
421, 439-40. The only justification they provided for their 
messing practices was legally insufficient (See p. 36 n.36, supra.)

7. THE COURT SHOULD ALSO AFFIRM ON 
ALTERNATE GROUNDS OF DISPARATE TREATMENT

The district court ruled the employees made a prima facie 
case of disparate treatment in skilled jobs, unskilled jobs, 
housing and messing (App. Cert 1:114,1:118-19), so the suf­
ficiency of the prima facie case is no longer at issue. A ikens, 460 
U.S. 711, 714-15; Bazemore, 440 U.S. 385, 398. Since the 
challenged practices—including segregated hiring channels 
coupled with express race-labelling of jobs and bunkhouses— 
are facially discriminatory (Domingo, 727 F.2d 1429,1436), the 
shifting burden analysis doesn’t apply. Trans World Airlines, 
Inc. v. Thurston, 469U.S. I l l ,  121 (1981).Butevenifitdid,Dr. 
Rees—the labor economist for WCP and BBS—drew the infer­
ence of discrimination in certain upper-level jobs, even after 
adjusting for defense contentions on skills, labor market and 
the propriety of separate hiring channels. For these jobs, WCP 
and BBS failed to rebut the prima facie case. See Burdine, 450 
U.S. 248, 254. The labor market showing WCP and BBS 
offered for other jobs was legally insufficient, since under 
Teamsters, hiring area statistics will not rebut a disparate treat­
ment showing based on job segregation statistics. Teamsters, 
431 U.S. 324, 342 n.23. Even so, the reasons WCP and BBS 
offered for statistical disparities were clear pretext, for they 
were based on qualifications prepared for litigation and mis­
readings of their labor contracts. See Domingo, 727 F.2d 1429, * 12

“ Nearly every cannery superintendent who testified on the issue said 
workers housed in the same bunkhouse had different call-out times. (J.A. 8-
12, 227-28, 230-35.) Employees who arrived for pre-season work often 
changed bunkhouses when the season started. (J.A. 235.)

170



50

1436. Finally, since the same individuals were responsible for 
hiring, housing and messing practices (App. Cert. 1:37), a re­
versal on disparate treatment claims in hiring would necessi­
tate a reversal on such claims in housing and messing.*1 Lilly v. 
Harris-Teeter Supermarket, 720 F.2d 326,338 (4th Cir. 1983), 
cert denied, 466 U.S. 951 (1984).

CONCLUSION
The Court should affirm on all disparate impact claims, 

except re-hire preferences, as to which it should reverse the 
finding of business necessity. Alternatively, the Court should 
affirm claims of discrimination in jobs, housing and messing on 
disparate treatment grounds.

Respectfully submitted,
Abraham A. Arditi*
Bobbe Jean Bridge

•Counsel of Record *

*‘ Affirming on disparate treatment grounds under 42 U.S.C. §1981 would 
affect claims Involving Ekuk and Alitak canneries. (S e e p. 1 n. I and P- 8
n. 10.) 171



A-l

APPENDIX A-l 

TABLE F*
HIRING IN JOB DEPARTMENTS BY RACE 

AT BUMBLE BEE CANNERY 1971-80

Number of Positions Percentage
Job Department By Race By Race

W NW %W %NW
Administrative 3 1 75% 25%
Machinist 144 0 100% 0%
Company Fishing Boat 160 0 100% 0%
Tender 136 3 98% 2%
Carpenter 86 2 98% 2%
Beach Gang 49 3 94% 6%
Clerical 39 4 91% 9%
Quality Control 8 4 67% 33%
Miscellaneous 107 8 93% 7%
Culinary 112 56 67% 33%
Laborer 72 38 65% 35%
Cannery Worker 501 719 41% 59%
TOTAL 1417 838 63% 37%

This chart shows hires by race. Each year-round employee is 
counted once. Each seasonal employee is counted once for each 
season he or she was hired, regardless of whether he or she had 
been hired in that department in previous years. When a person 
worked in more than one job in a given season, he or she was 
counted once for each job he or she held.

•This table la a verbatim reproduction of Exhibit 588 (E.R. 35), which was 
offered at trial by the employees. (Tr. 2231, 2281.)

172



A-2

APPENDIX A-2 

TABLE G*
HIRING IN JOB DEPARTMENTS BY RACE 

AT RED SALMON CANNERY 1971-80

Number of Positions Percentage
Job Department By Race By Race

W NW %W %NW

Administrative 4 0 100% 0%
Machinist 117 7 94% 6%
Company Fishing Boat 152 33 82% 18%
Tender 219 2 99% 1%
Carpenter 32 0 100% 0%
Beach Gang 60 16 79% 21%
Clerical 30 4 88% 12%
Quality Control 3 0 100% 0%
Miscellaneous 107 42 72% 28%
Culinary 131 35 79% 21%
Laborer 68 154 31% 69%
Cannery Worker 180 413 30% 70%

TOTAL 1103 707 61% 39%

This chart shows hires by race. Each year-round employee is 
counted once. Each seasonal employee is counted once for each 
season he or she was hired, regardless of whether he or she had 
been hired in that department in previous years. When a person 
worked in more than one job in a given season, he or she was 
counted once for each job he or she held.

•This table is a verbatim reproduction of Exhibit 689 (E.R. 36), which was 
offered at trial by the employees. (Tr. 2231, 2261.)

173



A-3

APPENDIX A-3 

TABLE H*
HIRING IN JOB DEPARTMENTS BY RACE 

AT WARDS COVE CANNERY 1971-60

Number of Positions Percentage
Job Department By Race By Race

W NW %W %NW

Administrative 2 0 100% 0%
Machinist 102 1 99% 1%
Tender 403 13 97% 3%
Clerical 25 1 93% 7%
Quality Control 9 0 100% 0%
Miscellaneous 54 1 98% 2%
Beach Gang 0 1 0% 100%
Culinary 40 41 49% 51%
Laborer 3 0 100% 0%
Cannery Worker 874 517 63% 37%
TOTAL 1512 576 72% 28%

This chart shows hires by race. Each year-round employee is 
counted once. Each seasonal employee is counted once for each 
season he or she was hired, regardless of whether he or she had 
been hired in that department in previous years. When a person 
worked in more than one job in a given season, he or she was 
counted once for each job he or she held.

•This table is a verbatim reproduction of Exhibit 590 (E.R. 37), which was 
offered at trial by the employees. (Tr. 2231, 2261.)

174



A-4

APPENDIX III (continued)

Natives constitute 4/5 of the 2,500 minorities who would be 
available under plaintiffs’ theory (Exhibit 631; Exhibit A-406, 
Thble 34, Row "All Jobs’’), then it would take 2,000 of the 400,000 
Filipinos and Alaska Natives in the civilian labor force to fill slots 
under Dr. Flanagan’s theory. That is, .5 percent of the total 
number of Filipinos and Alaska Natives in the civilian labor force 
are "available”.

The ratio of Filipinos and Alaska Natives to Whites is .5 
percent divided by .028 percent equals 17.85. That is, Filipinos 
and Alaska Natives are approximately 18 times as likely as whites 
to take the salmon canning jobs in a freely competitiive labor 
market under Dr. Flanagan's theory.

175



B-l

SUMMARY OF STATISTICAL TESTS FOR CASE 
NEW SEASONAL HIRES, 1971-80 

SOUTH NAKNEK [BUMBLE BEE] ONLY

APPENDIX B-l

JOB HIRES ACTjUALJ % WHITE

ADMIN. 0 0.000
BEACH GANG 52 90.385
CARPENTER 53 98.113
CULINARY 28 85.714
FISHERMAN 70 100.000
MACHINIST 70 100.000
MEDICAL • 6 83.333
OFFICE 7 85.714
RADIO 1 0.000
STOR/STCK 6 100.000
TENDER 78 94.872
CANNERY 767 47.718
LABORER 77 70.130
CANRY/LAB 844 49.763
AT ISSUE 536 84.142
GEN. SKILL 159 57.233
ALL JOBS 1380 63.116

•This table is an extract of Exhibit A-278 Table 4 S N  (E.R. 4), which was 
offered at trial by the employers. (Tr. 2646-47.)

176



SUMMARY OF STATISTICAL TESTS FOR CASE 
NEW SEASONAL HIRES, 1971-80 

RED SALMON ONLY

JOB HIRES ACT1UAL1 % WHITE
ADMIN. 0 0.000
BEACH GANG 41 78.049
CARPENTER 3 100.000
CULINARY 24 79.167
FISHERMAN 35 94.286
MACHINIST 29 79.310
MEDICAL 3 100.000
OFFICE 5 80.000
RADIO 3 100.000
STOR/STCK 0 0.000
TENDER 108 96.296
CANNERY 338 35.799
LABORER 163 34.356
CANRY/LAB 501 35.329
AT ISSUE 391 81.841
GEN. SKILL 140 68.571
ALL JOBS 892 55.717

•This table is an extract of Exhibit A-278 Table 4 RS (E.R. 3), which was 
offered at trial by the employers. (Tr. 2646-47.)

B-2

APPENDIX B-2

177



APPENDIX B-3

SUMMARY OF STATISTICAL TESTS FOR CASE 
NEW SEASONAL HIRES, 1971-80 

WARDS COVE ONLY

JOB HIRES ACTIUAL1 % WHITE
ADMIN. 0 0.000
BEACH GANG 1 0.000
CARPENTER 0 0.000
CULINARY 19 47.368
FISHERMAN 4 100.000
MACHINIST 28 100.000
MEDICAL 0 0.000
OFFICE 7 100.000
RADIO 0 0.000
STOR/STCK 0 0.000
TENDER 188 94.681
CANNERY 834 68.585
LABORER 3 100.000
CANRY/LAB 837 68.698
AT ISSUE 318 89.623
GEN. SKILL 71 83.099
ALL JOBS 1155 74.459

•This table is an extract of Exhibit A-278 Table 4 W C  (E.R. 2), which was 
offered at trial by the employers. (Tr. 2646-47.)

178



0 3

Joseph Brindle, superintendent at Wards Cove, referred to 
the "Filipino [bunkhouse roofl ridge” and “the Japanese bed­
room.” (Ex. 405; R.P.O. 16, 132, 154.) Gerald Steele, office 
manager at Wards Cove, referred to "the Filipino house.” (Ex. 
404; R.P.O. 132,154; Tr. 81.) Other cannery records referred 
to the "Filipino house,” "Japanese Apts,” "Filipino Bunk 
House,” "White Bunkhouse,” “Japanese Bunkhouse” and "Fil­
ipino Bunkhouse.” (Ex. 379, 401, 402, 450; R.P.O. 132, 153- 
54.)

Personnel records at Wards Cove refer to the Native Crew 
of 1971” (Ex. 375; R.P.O. 132, 153); classify the egg crew 
workers as “Supervisor,” “Orientals,” "Girls,” “Egg Depart­
ment-Girls,” "Egg Department-Fils,” and "Egg Department- 
Eskimos” (Ex. 358,509; R.P.O. 132,151; Tr. 2279); and cate­
gorize cannery workers as "Misc. cannery workers,” “Philip- 
pino’s,” "Eskimo’s” and "Female” (Ex. 358; R.P.O. 132, 151).

181



C-2

ipinoes,” "your FUipino crew” and "four Filipinos” (Ex. 355- 
357, 362, 378, 384, 386, 409, 410, 421A, 494, 500, 511, 513, 
515,518,519; R.P.0.132,151-53,155-56; Tr. 2279); cooks for 
the non-resident cannery workers as "your Fil cook” and the 
“FUipino cook” (Ex. 403, 424, 741; R.P.O. 132, 154, 156; Tr. 
2279); and other employees as “the colored feUow,” “the 2 
Samoans,” “the 4 natives for Vems crew,” "4 of the natives” 
and “the 4 Eskimo feUows” (Ex. 373,374,377,398,415; R.P.O. 
132, 152-55).

Don Ballard, office manager at Red Salmon, referred to 
non-resident cannery workers as “the FUs” and “the Phils” (Ex. 
354, 363, 396, 417, 498; R.P.O. 15, 132, 151, 154-55); and 
Local 37, ILWU as “the Fil union” (Ex. 499; Tr. 2279); resident 
cannery workers as “24 Eskimos” and “the 24 natives Cannery 
Workers” (Ex. 418, 454; R.P.O. 132, 155; Tr. 2022-23); and 
certain employees as “four of the natives” and the “natives” 
(Ex. 372,374; R.P.O. 132,152-54). Sinularly, Ballard wrote to 
the home office,

Hardy, could you check with Mayflower press about those 
little square preprinted cards for the buttons. We should 
have had them up here before now, we got 24 Eskimos in 
yesterday and I would like to get these things made up so I 
know who they are and also to keep the other bums out of 
the Mess Hall.

(App. Cert. 1:80; R.P.O. 15.)
Forms at Red Salmon cannery contain a blank for the race of 

each employee, which is often recorded. (Ex. 520, 523-25; Tr. 
2279.)

Management at Red Salmon referred to the “Eskimo bunk- 
house,” “Native bunkhouse,” “Filipino bunkhouse” and "Fili­
pino messhall and bunkhouse.” (Ex. 84; Dep. Lessley p. 12-15.) 
Don BaUard, office manager at Red Salmon, referred to the 
“FUs bathhouse.” (Ex. 354.)

Joseph Brindle, superintendent at Wards Cove, referred to 
non-resident cannery workers as “the Filipino(s).” (Ex. 422;
R.P.O. 16, 132, 156.) Harold Brindle, an officer of WCP, re­
ferred to labor agreements for resident cannery workers as “the 
Eskimo agreements.” (Ex. 487; Tr. 2279, 2765.) Personnel at 
the WCP home office spoke of the “FUipino crews,” “native 
crews” and “Eskimo crew.” (Dep. Parrish p. 65.)

180



C-3

Joseph Brindle, superintendent at Wards Cove, referred to 
the “Filipino (bunkhouse roof] ridge” and “the Japanese bed­
room.” (Ex. 405; R.P.O. 16, 132, 154.) Gerald Steele, office 
manager at Wards Cove, referred to “the Filipino house.” (Ex. 
404; R.P.O. 132,154; Tr. 81.) Other cannery records referred 
to the “Filipino house,” “Japanese Apts,” “Filipino Bunk 
House,” “White Bunkhouse,” “Japanese Bunkhouse” and “Fil­
ipino Bunkhouse.” (Ex. 379, 401, 402, 450; R.P.O. 132, 153- 
54.)

Personnel records at Wards Cove refer to the “Native Crew 
of 1971” (Ex. 375; R.P.O. 132, 153); classify the egg crew 
workers as “Supervisor,” “Orientals,” “Girls,” “Egg Depart­
ment-Girls,” "Egg Department-Fils,” and “Egg Department- 
Eskimos” (Ex. 358,509; R.P.O. 132,151; Tr. 2279); and cate­
gorize cannery workers as “Misc. cannery workers,” “Philip- 
pino’s,” “Eskimo’s” and “Female” (Ex. 358; R.P.O. 132, 151).

181



D-l

APPENDIX D
DeFrance’s "ability to” criteria include:

“(A)bility to use mechanic’s hand tools,” "ability to use seam 
micrometers [and] gauges,” "ability to understand mechanical 
drawings,” "ability to use... pipefitter’s tools,” “(mjust be able 
to understand and accurately complete required inspection 
and report forms,” "ability to check weights, record temper­
atures, and use basic mathematics through decimals” and "abil­
ity to accurately operate ten-key calculator.” (J.A. 500-07.)

DeFrance’s subjective qualifications include:
"(Ajbility to work with minimum supervision,” “|m|ust possess 
leadership skills,” “jajbility . . .  to communicate effectively in 
English,” "ability to handle the strain, responsibility and pres­
sure,” "capable of training a machnist helper-trainee,” "me­
chanical ability,” "(m)ust be flexible, willing to learn, and (able) 
to follow directions,” "(m]ust have ability to handle details,” 
“be reliable,” “(rjequires good health,” "ability to perform 
heavy work out of doors,” "be honest,” "ability to live in small 
quarters and function as an effective member of a small group” 
and “ability to work long hours on ocean-going vessel.” (J.A. 
500-07.)

Subjective qualifications cited by lay witnesses include:
“(A| good worker,” "somebody that’s sober,” "somebody that’s 
reliable,” “good people,” "|people who] want to work,” “(ajbil­
ity plus hands, head,” "family background,” “good guy,” “gets 
along with everybody,” "motivated to do this kind of work,” 
"people that we were sure you could depend on to stay on the 
job,” “(people who are) capable,” "we tried to stay away from 
drinkers,” "not a dirty person” and “personality.” (Dep. of 
AW. Brindle-1975 29; Dep. Leonardo-1975 22; Dep. Leo­
nardo-1978 47; Dep. W.F. Brindle-1978 68-69; Dep. H. Parrish 
18; Dep. Rohrer 43; Dep. Mullis 12.)

182



No. 87-1387

IN THE

Supreme Court of the United States

October Term, 1988

W ards Cove P acking Company. Inc., 
Castle & Cooke. Inc..

Petitioners,

v.

F rank  A tonio, et al, 
Respondents.

REPLY BRIEF OF PETITIONERS

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ikkelborg, B roz. 
W ells & F ryer 
Suite 3300
1001 Fourth Avenue Plaza 
Seattle, Washington 98154 
(206) 623-5890

* Counsel of Record Attorneys for Petitioners

December 5, 1988

183



1

Reply to Respondents’ Statement of Facts...........................1

1. Hiring Practices........................................................... 2

A. Hiring for At-Issue Jobs.................................... 2

B. Local 37 .............................................................. 3

C. The "Lock-In” Argument................................... 4

D. Hiring in Alaska................................................. 5

E. Labeling.............................................................. 6

2. Labor Supply...............................................................7

3. Skills............................................................................ 8

4. Nepotism.................................................................... 10

5. Housing and Messing................................................. 11

Argument in Reply...............................................................12

I. Proof of Work Force Imbalance Is Not and Should
Not Be Dispositive Proof of Disparate Impact............. 12

A. Work Force Imbalance is Not Per Se
Discriminatory...................................................12

B. Policy Considerations Strongly Disfavor
Internal Comparisons........................................14

C. Instances of Statistical Significance..................15

TABLE OF CONTENTS

Page

184



II

II. The Causation Gap......................................................... 16

III. Respondents’ Proof Does Not Establish an
Unrebuttable Presumption..............................................16

Conclusion..............................................................................20

Appendices:

Appendix I: Percentage of New Hires From
Washington, Oregon, and Alaska: All 
Alaska Facilities of Defendants, Except 
Icy Cape....................................................A-l

Appendix II: White Percentage of Civilian Labor Force
Over Age 18 From Farwest States............ A-2

Appendix III: Number and Percentage of White and 
Minority Components of Civilian Labor 
Force in Farwest Needed to Produce 
5,000 Employees Per Year for Salmon 
Industry................................................... A-3

Appendix IV: Instances of Statistically Significant 
Underrepresentation of Nonwhites in 
At-Issue Jobs in Petitioners' Labor 
Market Analysis....................................... A-5

TABLE OF CONTENTS, (continued)

Page

185



Ill

Albemarle Paper Ca v. Moody, 422 U.S. 405 (1975)............ 16

Allen v. Prince George's County, Md., 737 F.2d 1299
(4th Cir. 1984).................................................................. 15

Anderson v. Bessemer City, 470 U.S. 564 (1985).................... 1

Bazemore v. Friday, 478 U.S. 385 (1986)............................  14

Carpenter v. Stephen F. Austin State Univ., 706 F.2d
608 (5th Cir. 1983).......................................................... 13

Christie v. Callahan, 124 F.2d 825 (D.C. Cir. 1941) .............. 17

Connecticut v. Peal, 457 U.S. 440 (1982)............. ..........  13,16

Domingo v. New England Fish Ca, 445 F. Supp. 421 
(W.D. Wash. 1977), rev'd on other issues, 727 F.2d 1429, 
modified 742 F.2d 520 (9th Cir.) (1984)........................ 7,14

Dothard v. Rawlinson, 433 U.S. 321 (1977).........................  16

EEOC v. Federal Reserve Bank of Richmond, 698 F.2d 
633 (4th Cir. 1983), rev'd on other grounds sub nom 
Cooper v. Federal Reserve Bank of Richmond, 467 
U.S. 867 (1984)...............................................................  15

EEOC v. Sears, Roebuck & Ca, 839 F.2d 302 (7th Cir.
1988)............................................................................ 12,16

EEOC v. Western Elea Co., 713 F.2d 1011 (4th Cir.
1983)............................................................................... 15

TABLE OF AUTHORITIES

Cases Page

186



Cases, (continued) Page

Furnco Constr. Corp. v. Waters, 438 U.S. 567 (1978) 13

Gay v. Waiters' & Dairy Lunchmen's Union, Local No. 30,
694 F.2d 531 (9th Cir. 1982)..........................................  15

General Electric Co. v. Gilbert, 429 U.S. 125 (1976)...........  18

Green v. USX Corp., 843 F.2d 1511 (3d Cir. 1988), petition
for cert, filed July 23, 1988), No. 88-141.........................19

Griggs v. Duke Power Co., 401 U.S. 424 (1971) . . . 16,17,18,19

Hester v. Southern Railway Co., 497 F.2d 1374 (5th Cir.
1974)..................................................................................9

James v. Stockham Valves & Fittings Co., 559 F.2d 310
(5th Cir. 1977), cert, denied, 434 U.S. 1034 (1978)......... 13

Joseph Constantine S.S. Line, Ltd. v. Imperial Smelting
Corp., Ltd., (1942) A.C. 154............................................. 18

Markey v. Tenneco Oil Co., 635 F.2d 497 (5th Cir. 1981) . . . 14

Markey v. Rmneco Oil Co., 707 F.2d 172 (5th Cir. 1983)
(after remand).............................................................14,15

McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973) . . .  17

New York Thinsit Auth. v. Beazer, 440 U.S. 568
(1979)................................................................... 13-14,18

iv

TABLE OF AUTHORITIES, (continued)

NLRB v. TYansp. Management Corp., 462 U.S. 393 (1983). 17

187



V

Cases, (continued) Page

Paxton v. Union Nat'l Bank, 688 F.2d 552 (8th Cir. 1982),
cert, denied, 460 US. 1083 (1983)................................. 13

Payne v. Travenol Laboratories, Inc., 673 F.2d 798 (5th
Cir. 1982), cert, denied, 459 U.S. 1038 (1982)...............  13

Petition of New England Fish Company, 465 F. Supp.
1003 (W.D. Wash. 1979)............................................  .9

Rivera v. City of Wichita Falls, 665 F.2d 531 (5th Cir.
1982)...............................................................................  13

Ste. Marie v. Eastern R.R. Ass'n, 650 F.2d 395 (2d Cir.
1981) .................................................................................9

Tkamsters v. United States, 431 U.S. 324 (1977).................  14

Texas Dep't of Community Affairs v. Burdine, 450 U.S.
248 (1981).......................................................................  17

Watson v. Ft. Worth Bank & Trust, 487 U.S___ _ 108
S. Ct. 2777, 101 L. Ed. 2d 827 (1988)......................  12,20

Weissinger v. United States, 423 F.2d 795 (5th Cir.
1970)...............................................................................  19

Williams v. Owens Illinois, Inc., 665 F.2d 918 (9th Cir.
1982) .............................................................................. 14

Wright v. Rockefeller, 376 U.S. 52 (1964).............................  17

TABLE OF AUTHORITIES, (continued)

188



Statutes & Regulations Page

42 U.S.C. § 2000e et seq., Civil Rights Act of 1964 12

43 U.S.C. § 1601 et seq., Alaska Native Claims Settlement
Act, 85 Stat. 688 (1986)....................................................7

29 C.F.R. 1602.14(b)............................................................... 18

Other Authorities

W. Blackstone, 3 Commentaries * 340 (1900) 18

Scanlan, Illusions of Job Segregation, 93 The Public
Interest 54 (1988).............................................................14

E. Cleary, McCormick on Evidence § 337 (3d ed. 1984). . . . 18

J. Moore, 5 Federal Practice 1 41.13(4), p. 41-179 (2d ed.
1988)...............................................................................  19

J. Buzzard, 10 Phipson on Evidence, p. 36 (12th ed.
1976)...............................................................................  18

Cleary, Presuming and Pleading: An Essay on Juristic
Immaturity, 12 Stan. L. Rev. 5, 7 (1959)........................  17

Manual for Complex Litigation (2d ed. 1986).......................  19

vi

TABLE OF AUTHORITIES, (continued)

189



IN THE
Supreme Court of the United States

October Term, 1988

W a rd s  Cove P acking  Company, Inc., 
Castle  & Cooke, Inc.,

Petitioners,

v.

F rank  A tonio, et al., 
R espondents.

REPLY BRIEF OF PETITIONERS

Respondents’ brief is dominated by two recurring themes: 
a refusal to acknowledge or appreciate the District Court’s fact­
finding role and a disregard of the principles used to measure dis­
parate impact. Respondents opt, instead, for an exhaustive re­
argument of the facts as if the clearly erroneous rule did not exist 
and they assume that mere proof of imbalance in an employer’s 
work force establishes an unrebuttable case of disparate impact.

REPLY TO RESPONDENTS’ STATEMENT OF FACTS

Respondents’ densely packed brief presents a misleading 
factual picture. Tb avoid the heavy burden placed on them under 
the clearly erroneous rule,1 respondents simply ignore the findings 
and point only to evidence favorable to their case. This case is 
before this Court, however, not on a review of a summary judg­
ment but after a trial has been conducted.
1 Anderson u Bessemer City, 470 U.S. 564 (1985). Respondents do state 

there was no evidence that rehiring a satisfactory worker in the same 
job the next season was necessary. Resp. Br., p. 14. Not even the Ninth 
Circuit agreed with respondents on this. Pet. App. 111:56, VI:32, 33.

190



2

The superficial appeal of respondents’ evidence was found 
collectively by the District Court to provide an inference of 
intentional discrimination. That inference disappeared when the 
court considered petitioners’ evidence. The error of the Court of 
Appeals occurred when it considered this same inference created 
an unrebuttable presumption if the impact model were used. 
Pet. App. VI: 18.

1. Hiring Practices.

Respondents persistently mischaracterize the practices as 
“racially segregated.”1 The label does not make it a fact. The 
trial court found the practices were not racially motivated, were 
not a pretext for intentional discrimination, and the Ninth Circuit 
affirmed on this issua1 While respondents continue to insinuate 
that petitioners used Local 37 out of racial motivation, they lost 
that argument in both courts below.

A. Hiring for At-Issue Jobs.

Respondents do not dispute that petitioners hired many 
nonwhites in at-issue jobs — including at the highest levels.1 * * 4 
Petitioners collectively employed 24% nonwhites in these at-issue 
jobs. Respondents concede the relevance of these hiring results. 
(Resp. Br., p. 8, n. 10.)

1 The term "segregate” conjures visions of racial patterns in the
southern United States. It simply does not apply to petitioners. 
Being unfamiliar with the record, the amicus briefs in support of 
respondents are infected with the same rhetoric.

* Pet. App. 1:119, 129, 130. Moreover, this Court denied respondents’ 
Petition for Writ of Certiorari (No. 87-1388) on this issue and denied 
their Motion for Rehearing. Those rulings should not be disturbed.

4 E.R. 13 (790 nonwhite hires); R.T. 2862; J.A. 159 (director, vice presi­
dent, superintendent).

191



3

Respondents’ real argument is that petitioners should have 
hired 50% minorities in at-issue jobs and should have either 
targeted Local 37 as a source for these jobs and/or trained people 
for at-issue jobs.

B. L ocal 37.

Petitioners used Local 37 as a source of nonresident cannery 
workers simply because it held the contract. R.T. 1128. Respon­
dents now contend that the union has no formal role in selecting 
employees. The facts and the findings are otherwisa The cannery 
worker foreman* first determines who has a rehire preference, and 
the remaining employees (except for egg house workers)* are 
designated by the union president. R.T. 1127,1128.

None of the respondents were channeled by management into 
cannery worker jobs. Rather, each testified that he originally 
went to Local 37 to seek employment in the canneries.5 * 7 * The Dis­
trict Court found that management does not direct any cannery 
worker foremen to line up members of any raca (Pet. App. 1:33.)* 
Respondents do not mention these important findings.
5 While the union contract does state the cannery worker foreman is 

a company representative (see Local 37 contract in Ex. A-l through 
A-ll), Dr. Rees explained that this is a common provision in union 
contracts to avoid Thft-Hartley implications; that the person named 
on management payroll was acceptable to the union; and that his 
decisions were agreeable to the union. R.T. 1967, 1968.

* Some of the egg house workers are hired by management because of 
the union's refusal to fill those positions. Perhaps as a carry-over from 
World War II, the Filipino males refuse to work in the egg house with 
Japanese nationals without payment of overtime. R.T 1128.

7 R.T. 46 (Kido); R.T. 76 (Atonio); R.T. 874 (Baclig); R.T. 956 (C. Lew);
R.T. 205 (Kuramoto); R.T. 1042 (Arruiza); R.T. 1057 (R. del Fierro); R.T. 
2222 (A. Lew). See R.T. 160 (Della); R.T. 202 (Pascua); R.T. 795 (Daba) 
(class member witnesses).

* Petitioners cite a letter from the Alitak cannery worker foreman Fred 
Wong (Ex. 394), who said he could recruit whites or blacks if directed. 
Wong was concerned because of this suit that Local 37 would supply 
too many Asians. (R.T. 3122.) In line with company policy, he was 
never directed to do so and the District Court so found. Pfet. App. 1:33.

192



4

C. The “Lock-In” Argument.

Respondents contend that cannery workers are locked in or 
tainted by the fact that they are hired as cannery workers. They 
imply that cannery workers are cut off from other jobs. However, 
at the end of the season all employees complete their jobs and 
become part of the overall labor supply available for at-issue jobs. 
The District Court found that none were deterred, that all appli­
cants were evaluated on job-related criteria, that employees and 
non-employees are free to apply for any job and that similarly 
situated applicants are treated equally.9 Many did apply, 
including respondent Atonio, and were hired.10 * The statement 
“What’s wrong with being in the Filipino crew?” quoted by 
respondents (Resp. Br. 15) was flatly denied by its alleged author. 
(R.T. 2816.) Respondents also do not mention the finding that 
there has been a general lack of interest by cannery workers in 
applying for non-cannery worker jobs. (Pet. App. 1:40.)

There are also substantial business reasons why petitioners 
do not promote during the season, and the finding (and sup­
porting evidence) that there is a lack of time to train during the 
season is unchallenged. (Pet. App. 1:19,34,46.) Thus, it is clear 
that both by necessity and by choice these employers do not 
promote from within.11

9 Pet. App. 1:33, 122, 123.
10 Filipinos and other minorities are not “channeled” into Local 37 

cannery worker jobs. Those who actually apply for at-issue jobs are 
hired in those jobs when qualified for an existing opening, rather than 
being sent to Local 37. See, e.g., R.T. 2715-16; J.A. 159-60; R.T. 2889; 
J.A. 614; R.T. 2833. Minorities who have worked as part of the Local 
37 crew and have applied for at-issue jobs in the off season have been 
hired in those jobs. See Pet. App. 1:88 (Peters, a/k/a Atonio); J.A. 
463-64; R.T. 2771-72.

“ Respondents argue that Ex. 614 shows promotion discrimination, but 
the trial court found otherwise Moreover, this exhibit only accounted 
for a minute portion of the jobs filled at the five canneries combined 
and did not address the issues presented.

193



5

D. Hiring in Alaska.

Respondents argue that petitioners targeted villages in 
Alaska for racial reasons.11 The hiring patterns in Alaska were 
dictated by geography, as the District court found. (Pet. App. 
1:38, 39.)11 The bush country in Alaska, which is dominated by 
Alaska Natives, is an area of no roads and no telephones; it is 
sparsely populated; and there is no public transportation other 
than air strips (J.A. 481-485). Bush pilots may have the ability 
to hire unskilled workers but have no idea as to the requirements 
for other jobs (R.T. 1125). Petitioners also tap Alaska sources for 
cannery workers that because of geography are heavily white, e.g., 
the cities of Kenai and Ketchikan and the Air Force base at King 
Salmon in Bristol Bay. (Pet. App. 1:38-39; E.R. 16; J.A. 617.)

The companies hired relatively few employees for any at-issue 
jobs from Alaska (probably because of its distance from the hiring 
centers in Seattle and Astoria), but to the extent they did, Alaska 
Natives received the vast majority of the jobs. Pet. Br., p. 8, n.13. 
Certain areas of Alaska do provide skilled workers, such as the 
native boat building community on Kodiak Island which is tapped 
by the CWF Port Bailey and Alitak canneries. (R.T. 1126.)M
11 Resp. Br., p. 7.
'* Respondents make much of the recruitment of unskilled workers in the 

Alaska Native villages and imply that the practice had a significant 
role in recruiting new cannery workers. Resp. Br., pp. 7-10. In fact, 
Wards Cove never used the practice; at Red Salmon only 5% of the 
new cannery workers hired 1971-80 were Alaska Native (17/338); and 
at Bumble Bee (South Naknek) only 17% of the new cannery workers 
were Alaska Natives (129/767) and, of those, three-fourths were hired 
1971-73, at which point Bumble Bee decided not to use the practice 
any longer. Ex. A-403, Tbl. 1, 2, 5 row "cannery worker"; Ex. A-64 
(Bumble Bee employees, years 1971-80); See E.R. 16; R.T. 1515-1516. 
At Red Salmon 11% of the new hires in at-issue jobs were Alaska 
Natives; at Bumble Bee they were 10%. E.R. 11,12.

M Many of the Alaska Natives hired in the at-issue jobs came from the 
very villages where some of the bush pilot recruiting took place.

(footnote continued on following page)

194



6

Geography also takes some potential employees out of con­
sideration. For example, all fishermen at Ekuk cannery have been 
independent1* since 1959 (R.T. 2437). Most native fishermen prefer 
the Nushagak River in Bristol Bay because of its longer, more 
consistent run and the presence of King salmon (R.T. 2436, 2437). 
Thus, they fish at Ekuk, which is located on that river, rather 
than being available as company (employee) fishermen at Naknek 
(Red Salmon, Bumble Bee). (R.T. 1142.)'® In addition, there was 
substantial evidence that many Alaska residents prefer fishing 
to cannery work. (R.T. 2437, 2347-48, 1141-42.)

E. Labeling.
The District Court reviewed the instances of race labeling 

and found that, although not laudable, it did not deter minor­
ities in employment. (Pet. App. 1:123.) Respondents have pointed 
to every instance of labeling in the ten-year case period. Much 
of it, as the District Court found, is not as sinister as might 
appear.* 15 * 17 The “native cook” at Bumble Bee was white. (J.A. 623.)

See, e.g., Exhibit 309 (employees Nos. 2, 6, 8, 16-20, 25-26, 31, 34, 
35 are all Alaska Natives (Ex. A-65. pp. 767-770); Ex. A-382 (Tbl. 1 
Ekuk) (showing residence of 100 Alaska Natives among “all hires” 
in at-issue jobs over a six-year sample period). Indeed, respondents 
themselves presented application exhibits for at-issue jobs from 
Alaska Native villages. Exs. 693, 695. Many Alaska Natives were 
hired in the at-issue jobs at all the remote canneries. Ex. A-403, 
Tbls. 2-5, Col. "Ak. Nat."

15 Independent means they are not employees but are in business for 
themselves (R.T. 2436, 2473). Their earning potential is considerably 
higher than that of the cannery superintendedent. Id. Respondents 
omit that the letter they cite as nepotism evidence (Ex. 464, Resp. 
Br. p. 9) was referring to a job with an independent fisherman at Ekuk. 
(R.T. 2436-37.)

18 Seventeen natives fished as a group at CWF-Egegik as company 
fishermen; although hired by and counted as Red Salmon fishermen 
for accounting purposes, they do not show up in the comparative 
statistics. (R.T. 1142.)

17 Because most of the labeling was in internal company records, most 
class members would not have even seen it. Thus, respondents cited 
it as evidence of intent, not impact. Plaintiffs' (Respondents’) Final 
Argument, pp. 70-72.

195



7

Filipinos refer to themselves as such (J.A. 181); the Wards Cove 
Packing Co. president told the affirmative action representative 
(he was Filipino) to stop using the term “Filipino Bunkhouse,” 
but he continued to use it anyway (J.A. 182). The "impossible” 
Eskimos referred to cannery workers who had just gone on strike 
for wages higher than the contract during one of the biggest 
salmon runs in 30 years (R.T. 1144). Alaska Natives commonly 
refer to themselves as "native” and have so labeled many of then- 
organizations, e.g., Bristol Bay Native Corporation (J.A. 182); 
cf. Alaska Native Claims Settlement Act, 85 Stat. 688, 43 U.S.C. 
§ 1601 et seq. The "Japanese” were Japanese nationals, not 
United States citizens, employed not by petitioners but by the 
companies who purchased the salmon eggs. (R.T. 1128.)

2. Labor Supply.

Unquestionably, selecting the relevant labor market and 
resolving statistical conflicts are fact-finding functions. (Br. of 
Pfet., p. 18.)u Respondents profess shock that the District Court 
found the available labor supply is 10% nonwhite. They say that 
because half the employees hired are nonwhite, the labor supply 
must be 50% nonwhite. Resp. Br., pp. 16-18. This credibility 
argument was rejected by the District Court (Pet. App. 1:35-42, 
110-111)1* which believed Dr. Rees’ theory. See, e.g., J.A. 268. *

,8 Respondents champion appellate fact finding on the labor supply 
issua Resp. Br., p. 16 (“believed was relevant”).

18 Respondents argue that their own labor market analysis should have 
been adopted by the District Court. Resp. Br., p. 16. Plainly, there 
was substantial evidence to support the trial court's rejection of 
respondents’ theory on this hotly contested issue (see, e.g., testimony 
of Dr. Rees, J.A. 291-294; cross-examination of Dr. Flanagan, R.T. 
2065-2116). Respondents omit the fact that Dr. Flanagan’s theory 
was also rejected in the Domingo case that is cited so often in their 
brief. Domingo u New England Fish Ca, 445 F. Supp. 421, 433 (W.D. 
Wash. 1977). See Resp. Br., p. 29.

196



8

Dr. Rees did not “drop” the percent nonwhite from a starting 
point of “47% to roughly 10%.” Resp. Br., p. 18. More appro­
priately, he started with an unweighted civilian labor force that 
was from 92% to 96% white and, if anything, his adjustments 
raised, not lowered, the nonwhite availability percentage.20 (J.A. 
268-271; R.T. 1878.) The logical conclusion is that nonwhites, 
particularly Filipinos and Alaska Natives, receive dispropor­
tionately more, not less of available job opportunities.21

3. Skills.
Respondents contend that there should have been no skills 

adjustment to the statistics since (1) the jobs were unskilled and 
(2) petitioners’ qualifications were subjective, and not shown to 
have been imposed. Resolution of this question does not change 
the result because the District Court found that the unskilled 
labor market was 90% white. (Pet. App. 1:37.)22
20 Petitioners hired 83% of their employees from three states (Reply Br. 

App. I) where the civilian labor force is 96.3% white; it is 92% white 
if California is included. Reply Br. App. II. Moreover, Filipinos and 
Alaska natives, who combined were 88% of the class members (Ex. 
A-476), only made up a total of 4% of the civilian labor force in the 
overall geographical area drawn on by petitioners. See Reply Br. App. 
Ill; Ex. A-278-A-281, Labor Pool Tbls. 5b, 6b, 7b, 8b (“Everyone in 
Work Force”), CoL “All Depts.” They were less than 2% of the popula­
tion of Alaska, Washington, and Oregon combined. J.A. 295, 296; 
Ex. A-35, Tbl. 17 (at p. 32); Ex. A-36,Tbl. 17 (at pp. 39-41,42); Ex. A-37, 
Tbl. 17 (at pp. 39-45, 46).

21 Despite these statistics, respondents still argue that white and non­
whites (i.e., Alaska Natives and Filipinos primarily), are available in 
equal numbers for all of the at-issue jobs. This rejected claim assumes 
that nonwhites are eight times more likely than whites to make 
themselves available for a job in the salmon canning industry and 
Filipinos and Alaska Natives combined are / 7 times more likely than 
whites to be available for such jobs. See Reply Br. App. III.

22 The statistical labor market analysis incorporating the skills adjust­
ment in the tables preferred by Dr. Rees is set forth in E.R. 2-9. In 
nine out of 13 at-issue job families (including at-issue combined), 
nonwhite availability was 10% or higher after making the skills 
adjustment. E.R. 8-9. Also see Ex. A-278, Tbl. 5, for each tab, cited 
at Pet. Br., p. 20, n. 29.

197



9

Nevertheless, both arguments are wrong and very carefully 
omit the fact that numerous witnesses testified and the court 
found that most of the at-issue jobs require prior skill and ex­
perience, whether expressly articulated or not, which petitioners 
sought in evaluating applicants.23 The isolated instances of lesser 
qualified persons being hired cited by respondents (Resp. Br. 13) 
were not sufficient to overcome the evidence that supported the 
Court’s findings.24 Ste. Marie v. EasternR.R. Ass'n, 650 F.2d 395, 
401, n.6 (2d Cir. 1981); Hester v. Southern Ry. Ca, 497 F.2d 1374, 
1379, n. 6 (5th Cir. 1974).25

The court also found that the at-issue jobs were not fungible 
with the cannery worker jobs, that they required skills not readily 
acquired on the job, that petitioners needed experienced personnel 
in the jobs and hired on that basis, and that nearly all of the
23 Pet. App. 1:35, 45-47, 55-76, 112-123 (FF 104, 124, 126,130,134).

See, eg., R.T. 636, Ins. 21-24; R.T. 1013, 1019-21; R.T. 2181-82; R.T. 
2314-22; R.T. 2358-62. 2368-69; R.T. 2434-40; R.T. 2541-44; R.T.
2553-54; R.T. 2559-65; R.T. 2605-12, 2620-24; R.T. 2625-28; R.T.
2636-41; 2643-44; R.T. 2715-17; R.T. 2736-52; R.T. 2771-72; R.T.
2854-64; R.T. 2887-91, 2898-99; R.T. 2941-74, 2987-3002; R.T.
3153-58; Dep. of Robertson, pp. 3, 13-21 (following R.T. 3151); R.T. 
3208-15; R.T. 3235-69; R.T. 3311-13, 3316-20; R.T. 1114-35, 1141-42, 
1144-50; R.T. 1513-15.

24 Respondents cite testimony of two tender engineers (J.A. 19-24, 
60-62) but omit their cross-examination and other evidence. They 
worked on the one tender that was the best maintained in the fleet, 
and did not venture far from the cannery. One had significant prior 
skill, including engine overhauls. (R.T. 126-136; see also R.T. 780-790, 
915-919). They also cite the "nephew” (Resp. Br. 13) who became a 
seamer machinist. His qualifications are shown at R.T. 705-740.

23 Respondents’ primary skills evidence was an industrial psychologist 
(Latham), who demonstrated a complete lack of knowledge of the re­
quirements to operate a salmon cannery. For example, he opined that 
a tender could be safely and efficiently operated with a crew consisting 
of one skilled captain and three green deckhands (R.T. 2144), a crew 
held by another court to be unseaworthy. Petition of New England 
Fish Company, 465 F. Supp. 1003 (W.D. Wash. 1979). The District 
Court rejected his testimony at the close of evidence. (R.T. 2042.)

198



10

at-issue jobs required pre-existing skill and experienca Pet. App. 
1:35; 45-75.”

Mr. DeFrance did not present “an entirely different set of 
qualifications” (Resp. Br. 12), but gave his opinion as to the skills 
and requirements necessary to do the jobs. That opinion (R.T. 
2948-2974; J.A. 470-576) was evidence serving to corroborate 
petitioners’ evidence and the District Court’s findings that prior 
skill and experienca pre-season availability and fluency in English 
are all required. This was also corroborated by evidence in Exs. 
68-72, petitioners’ interrogatory answers on qualifications, intro­
duced by respondents.17

Respondents complain that the District Court did not state 
what “kind” of experience or skill was required for a number of 
jobs. Resp. Br., p. 12, n.13. This is a strained reading of the 
opinion. It is only common sense that when the judge (or a 
witness) says the carpenter job needs "substantial prior skill and 
experience,” he means skill and work experience in carpentry, not 
as a bookkeeper, cook, or engineer.

4. Nepotism.
Respondents grudgingly concede that their nepotism tables 

involve double counting (Resp. Br. 38, n. 39). However, that is 
only one of the many flaws the District Court could consider in 
rejecting them (Pet. Br. 26), none of which respondents rebutted. 
There were other flaws, including counting relationships with 
non-employees, such as independent fishermen. Ex. 604. 18

18 Although the A.C.L.U. does not agree (Br. p. 23), the unrebutted 
evidence was that prior truck driving experience was required for the 
set net pickup driver. Ex. 68, 71; Pet. App. 1:108; R.T. 2965. Tb be 
"readily acquirable,” the skill must be able to be learned within a 
matter of days. R.T. 1129, 1 58.

*7 E.g., Ex. 68 (qualification for a commercial fisherman is "experience ’; 
for machinist is "skilled machinists and mechanics, welders and pipe

(footnote continued on next page)

199



11

5. Housing and Messing.

Respondents refuse to accept the District Court’s finding that 
housing was not assigned on the basis of race, but on job crew 
and time of arrival. Resp. Br., pp. 19-20, 35-37.”

Respondents also refuse to accept the finding that employees 
ate their main meals with their own crew and were not grouped 
only with persons of their own race. Resp. Br. pp. 20-21. Their 
citations solely to their own evidence of labeling and isolated 
instances or anecdotes are nothing more than an attempt to prove 
pretext. Id., in which they failed. Pet. App. 1:119. * *•

fitters”; for carpenters is "skilled carpenters.. .”); Ex. 69 (qualifica­
tions for port engineer are “journeyman training and experience. . 
for bookkeeper is "knowledge of accounting required”; for iron chink 
man, "mechanical experience required”); Ex. 71 (beach boss requires 
"experience and training in all phases of rigging, dock instruction, 
pile-driving, heavy equipment, equipment operation, salvage and 
maintenance”; for a baker, "considerable baking experience. ..”)

*• There is substantial evidence to support the trial court’s findings. 
Workers are generally housed according to job department and time 
of arrival. Pet. App. 1:83 (F.F. 149(a)); R.T. 1137-38,11 79-84, 86; R.T. 
2832. 1 20; R.T. 3273, H 11, 13; R.T. 3275,1 22; R.T. 2891,1 20; R.T. 
3168,1 17. Minorities and whites in the same job categories are not 
segregated by race, but are housed together. R.T. 1137-38. The 
Filipino workers did not want their crew to stay with the Alaska 
Natives because they thought they were dirty and did not keep their 
quarters clean. (R.T. 2578.) See Ex. A-97-100. Workers in the at-issue 
jobs who arrive in the pre-season are housed together regardless of 
race. E.g., R.T. 3215-16,1 36; R.T. 2360,1 11; R.T. 2378-79,11 40-44;
R.T. 754-55; see R.T. 3333,11 56-57; R.T 2440; R.T. 2056; R.T. 3308. 
Minority and white male cannery workers are also housed together.
R.T. 1082-83; R.T. 824, 827; R.T. 250, In. 24 to R.T. 242, In. 20. Simi­
larly, minority and white female cannery workers are also housed 
together, rather than separately. E.g., R.T. 2212; R.T. 700; R.T. 701.

200



12

ARGUMENT IN REPLY

I. Proof of Work Force Imbalance Is Not and Should Not Be 
Dispositive Proof of Disparate Impact.

A. Work Force Imbalance is Not Per Se Discriminatory.
Respondents contend that work force imbalance statistics 

constitute an unrebuttable presumption of disparate impact, that 
an employer’s evidence is not relevant to the impact assessment 
and that the District Court's resolution of the conflicting evidence 
is not a function of the trier of fact; in short, they contend that 
work force imbalance is per se impact and that the work force 
itself, not the labor market, is the sole measure of discrimination.

Title VII itself recognizes that imbalance is not equivalent 
to discrimination. 42 U.S.C. 2000e-2(j); Watson v. Ft. Worth
Bank & Trust, 487 U.S.___ 108 S.Ct. 2777, 101 L. Ed. 2d 827,
843 (1988). Tb accept respondents’ position would require the 
adoption of racial quotas, in this case 50% nonwhite in every job 
department. That balance would have to be maintained with new 
hiring. The employers could not draw from the external labor 
market without skewing recruitment to make sure that for every 
white hired a nonwhite was hired. This is not what Congress 
intended. Id. at 844.

Respondents simply ignore precedent holding that internal 
work force comparisons are to be rejected in favor of credible labor 
market evidence.1* They mention neither the decisions of this 
Court and the circuit courts requiring consideration of petitioners’ 
rebuttal evidence before determining the impact,* 30 nor those 
holding that it is a fact-finding function of the trial court to 
determine the relevant statistical comparison.31
10 Cases cited at Pet. Br. p. 20.
30 Cases cited at Pet. Br. p. 17 and n. 23
31 Cases cited at Pet. Br. p.20-21 and n. 31. Accord, EEOC v. Sears, 

Roebuck A Ca, 839 F.2d 302, 309-10 (7th Cir. 1988) (“especially where 
statistical evidence is involved, great deference is due the district 
court’s determination .’’).

201



13

This is not a case like Paxton v. Union N atl Bank, 688 F.2d 
552 (8th Cir. 1982), cert denied, 460 U.S. 1083 (1983); or James 
v. Stockham Valves and Fittings Co., 559 F.2d 310 (5th Cir. 1977), 
cert denied, 434 U.S. 1034 (1978), where the employer trains and 
promotes from within.” Nor is this case like Carpenter v. Stephen
F. Austin S t Univ., 706 F.2d 608 (5th Cir. 1983), where the 
employer intentionally assigned minorities to low level jobs which 
they could not change. 706 F.2d at 623-25.

In spite of respondents’ incantations about “channeling,” 
the facts here are that the at-issue jobs are filled first over a 
period of several months from an external labor market which 
includes all of the persons subsequently hired as cannery workers. 
Thereafter, the Local 37 dispatch is utilized.

Nor can respondents find solace in Thai* 33 which is directly 
at odds with their position. Resp. Br., pp. 21, 27. There, one stage 
of a multi-component test was shown to have impact and the 
defense of "bottom line” statistics rejected. Here, respondents 
cannot show any single causal effect, yet seek to challenge the 
entire selection system on their own “bottom line” statistics, while 
asserting petitioners cannot respond in kind.34 * *
” James is explained by the Fifth Circuit in Rivera u City of Wichita 

Falls, 665 F.2d 531, 541, n. 16 (5th Cir. 1982). Payne v. Travenol 
Laboratories, Inc., 673 F.2d 798 (5th Cir. 1982) cited at Resp. Br., 
p. 27, n. 26, simply held that crediting plaintiffs’ applicant flow 
statistics was not clearly erroneous. 673 F.2d at 823. But Payne also 
rejected "bare work force statistics” on plaintiffs’ claim of dis­
crimination in initial job assignments. Id. at 824-25. Using these 
decisions is an attempt to pound a square peg into a round hole.

33 Connecticut u Thai, 457 U.S. 440 (1982). The United States in its brief, 
p. 22, suggests that certain multi-component selection devices may 
be challenged or defended as a whola

34 The remaining cases cited by respondents are inapposite. Furnco 
Constr. Ca u Waters, 438 U.S. 567, 579 (1978), is taken out of context.
There, the court stated an individual claim could not be rebutted by 
proportional representation statistics. Neither New York Transit

(footnote continued on next page)

202



14

Respondents contend that the District Court made errors of 
law in arriving at its findings, but then they proceed to argue 
the facts. Resp. Br., pp. 30-31. They state that there are cases 
where recruitment from heavily white sources can distort the 
labor market,35 but cannot show distortion in this case. Dr. Rees 
used the large population areas from which petitioners draw their 
sources. The trial court’s acceptance of this careful analysis is 
a factual finding.

B. Policy Considerations Strongly Disfavor Internal 
Comparisons.

Without authority to support their position, respondents 
urge that comparative statistics should be preferred over labor 
market data on the ground that they afford "certainty, simplicity, 
and ease of use.” (Resp. Br. 21, 29.)

Simplicity may lead to clearly erroneous conclusions because 
the internal work force comparisons tell us nothing about appli­
cant flow or qualifications.3* It leads to rigid quota-based hiring 
and leaves the employer in total command of the standard by 
which his practices will be measured. * 38

Authority v. Beazer, 440 U.S. 568 (1979), nor Bazemore v. Friday, 478 
U.S. 385 (1986), dealt with the factual situation here, nor did they hold 
a labor market analysis was irrelevant. 7hamsters involved fungible 
jobs and non-whites were excluded from ona 431 U.S. 324 (1977).

38 Domingo v. New England Fish Co., 445 F. Supp. 421 (W.D. Wash. 
1977), rev'd on other issues, 727 F.2d 1429, modified 742
F.2d 520 (9th Cir.) (1984); Williams v. Owens Illinois, Inc., 665 F.2d 
918 (9th Cir. 1982); Markey v. Tknneco Oil Co, 635 F.2d 497 (5th Cir. 
1981). Domingo was a treatment case where because of that 
employer’s practices the findings went the other way. The District 
Court was, however, obviously troubled by institutional factors. 445
F. Supp. at 433. The Ninth Circuit in this case noted the limited 
applicability of Domingo. (Pet. App. 111:18-19, 29-31.) Williams and 
Markey confirm that the relevant labor market is a fact issua 707 
F.2d 172, 175 (5th Cir. 1983) (after remand).

38 One author illustrates how comparative statistics can either hide 
discrimination or show it when none exists. Scanlan, Illusions of 
Job Segregation, 93 The Public Interest 54 (1988).

203



15

Respondents say (Resp. Br., p. 29) that this just means 
employers will discriminate — they miss the point. The vice of 
the imbalance theory is that it allows such behavior as long as 
the work force is in balance. On the other hand, if the employer 
exceeds the balance for minorities in one job classification, he 
does so at his peril. The imbalance theory in effect punishes the 
socially responsible employer, offers no independent, objective 
external measure of his practices, and leaves no room for 
traditional management prerogatives. This is an extraordinarily 
high price to pay for a theory that promises only “certainty, 
simplicity, and ease of use."37 38

C. Instances of Statistical Significance.

Respondents contend that by isolating four job family evalua­
tions from the whole,38 the statistical significance of those groups 
establishes a prima facie case. Resp. Br., 17, 25.39 The four job 
groups have been selected out of 138 (Reply Br., App. IV), but 
pure chance could account for this since one would expect two 
standard deviations in seven out of 138 categories (5%) — even 
from a nondiscriminatory employer. R.T. 1725, 137 (Dr. Wise).

But there were reasons other than chance to explain such de­
viations. The Court found a lack of cannery worker interest in
37 "Many a promising theory founders on the facts. . . and those in our 

record demonstrate that word-of-mouth recruitment at this refinery 
has not proved a discriminatory practice” Markey, supra, 707 F.2d 
at 174.

38 Tfender jobs at Red Salmon; machinist and fishermen at Bumble Bee; 
the tender jobs for Wards Cove and Red Salmon combined ("at WCP”).

38 Statistical significance is not the equivalent of legal significance. 
EEOC v. Federal Reserve Bank of Richmond, 698 F.2d 633, 648 (4th 
Cir. 1983); Gay v. Waiters' & Dairy Lunchmen's Union, Local No. 30, 
694 F.2d 531, 551, 552-55 (9th Cir. 1982) (waiter job). 
See Allen v. Prince George’s County Md., 737 F.2d 1299, 1307 (4th 
Cir. 1984) (no liability where four out of eight classifications exceeded 
two standard deviations); EEOC v. Western Elec. Co., 713 F.2d 1011, 
1019-23 (4th Cir. 1983) (age case).

204



16

applying40 and students (often cannery workers) were not avail­
able. There was substantial evidence that Alaska Natives pre­
ferred fishing to cannery jobs, including tender jobs (R.T. 2437, 
2772; J.A. 163-4). The Native fishermen in Bristol Bay for the 
most part preferred to fish at Ekuk rather than Bumble Bee 
(J.A. 179).4'

II. The Causation Gap.

Respondents concede an inability to prove separate causa­
tion. They assert that petitioners admitted causation with the 
“asserted” qualifications and the use of Local 37 for cannery 
worker jobs. But the petitioners’ labor market data showed that 
elimination of the “skills adjustment” could not change the result. 
(See supra, p. 8.) Neither Local 37 nor hiring from the villages 
could have any effect on at-issue jobs.41 * Respondents’ contention 
that proof of causation is not required (Resp. Br. 41) is directly 
contrary to existing case law. (Pet. Br. 30-33.)

III. Respondents’ Proof Does Not Establish an 
Unrebuttable Presumption.

In respondents’ view, an inference of disparate impact can 
only be defended by proving business necessity as an affirmative 
defense. They rely on Griggs and its progeny.43 The problem in
40 This is an absolute defense to segregation claims. Cases cited in Equal 

Employment Advisory Council Am. Br. p. 14. See EEOC v. Sears, 
Roebuck & Co., supra, n.31 (such lack of interest undermined statistics).

41 If the native fishermen at CWF-Egegik are added to the Bristol Bay
pool for all facilities, the nonwhite percentage of company fishermen 
rises to 26.63% (Ex. A-403, Thble 22). It is logical that they are 
included. (J.A. 179-80; R.T. 1142, 2861.)

43 Respondents do argue that Local 37 has no control over the dispatch. 
The facts and findings are otherwise, but it is not relevant in an impact 
case involving other jobs in any event.

43 Griggs v. Duke Power, 401 U.S. 424 (1971); Dothard v. Rawlinson, 433 
U.S. 321 (1977); Connecticut v. Teal, 457 U.S. 440 (1982); Albemarle 
Paper Co. u. Moody, 422 U.S. 405 (1975).

205



17

automatically applying such cases at all is that in each the 
plaintiff had proved a discrete, single, objective, facially neutral 
employment practice to have actually caused the disproportionate 
exclusion of the protected group.

Though denominated a prima facie showing, the Griggs proof 
was much more than that. The prima facie case is normally the 
threshold of evidence which will permit, but not require, the trier 
of fact to find in plaintiff's favor. Wright v. Rockefeller, 376 U.S. 
52 (1964). See Christie v. Callahan, 124 F.2d 825, 827, 840 (D.C. 
Cir. 1941). In Burdine,44 * * this Court raised the threshold one step 
further, and held that the McDonnell Douglas4S elements of proof 
in a disparate treatment case would create a rebuttable pre­
sumption — provided the evidence is believed.48

Courts have not extended prima facie evidence to the level 
respondents suggest — an unrebuttable presumption which limits 
the only defense to an affirmative one.47

In each of the decisions of this Court which put an employer 
to the proof of business necessity, not only had the plaintiff 
established that an objective, rigidly applied head wind had 
automatically been applied, but the employers basically admitted 
it. They chose to defend on the ground that the head wind was

44 Thxas Dep't of Community Affairs v. Burdine, 450 U.S. 248 (1981).
48 McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973).
48 This Court identified the two distinct burdens. 450 U.S. at 254. The 

creation of the rebuttable presumption does not arise until the plaintiff 
establishes the essential elements of his McDonnell Douglas case. 
Belief and credibilty still remain a function exclusively of the trier 
of fact.

47 The plaintiff must always establish the elements of his case. See
Cleary, Presuming and Pleading: An Essay on Juristic Immaturity, 
12 Stan. L. Rev. 5, 7 (1959). NLRB u. TYansp. Management Corp., 462 
U.S. 393 (1983), cited by respondents for the proposition that the court 
may change the burden of persuasion on an issue, expressly holds the 
plaintiff must prove the elements of his casa

206



18

justified, as in Griggs, where the employer thought it would 
simply improve the overall work force.4®

Respondents’ case rests on far weaker grounds; they cannot 
even prove causation and urge the Court to place that burden 
on the employer also. There are fundamental and sound reasons 
why the ultimate burden of persuasion should remain with the 
plaintiff. It is based on the principles that (1) he who seeks to 
change the present state of affairs should bear the risk of non­
persuasion; and (2) he who denies a fact cannot normally produce 
any proof.49 Here, the respondents seek both to change the state 
of affairs and to require petitioners to prove the negative of several 
possible causes of impact: (1) failure to recruit; (2) failure to post;
(3) failure to use objective criteria; and (4) failure to promote. 
(Resp. Br. 25, 40.)

Respondents and their amici also contend that the burden 
should be on the employer to prove lack of causation on grounds 
that the employer normally maintains the employment records.50
49 In New York Transit Auth. v. Beazer, 440 U.S. 568, 585 (1979), the 

employer defended by attacking plaintiffs’ statistics as well.
49 E. Cleary, McCormick on Evidence, § 337 (3d Ed., 1984); J. Buzzard, 

10 Phipson on Evidence 36 (12th Ed. 1976); Joseph Constantine S.S. 
Line, Ltd. v. Imperial Smelting Corp., Ltd. (1942| A.C. 154.

It is an ancient rule and founded on considerations of good sense 
and should not be departed from without strong reasons.

Id. at 174. The rule is traced to the era when if a trial by combat 
ended in a draw, the plaintiff lost. W. Blackstone, 3 Commentaries, 
*340 (1900).

50 They also contend that certain application records were destroyed. 
Petitioners routinely disposed of applications until two years into 
the case period, when plaintiffs requested them for the first time.
R.T. 1143. They were then retained but were not useful since most 
of them did not contain the race of the applicant. E.g., Ex. 693. 
Since respondents are involved in a seasonal industry, they are 
specifically exempt from keeping such records for even six months. 
29 C.F.R. 1602.14(b). Nor do the Guidelines have the force of law 
as respondents contend. General Electric Co. v. Gilbert, 429 U.S. 
125, 141-42 (1976).

207



19

The court in Green v. USX Corp., 843 F.2d 1511 (3d Cir. 1988), 
Petition for Cert, filed (No. 88-141), states that the employer is 
in the best position to know. But the same logic can be applied 
to any tortfeasor, he who collides with a car is in the best position 
to know if he is negligent but should he have the burden of proving 
no negligence? The concept is neither practical nor in accordance 
with existing law.®1

The distinction between a prima facie case and an unrebut- 
table presumption is an important one for it explains the Ninth 
Circuit’s error in this case. The Court of Appeals took the respon­
dents’ collective evidence and determined that:

The statistics show only racial stratification by job category. 
This is sufficient to raise an inference that some practice or 
combination of practices has caused the distribution of em­
ployees by race and to place the burden on the employer to 
justify the business necessity of the practices identified by 
the plaintiffs. (Emphasis supplied.)

Pet. App. V!:18. The court, having recognized that the treatment 
inference created by the totality of respondents’ evidence had been 
rebutted, determined that the same evidence was now unrebut- 
table. Only the name of the analytic model had changed.

Respondents and their amici suggest that an unrebuttable 
presumption is created at any time imbalance in the work force 
can be shown, and the burden of proving affirmative defenses ex­
tends to each practice named. This is clearly not the teaching of 
Griggs. If indeed respondents did make a prima facie showing,®* 
it was answered. 51
51 Civil discovery procedures with modern computer aided technology 

are formidable tools in the hands of a skilled plaintiff’s lawyer who 
can select the information he chooses. Petitioners spent thousands 
of hours in compiling information answering respondents’ first wave 
of discovery. Routinely, such techniques are employed in a wide 
variety of complex litigation. Manual for Complex Litigation (1986).

®* Respondents wrongly contend that the existence of a prima facie case 
cannot be challenged after denial of a motion to dismiss. Denial of 
such a motion may be discretionary, is only tentative, and does not 
preclude later findings and determinations inconsistent with it. 
Weissinger u United States, 423 F.2d 795, 797-798 (5th Cir. 1970): 
J. Moore, 5 Federal Practice § 41.13(4), p. 41-179.

208



20

The plurality of this court in Watson determined there should 
be one analytic model in discrimination cases; this case should 
set forth the elements of a single order of proof to redefine in 
discrimination cases that a prima facie case is subject to rebuttal 
before the necessity of an affirmative defense.

CONCLUSION

Respondents’ imbalance theory of impact should be rejected. 
It is contrary to law, wholly impractical, and seriously undermines 
the goals of Title VII. Without it respondents have no casa For 
the same reasons the Court should reject respondents’ business 
necessity argument and their argument that the impact theory 
can be used to challenge multiple practices without separate proof 
of causation.

Respectfully submitted,

Douglas M. Fryer* 
Douglas M. Duncan 
Richard L. Phillips 
M ikkelborg . B roz. 
W ells & F ryer

* Counsel of Record Attorneys for Petitioners

209



APPENDIX I

PERCENTAGE OF NEW HIRES FROM WASHINGTON, 
OREGON, AND ALASKA: ALL ALASKA FACILITIES 
OF DEFENDANTS, EXCEPT ICY CAPE*

State of Residence:

Alaska Oregon Washington Other
Number of
New Hires: 3356 943 3694 1610

Percent From 
Alaska + 

Washington 
+  Oregon

83.2

* This table is a summary drawn from Exhibits A-63 through A-69, 
and A-71 through A-74.

210



A-2

APPENDIX II

WHITE PERCENTAGE OF CIVILIAN LABOR FORCE 
OVER AGE 18 FROM FARWEST STATES*

Tbtal Number 
In Civilian 

Labor Force
Number of 
Whites (%)

Alaska, 
Washington, 
and Oregon 2,204,068 2,122,011 (96.3%)

Alaska 98,296 84,970 (86.4%)

Washington 1,295,958 1,246,620 (96.2%)

Oregon 809,814 790,421 (97.6%)

California 7,778,047 7,004,757 (89.9%)

Tbtal (AK, WA, 
OR and CA) 9,992,115 9,126,768 (91.99%)

* This is a s u m m a r y  drawn from Exhibits A-35 (Alaska), Thble 53, 
p.100 (over age 16); A-36 (Washington), Thble 46, p.138; A-37 
(Oregon), Tbble 46. p.129; and A-38 (California), Thble 46. p.383.

211



APPENDIX III

NUMBER AND PERCENTAGE OF WHITE AND 
MINORITY COMPONENTS OF CIVILIAN LABOR FORCE 
IN FARWEST NEEDED TO PRODUCE 5,000 EMPLOYEES 

PER YEAR FOR SALMON INDUSTRY

1. Assumed Facts: Entire civilian labor force in geographical 
area drawn on by defendants is 89% white, 11% nonwhite (non­
white: 4% Filipino and Alaska Native; 7% Asian and other 
Minority). Exhibit A-281, Labor Pool Thble 5b, Col. “All Depart­
ments".* (Unweighted, the percent white is greater than 91%. 
See Reply Br. App. II and Exhibits A-35 through A-38.) The 
civilian labor force numbers approximately 10 million persons. 
Reply Br. App. II. Thus, there are 8.9 million whites and 1.1 
million nonwhite (400,000 Filipino and Alaska Native; 700,000 
Asian and Other Minority).

The industry will employ 5,000 persons (testimony Dr. 
Smith) and the number of whites and nonwhites who would 
make themselves available for those 5,000 is equal. That is, 
Dr. Flanagan’s theory that the labor supply is 50% nonwhite is 
assumed. Thus, it will take 2,500 whites and 2,500 nonwhites to 
fill the industry jobs.

2. Comparison of White vs. NonWhite: 2,500 of 8.9 million 
whites and 2,500 out of 1.1 million nonwhites are “available". 
That is, .028 percent of whites are available, but .22 percent of 
nonwhites are available.

The ratio of nonwhites to whites is .22 percent divided by 
.028 percent equals 7.85. That is, nonwhites are approximately 
eight times as likely as whites to take the jobs in a freely com­
petitive labor market under Dr. Flanagan's theory.

3. Comparison of Whites to Filipinos and Alaska Natives 
Combined: Assuming that between them Filipinos and Alaska
* This table uses no skill differentiation (“all workers”), the broadest 

availability class (“everyone in work force”) and weighting based on 
plaintiffs’ preferred method of counting, “All Hires”. The percent white 
is slightly higher in Exhibit A-278 (“New Seasonal” Hires).

212



A-4

APPENDIX III (continued)

Natives constitute 4/5 of the 2,500 minorities who would be 
available under plaintiffs’ theory (Exhibit 631; Exhibit A-406, 
Thble 34, Row "All Jobs”), then it would take 2,000 of the 400,000 
Filipinos and Alaska Natives in the civilian labor force to fill slots 
under Dr. Flanagan’s theory. That is, .5 percent of the total 
number of Filipinos and Alaska Natives in the civilian labor force 
are “available”.

The ratio of Filipinos and Alaska Natives to Whites is .5 
percent divided by .028 percent equals 17.85. That is, Filipinos 
and Alaska Natives are approximately 18 times as likely as whites 
to take the salmon canning jobs in a freely competitiive labor 
market under Dr. Flanagan's theory.

213



A-5

APPENDIX IV

INSTANCES OF STATISTICALLY SIGNIFICANT 
UNDERREPRESENTATION OF NONWHITES IN ATISSUE 
JOBS IN PETITIONERS’ LABOR MARKET ANALYSIS*

Name of Cannery No. of At-Issue Number Showing Number Showing 
or Combination Job Families GREATER THAN LESS THAN
of Facilities lEvaluated** 1.96 Std. Dev. 1.96 Std. Dev.

Wards Cove 8 0 8
Red Salmon 11 1 10
Alitak 11 0 11
Ekuk 11 0 11
South Naknek 12 2 10
Wards Cove & 

Red Salmon 11 1 10
WCP Interests 

(Class Facilities) 12 0 12
WCP Interests 

(All Alaska 
Operations except 
Icy Cape) 13 0 13

CWF (Alitak 
& Ekuk) 11 0 11

CWF (Alaska 
Operations except 
Icy Cape) 13 0 13

Castle & Cooke 
Interests (Class 
Facilities) 12 0 12

Castle & Cooke 
Interests (All 
Alaska Operations 
except Icy Cape) 13 0 13

TOTAL 138 4 134

* This is a summary drawn from Ex. A-278, Thble 4 for each Thb.
** Job families, including "at issue combined", with one or more employees.

214



No. 87-1387

IN THE

gatprmnc (Eourt of tlit lltniteii
October Term, 1988

Wards Cove Packing Company, Inc.,
Castle & Cove, Inc.,

Petitioners,
—v.—

FRANK ATONIO, ET AL.,
Respondents.

ON WRIT OF CERTIORARI TO THE UNITED STATES 
COURT OF APPEALS FOR THE NINTH CIRCUIT

BRIEF FOR THE AMERICAN CIVIL LIBERTIES 
UNION, NATIONAL WOMEN’S LAW CENTER, 

NOW LEGAL DEFENSE AND EDUCATION FUND, 
WOMEN’S LEGAL DEFENSE FUND,

AM ICI CURIAE, ON BEHALF OF RESPONDENTS

Joan E. Bertin 
Counsel o f  Record 

Kary L. Moss 
Isabelle Katz Pinzler 
John A. Powell 
American Civil Liberties 

Union Foundation 
132 West 43rd Street 
New York, New York 10036 
(212) 944-9800

215



TABLE OF CONTENTS

Page

TABLE OF AUTHORITIES . • • • i l l

INTEREST OF AMICI CURIAE 1

STATEMENT OF THE CASE 2

SUMMARY OF ARGUMENT 10

ARGUMENT 13
Introduction 13

I. THE TOTALITY OF THE EVIDENCEINCLUDING STATISTICAL EVIDENCE OF 
RACIAL STRATIFICATION, EASILY 
ESTABLISHES A PRIMA FACIE 
CASE OF DISPARATE IMPACT 
DISCRIMINATION..................17
A. Petitioners Have Failed 

to Demonstrate that 
Observed Disparities 
Are Attributable to 
Valid Skill or Qualifi­cations Requirements . . . .  21

B. The Data Comparing Non- White Representation In 
Cannery and Non-Cannery 
Jobs, Along with Other 
Evidence, Demonstrates 
the Impact of the Employers' Practices . . . .  30

— i  —

216



Page

C. The Employers' Use of An 
Undifferentiated Hiring 
Procedure Makes the End 
Result Appropriate For a 
Measurement of Disparate 
Impact..................... 36

II. TITLE VII REQUIRES MORE THAN A 
SHOWING OF BUSINESS-RELATED 
PURPOSES TO DEFEND APPARENTLY 
DISCRIMINATORY EMPLOYMENT 
PRACTICES....................... 38
A. Congress Has Ratified ThisCourt's Requirement That 

Employers Must Demonstrate 
the Business Necessity of 
Practices Which Result in 
Discrimination.............. 41

B. Decisions of This Court 
Reveal That the Business 
Necessity Defense Is An 
Essential Element of Dis­
parate Impact Analysis . . .  46

C. Cost Considerations Do Not Establish Business Necessity 
Or Excuse The Necessity To
Prove I t ................... 56

III. PRINCIPLES OF STARE DECISISPRECLUDE ABANDONING OR ALTER­
ING THE BUSINESS NECESSITY 
DEFENSE......................... 59

CONCLUSION........................... 63

-11-

217



TABLE OF AUTHORITIES

Page

Cases
Albemarle Paper Co. v. Moody,
422 U.S. 405 (1975)..............passim
Alexander v. Choate,
469 U.S. 287 (1985).................. 45
Arizona Governing Comm, v. Norris.
463 U.S. 1073 (1983)............ 32, 58
Bazemore v. Friday.478 U.S. 335 (1986)..............passim
Beazer v. New York City 
Transit Authority.
440 U.S. 568 (1979).................. 51
Boston Chapter. NAACP, Inc, v. 
Beecher.504 F.2d 1017 (1st Cir. 1974), 
cert, denied,421 U.S. 910 (1975).............. 47, 49
Catlett v. Missouri Highway & Transp.
Comm1n. 828 F.2d 1260 (8th Cir. 1987),
cert, denied. __  U.S. __ ,
108 S.Ct. 1574 (1988)........ 16, 28, 54
Caviale v. State of Wisconsin 
Department of Health and 
Social Security,
744 F. 2d 1289 (7th Cir. 1984)........ 48

— i i i —

218



. . 54

Coble v. Hot Springs School 
District No. 6.682 F.2d 721 (8th Cir. 1982) . . .
Connecticut v. Teal,
457 U.S. 440 (1982)..............
Craig v. Alabama State Univ.,
804 F.2d 682 (11th Cir. 1986). . .
Crawford v. Western Electric Co., 
614 F.2d 1300 (5th Cir. 1980). . .
Davis v. Califano.613 F.2d 957 (D.C.Cir. 1979) . . .
De Medina v. Reinhardt,686 F.2d 997 (D.C.Cir. 1982) . . .
Dothard v. Rawlinson.433 U.S. 321 (1977).......... 40,
EEOC v. Radiator Specialty Co.,
610 F.2d 178(4th Cir. 1979).............. 24,
EEOC v. Rath Packing Co.,787 F.2d 318 (8th Cir.), 
cert. denied, 479 U.S. 910
(1986) ........................
Firefighters Institute for Racial 
Eoualitv v. City of St. Louis,
549 F.2d 506 (8th Cir.), 
cert, denied,434 U.S. 819 (1977)............

-  i  v-

passim 

. . 10

. . 27

24 , 25 

26, 28 

49, 50

29, 35 

16, 26

. . 48

219



Page

Fisher v. Procter and Gamble 
Manufacturing Co..613 F.2d 527 (5th Cir. 1980), 
cert, denied.449 U.S. 1115 (1981)........ 23, 47, 48
Franks v. Bowman Transportation Co..
495 F.2d 398 (5th Cir. 1974),rev'd. 424 U.S. 747 (1976) . . . .  14, 24
Gamble v. Birmingham Southern R. Co..
514 F. 2d 678 (5th Cir. 1975). . . 15, 22
Gathercole v. Global Associates.545 F.Supp. 1280 (N.D.Cal. 1982) . . .  59
Geller v. Markham.
635 F.2d 1027 (2d Cir. 1980),
cert, denied. 451 U.S. 945 (1981). . . 47
General Bid. Contractors Ass'n v.
Pennsylvania.458 U.S. 375 (1982).................. 32
General Electric Co. v. Gilbert.
429 U.S. 125 (1976) ............ 49, 61
Grant v. Bethlehem Steel Core..
635 F.2d 1007 (2d Cir. 1980), 
cert, denied.
452 U.S. 940 (1981).............. 47,49
Griggs v. Duke Power Co..
401 U.S. 424 ( 1971)..............passim
Grove City College v. Bell,
465 U.S. 555 (1984)..................61

-v-

220



Guardians Ass'n of the New York 
City Police Dep't v. Civil Service 
Commission.463 U.S. 582 (1983)..................60
Harless v. Duck,619 F.2d 611 (6th Cir.),cert. denied, 449 U.S. 872 (1980). . . 48
Harrell v. Northern Elec. Co.,672 F.2d 444 (5th Cir. 1982) . . .  24, 25
Hawkins v. Bounds,752 F.2d 500 (10th Cir. 1985). . . 48, 49
Haves v. Shelby Memorial Hospital,726 F.2d 1543 (11th Cir. 1984) . . 48, 58
Illinois Brick Co. v. Illinois,
431 U.S. 720 (1977).................. 60
International Brotherhood of 
Teamsters v. United States,431 U.S. 324 (1977)..............passim
Jones v. Lee Wav Freight Inc.,431 F.2d 245 (10th Cir. 1970),cert. denied, 401 U.S. 954 (1971). . . 14
Kilao v. Bowman Transp. Inc.,789 F.2d 859 (11th Cir. 1986). . . 14, 16
Kinsey v. First Regional Sec.,
557 F. 2d 830 (D.C.Cir. 1977)........ 27
Kirbv v. Colony Furniture Co.,
613 F. 2d 696 (8th Cir. 1980)........ 49

_ v i _

221



Page

Kirkland v. NY State Dep't of 
Correctional Services,
520 F.2d 420 (2d Cir. 1975),cert, denied. 429 U.S. 823 (1976). . . .  49
Kniaht v. Nassau County Council 
Service Commission.
649 F.2d 157 (2d Cir. 1981),cert. denied, 454 U.S. 818 (1982). . . .  22
Lewis v. Bloorosburg Hills. Inc.,
773 F. 2d 561 (4th Cir. 1985) ........ 50
Los Angeles Dep't of Water &
Power v. Manhart.435 U.S. 702 (1978)................52, 58
Marsh v. Eaton Corp.,639 F. 2d 328 (6th Cir. 1981)..........15
McDonnell Douglas Corp. v. Green,
411 U.S. 792 (1973).................... 49
Mobile v. Bolden.446 U.S. 55 (1980)....................61
Monell v. Deo11 of Social Services,
436 U.S. 658 (1978).................... 60
Muller v. U.S. Steel Corp..
509 F.2d 923 (10th Cir.),
cert, denied.423 U.S. 825 ( 1975)................ 22 , 35
Nash v. Consolidated City of 
Jacksonville.837 F.2d 1534 (11th Cir. 1988) . . . .  50

-vii-

222



NLRB v .  I n t e r n a t i o n a l  L o n g s h o re m e n 's
Ass'n, AFL-CIO,
473 U.S. 61 (1985).................. 60
Nashville Gas Co. v. Sattv.
434 U.S. 136 (1977).................  49
Newport News v. EEOC.
462 U.S. 669 (1983).................  58
New York City Transit 
Authority v.Beazer,440 U.S. 568 (1979).................  51
Orr v. Orr,440 U.S. 268 (1979).................  53
Parson v. Kaiser Aluminum &
Chemical Corporation,
575 F.2d 1374 (5th Cir. 1978), 
cert, denied,411 U.S. 968 (1979).................. 23
Patsv v. Florida Board of Regents,
457 U.S. 496 (1982).................. 60
Robinson v. Lorillard Corp.,
444 F.2d 791 (4th Cir.), 
cert, dismissed,404 U.S. 1006 (1971)........ 47, 49, 58
Rowe v. General Motors Corp.,
457 F.2d 348 (5th Cir. 1972) . . . 23, 48
Shidaker v. Tisch,833 F.2d 627 (7th Cir. 1986), 
cert. denied, __  U . S .  __ ,108 S.Ct. 2900 (1988)................ 35

-viii-
223



Page

Smallwood v. United Airlines Inc.,
661 F.2d 303 (4th Cir. 1981), 
cert, denied,456 U.S. 1007 (1982)............ 58, 59
Spurlock v. United Airlines. Inc.,
475 F. 2d 216 (10th Cir. 1972)........ 26
Texas Deo't of Community Affairs 
v. Burdine.450 U.S. 248 (1981).................. 56
Transworld Airlines. Inc, v. Thurston,
469 U.S. Ill (1985).................. 30
United States v. County of Fairfax.,629 F.2d 932 (4th Cir. 1980) . . . 15, 28
Walker v. Jefferson County Home,726 F.2d 1554 (11th Cir. 1984) . . 48, 49
Washington v. Davis.
426 U.S. 229 (1976).................. 47
Watson v. Fort Worth Bank & Trust Co..
487 U.S. __ ,108 S.Ct. 2777 (1988)............ 17, 56
Weeks v. Southern Bell Tel. & Tel. Co., 
408 F. 2d 228 (5th Cir. 1969)........ 51
Wengler v. Druggists Mut. Ins. Co.,
446 U.S. 142 (1980).................. 53
Williams v. Colorado Springs School 
District.
641 F.2d 835(10th Cir. 1981)............ 22, 23 , 49

-  i x -

224



Page

Statutes
Civil Rights Restoration Act,
Pub. L. 100-259 (1988).............. 61
Pregnancy Discrimination Act,
42 U.S.C. §2000e(k)................... 57
Section 504 of the Rehabilitation Act,
29 U.S.C. §794 (1982)................ 45
Title VII of the Civil Rights Act 
of 1964,42 U.S.C. §2000 et sejg...........passim
Voting Rights Act,42 U.S.C. §1973 ...................... 45

Legislative History and Other Authorities
Alaskan Fisheries Hearings: Hearings
Before the Subcommittee on Alaskan 
Fisheries, Committee on the Merchant 
Marine and Fisheries,76th Cong., 1st Sess. (1939) . . 7, 9, 10
Committee on Labor and Human Resources, 
U.S. Senate, 96th Cong., 2d Sess., 
Legislative History of the Pregnancy 
Discrimination Act of 1978 (1979). . . 58
117 Cong. Rec. 31960 ( 1971).......... 44
117 Cong. Rec. 32095 (1971).......... 44
117 Cong. Rec. 32097 (1971).......... 44

- x -

225



117 Cong. Rec. 32101 (1971).......... 44
118 Cong Rec. 7166 (1972)............ 43
House Report No. 238,92d Cong., 1st Sess. (1971 . . 43, 44, 45
Liljeblad, Filipino Alaska: A Heritage
(1980)(Alaska Historical Commission 
Studies in History No. 9)..............9
S. Rep. No. 97-417, 97th Cong., 2d 
Sess., reprinted in 1982 U.S. Code 
Cong, and Ad. News 1 ................45

Page

- x  i -

226



INTEREST OF AMICI CURIAE^
Amici curiae are non-profit legal, 

education and research organizations 
concerned about the legal rights and 
economic status of women and minority 
workers.^/

Amici believe that this case has 
potentially far-reaching implications for 
the rights secured by Title VII of the 
Civil Rights Act of 1964, as amended, 42 
U.S.C. § §2000e, et sea. ("Title VII"). In 
particular, arguments pressed by petition­
ers and the United States, as amicus 
curiae, would, if accepted, seriously 
undermine the ability of any worker who has

1/ The parties have consented to the filing 
of this brief, and the letters of consent are being 
filed with the Clerk of the Court pursuant to Rule 
36.2 of the Rules of this Court.

2 J The interest of each individual amicus 
curiae is set forth in the Appendix to this brief.

-1-

227



jected to discriminatory employment prac­
tices to receive the remedy Congress 
intended. We write especially to urge the 
Court to reject this unwarranted departure 
from established precedent.
Statement of the Case

This class action, filed in 1974, 
alleges employment practices that, 
individually and in combination, have 
created a patently racially stratified 
work environment at three Alaska salmon 
canneries. Among the elements contributing 
to this discriminatory result are (1) a 
history of job segregation; (2) recruitment 
practices which targetted non-whites for 
lower-paying jobs, while applicants for 
better jobs were sought from a predomi­
nantly white labor force; (3) rehire 
preferences, word-of-mouth hiring and 
nepotistic practices; (4) subjective hiring

2

228



practices; (5) segregation in the provision 
of housing and meals; and (6) common use of 
overt racial designations and characteriza­
tions .

The evidence reveals that non-whites 
were concentrated in lower-paying cannery 
jobs, and whites predominated in higher- 
paid positions. At Bumble Bee Cannery, 
more than 90% of all hires over a nine year 
period in seven of twelve departments were 
white. Non-whites predominated in only one 
department —  cannery worker —  and 
represented a third of hires in three other 
departments (laborer, culinary, quality 
control). Joint Excerpt of Record 
(hereinafter "ER") at 35. The same kind of 
stratification was evident at Red Salmon: 
whites obtained more than 75% of jobs in 
nine of twelve departments. Non-whites 
filled the majority of the laborer and

3

229



cannery worker positions. ER 36. At Wards 
Cove and Red Salmon, between a third and 
two-fifths of all hires were non-white, 
but they were largely confined to two 
departments —  cannery worker and culinary. 
ER 36, 37.

Even within apparently "integrated" 
departments, there was job segregation. At 
Bumble Bee, in the Fish House and Cannery 
departments, butcher and slimer jobs were 
filled exclusively by non-whites, and 
filler feeder and retort jobs were held 
almost exclusively by whites. ER 42. At 
Red Salmon, whites in the same department 
worked only in cold storage, fish weigher, 
and egg puller jobs, and non-whites held 
all other positions. ER 50; see also ER 
54. At the Warehouse department at Wards 
Cove, whites represented 48% of the hires 
(N=28), but all except one were assigned to

4

230



the warehouse job category, while non­
whites were spread throughout seven job 
categories. ER 55.

The "inexorable zero" is evident in 
the record as well. At Bumble Bee, in two 
job categories comprising 304 hires, no 
non-whites were hired at all from 1971 to 
1980, and in five other categories, with 
437 jobs at stake, non-whites obtained only 
20 positions, fewer than 7% of the total.
ER 35. At Red Salmon, no non-whites were 
hired into two job categories, involving 39 
jobs; non-whites held only 11 out of 379 
positions in three other departments. ER 
36. At Wards Cove, in six job categories 
non-whites got only 16 of 614 jobs. ER 37.

The industry has traditionally 
employed non-white laborers for the 
hardest, least lucrative positions, a 
pattern that persisted well past the

5

231



passage of Title VII. The recruitment 
practices at issue in this case are 
particularly instructive. Non-whites were 
recruited specifically for cannery work, 
although there is no apparent reason why 
the employers did not make the full range 
of employment opportunities available to 
all potential applicants. Undoubtedly, 
many of the native Alaskans and Filipino 
workers recruited for cannery work would 
have preferred other jobs, especially if 
the pay and working conditions were better. 
But the preferred jobs were not offered, 
and inquiries about the availability of 
other positions were met by a variety of 
evasive responses.—/ Higher-paying "at

/̂ For example, recruiters at Alaskan vil­lages in the remote areas near the canneries were 
not authorized to accept applications for non­
cannery work, Joint Appendix (hereafter "JA") at 
163; and non-whites were actively discouraged from 
applying. JA 38-42; 52; 56-60; 63-67; 71-73; 75- 77; 85-86; 125-126.

6

232



issue" jobs were generally filled through 
offices in Washington and Oregon, that 
drew from a predominantly white labor 
force, a practice that was common in the 
industry well before the institution of 
this lawsuit.4/ Thus, by selectively con­
trolling the labor market from which 
employees in different positions were 
drawn, the employers controlled the racial 
make-up of both the applicant pool and 
hires in various jobs.^/

Although certain skills are claimed as 
necessary for some "at issue" jobs, no

4/ Alaskan Fisheries Hearings: Hearings
Before the Subcommittee on Alaskan Fisheries, 
Committee on Merchant Marine and Fisheries,
76th Cong., 1st Sess. 58 (1939) (hereafter 
"Hearings").

5/ By this practice, if the employers' 
argument is accepted, an employer may insulate 
itself from liability for discrimination by 
manipulating the labor pool: the comparative base
used to measure discriminatory inpact is the very 
labor force selected by the employer that has 
already produced the challenged result.

7

233



skill or education requirements were ever 
actually imposed.V  The desirable skills 
were identified only in preparation for 
litigation, and many of the incumbents did 
not possess them. Subjective judgments 
thus clearly controlled the selection of 
employees for "at issue" jobs, although the 
pool from which such employees were drawn 
largely excluded non-whites.

Overt discrimination was evident in 
the housing and meals arrangements and in 
the race-typing of jobs and workers. Many 
of the employment practices have changed 
little since the days when the cannery 
owners openly embraced and espoused race- 
based practices. A report by the Alaska 
Historical Commission observed:

^  Young inexperienced whites were commonly given jobs that petitioners now claim are 
"skilled." See, e^. , JA 19-24, 25-29, 35-37, 60- 
62, 78-79, 110-11, 114-22, 123-24, 131-36.

8

234



The cannery workers were divided into 
two groups, or crews, those that 
worked inside the canneries processing 
fish and those in cannery maintenance 
and operations.... [t]he fish and 
processing crew was mainly composed of 
racial and ethnic minorities, and 
after [the arrival of] Chinese 
contract laborers, it retained a 
primarily Oriental composition for 
decades. Sometimes Caucasians might 
be found within the "China" crew, yet 
the opposite was seldom found.

U  sue Liljeblad, Filipino Alaska: A
Heritage (1980) (Alaska Historical Commission 
Studies in History No. 9) (hereafter "Historical 
Report"). Race labeling and stereotyping dominated industry practices. Chinese were valued 
as "willing to work excessively long hours without 
grumbling, and are content to live in miserable 
quarters and the cheapest food." Id. at 100. A 
representative from the Territory of Alaska 
testified in Congress in 1939: "The oriental is
not able physically to do the work of a white man, and I am sure it is the desire of the packer now to 
get rid of the Filipino.... If he has to pay a 
wage like that the packer feels he can get more 
work out of a white man and perhaps with less 
trouble...." Hearings, supra n.4, at 29. At the 
same time the white man was viewed as too good for the dirtiest work. "Some of the operators are of 
the opinion that white people would not generally 
prove satisfactory in the butchering crews. That 
is, the men who feed the iron chinks and slime the
fish__  Some of them believe it would be
necessary to use oriental labor in the butcher 
room, even if white labor were available and there 
was no restriction on employment." Id. at 346-47.

9

235



Originally shaped by intentionally 
discriminatory practices, the system 
challenged here incorporated elements of 
intentional discrimination, both covert and 
overt, along with identifiable neutral 
practices applied alike to whites and non­
whites, that served to maintain the status 
q u o .^ /

SUMMARY OF ARGUMENT 
The record in this case is replete 

with evidence that petitioners' employment 
practices have operated to freeze 
historical patterns of race discrimination. 
See Griggs v. Duke Power Co.. 401 U.S. 424

Living accommodations were distributed on the basis 
of race or ethnicity with whites receiving the newest bunkhouses with the best lighting and 
conditions. Id. at 115-16.

For example, the rehire preference clearly 
serves to perpetuate current staffing patterns, 
Craig v. Alabama State Univ.. 804 F.2d 682, 687 n.7 
(11th Cir. 1986), and whites benefitted overwhelm­
ingly from the nepotistic practices. ER 57-101.

10

236



(1971), Albemarle Paper Co. v. Moody. 422 
U.S. 405 (1975). Yet petitioners ignore 
this overwhelming body of evidence, 
focussing their attack on only a small 
piece, and in so doing seriously distort 
the reality that workers in these Alaskan 
Salmon canneries experience.

This case recalls an earlier era of 
Title VII enforcement when race and sex 
typing in employment was rampant, often 
overt and institutionalized. Although the 
law has developed to address more sophisti­
cated and subtle forms of discrimination, 
there is nothing sophisticated or subtle 
about this case. Only by viewing the 
totality of the evidence, Bazemore v. 
Friday. 478 U.S. 385 (1986), rather than 
forcing the facts into a legal straight 
jacket never intended for this kind of

11

237



situation, can this case be properly 
analyzed. See Point I, infra.

The rules and definitions established 
in Griggs. refined in Albemarle, and 
reaffirmed in Connecticut v. Teal. 457 U.S. 
440 (1982), directly govern this case, and 
compel the conclusion that petitioners have 
wholly failed to rebut the evidence of 
discrimination or to demonstrate that their 
practices are justifiable. See also 
International Brotherhood of Teamsters v. 
United States. 431 U.S. 324 (1977). The
defense consists of speculative attacks on 
plaintiffs' evidence, unsupported asser­
tions of necessity, and hypothetical 
allegations as to the cost of improvements. 
Thus, petitioners and the United States 
propose dispensing with the authoritative 
interpretation of Title VII that has 
governed for 17 years, even though Congress

12

238



has expressly ratified this application of 
the statute. See Point II, infra.

Principles of stare decisis preclude 
the result petitioners urge. The Court 
should instead affirm its long-standing 
rule that practices "fair in form, but 
discriminatory in operation" are unlawful 
unless affirmatively justified as necessary 
to the business. See Point III, infra.

ARGUMENT

Introduction
Many industries have been marked by 

the pervasive stratification and stereo­
typing apparent in this case. Historical­
ly, for example, the trucking industry was 
stratified by race and sex. Whites 
predominated in the lucrative over-the-road 
("OTR") tractor trailer driving jobs, while 
minorities were initially confined to 
certain shop positions and later obtained

13

239



employment driving local routes and working 
on the loading dock. See Teamsters« 431 
U.S. at 337-38; Franks v. Bowman Transp. 
Co., 495 F.2d 398, 410 n.10 (5th Cir.
1974), rev 'd on other grounds. 424 U.S. 747
(1976); Jones v. Lee Wav Motor Freight,
Inc.. 431 F.2d 245 (10th Cir. 1970), cert. 
denied. 401 U.S. 954 (1971). This
allocation of jobs was maintained by a 
system of subjective and discretionary 
hiring practices, often in combination with 
neutral rules, like no-transfer provisions, 
that operated to freeze and perpetuate the 
discriminatory practices. See also Kilgo 
v. Bowman Transp. Inc.. 789 F.2d 859, 869- 
75 (11th Cir. 1986) (women excluded from 
OTR driving jobs as a result of a neutral 
prior experience requirement and other 
practices, including unequal and 
inconsistent application of job require­

14

240



ments, failure to provide adequate bathroom 
and bunk facilities, and other practices 
designed to deter women applicants).

The same patterns occurred in many 
settings. Whites entered the preferred 
lines of progression, and minorities and 
women were routed to race and sex-typed 
positions, often the lowest paid, with the 
least likelihood of advancement.^ Subjec­
tive assessments, selective recruitment, 
experience and education requirements, and 
paper and pencil tests were all employed, 
often simultaneously.

2/ E.g., Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975) ("racial identifiability" of lines 
of progression; blacks in worst jobs); United 
States v. County of Fairfax. Va., 629 F.2d 932, 
937-38 (4th Cir. 1980) (women clustered in clerical 
jobs, minorities in maintenance); Gamble v. 
Birmingham Southern Railroad Co.. 514 F.2d 678 (5th Cir. 1975) (blacks confined to switchman position); 
March v. Eaton Corp., 639 F.2d 328 (6th Cir. 1981) 
(women channeled into specific jobs).

15

241



Vigorous enforcement of Title VII, 
and judicial scrutiny of employment prac­
tices that resulted in marked workforce 
stratification has made this a less 
familiar, but by no means non-existent, 
pattern. Women workers, in particular, are 
still subject to stereotypes about their 
interests and skills and to pervasive 
discrimination in male-dominated industries 
and professions. See. e. q. . Catlett v. 
Missouri Highway & Transp. Comm'n. 828
F.2d 1260 (8th Cir. 1987) cert. denied, __
U.S. __ , 108 S.Ct. 1574 (1988); EEOC v.
Rath Packing Co.. 787 F.2d 318, 328 (8th 
Cir.), cert, denied. 479 U.S. 910 (1986); 
Kilgo v. Bowman Transp. Inc.. 789 F.2d 859. 
No one familiar with the efforts to eradi­
cate widespread employment discrimination, 
to which Title VII was originally 
dedicated, can fail to recognize the

242
16



pattern present in this case. This case, 
like Griggs. 401 U.S. 424, and Albemarle, 
422 U.S. 405, had its genesis in the 
overtly discriminatory practices pursued 
widely in this industry prior to the 
enactment of Title VII. And, like Griggs 
and Albemarle, this case also involves 
neutral practices, applied to all, that 
fall more harshly on one group and that 
"operate to 'freeze' the status quo of 
prior discriminatory employment practices." 
Griggs. 401 U.S. at 430.i^/

I. THE TOTALITY OF THE EVIDENCE,INCLUDING STATISTICAL EVIDENCE OF 
RACIAL STRATIFICATION, EASILY ES­
TABLISHES A PRIMA FACIE CASE OF DISPARATE IMPACT DISCRIMINATION.

3zQ/ There can be no doubt in this case that 
the "racial identifiability" of jobs, Albemarle,
422 U.S. at 409, did not arise from a "myriad of 
innocent causes," Watson v. Fort Worth Bank & Trust
Co., 487 U.S. __, 108 S.Ct. 2777 (1988), but froman extended history of specific discriminatory practices.

17

243



The parable of the blind describing an 
elephantii/ is an apt metaphor for the 
mischaracterization of this case by 
petitioners and some supporting amici.
By focussing their attack on a single 
element of the case, namely the comparison 
between the proportion of non-whites in 
cannery jobs and non-whites in better 
paying, more desirable "at issue" jobs, as 
if no other evidence exists, the employers 
distort reality. The record establishes 
pervasive evidence of discrimination, of 
which this particular statistical

ii/ Several blind people stationed at various 
points around an elephant are each separately asked 
to describe the beast. The one at the trunk 
reports that an elephant is a large hose; the one 
at the tusk describes a curved spike; another at 
the ear says the elephant resembles a sail; the one 
feeling the body says an elephant is a large hairy 
wall, and according to the one at the tail an 
elephant seems to be a sort of rope with a tassel 
at the end. Here, the employers and their amici 
are all feeling the tail; they perceive no 
elephant, only a rope.

18

244



comparison is only a part: discrimination
is apparant "by the manner in which [an 
employer] publicizes vacancies, his 
recruitment techniques, his responses to 
casual or tentative inquiries, and even by 
the racial or ethnic composition of that 
part of his workforce from which he has 
discriminatorily excluded members of 
minority groups." Teamsters, 431 U.S. at 
365.

Petitioners do not contest many of 
the facts upon which respondents rely, but 
they question the relevance and probative 
value of this evidence and assert that some 
of the practices are defensible. Specifi­
cally, the employers do not deny that 
general hiring, hiring of relatives, 
recruitment, rehire, bunking and messing 
practices occur essentially as respondents 
describe them. They claim that there is no

19

245



statistical evidence of discrimination, 
because they allegedly hire in proportion 
to the "qualified" labor pool.

Both parties attempted to define the 
appropriate comparative reference to 
measure statistically the significance of 
the obvious racial stratification in the 
labor force, and there was and is sharp 
dispute on this issue. This dispute, 
however, should not obscure the agreement 
regarding the other factual elements of the 
employment process (even if the legal 
implications are debated), and the fact 
that these elements can and should be 
considered in determining the existence of 
the prima facie case. Cf. Bazemore v. 
Friday. 478 U.S. at 400.

The dispute over the statistical 
evidence of impact ultimately involves two 
questions: (1) whether qualifications that

20
246



have not actually been uniformly applied in 
the selection of employees and that them­
selves may disproportionately exclude non­
whites may be used to define the relevant 
labor pool for purposes of statistical 
evidence of impact; and (2) whether the 
employers may themselves manipulate the 
labor pool, by selectively seeking employ­
ees for some jobs in a predominantly white 
area and employees for other jobs in 
predominantly non-white areas.
A. Petitioners Have Failed to Demonstrate 

that Observed Disparities Are Attrib­
utable to Valid Skill or Qualifi­
cations Requirements
Evidence of racial, ethnic, or sex- 

based stratification in employment compell- 
ingly suggests the presence of discriminat­
ory practices or the residue of intention-

21

247



al, historical discrimination.-3-2/ Such 
Such stratification is unlikely to occur as 
a result of choice when the jobs to which 
minorities or women are confined are the 
least desirable ones. It is equally 
unlikely that qualifications or skills 
fully account for the disparities, espe­
cially in the jobs "at issue" in this case: 
until this litigation, the employer had no 
defined requirements for the jobs, and many 
of the incumbents were unskilled. The 
selection process for at issue jobs was 
highly subjective.^/ Under these

-3̂ / E.q.. Gamble v. Birmingham So.R. Co.. 514 
F.2d 678 (5th Cir. 1978); Muller v. U.S. Steel 
Corp., 509 F.2d 923 (10th Cir.), cert, denied. 423 
U.S. 825 (1975). See also Albemarle. 422 U.S. 405 
and Griggs. 401 U.S. 424.

3̂/ ihe failure to articulate objective 
criteria, and the reliance on unguided subjective 
decisions, has often been identified as a ready 
mechanism for discrimination. Knight v. Nassau 
County Civil Service Commission. 649 F.2d 157, 161 
(2d Cir.), cert, denied. 454 U.S. 818 (1981); 
Williams v. Colorado Springs School District. 641

22

248



circumstances, the skill level, for 
purposes of identifying comparative labor 
force data, must be based on the "non- 
discriminatory standards actually applied 
... to individuals who were in fact hired." 
Franks v. Bowman Transp. Co.. 424 U.S. at 
772 n.32. See also Albemarle. 422 U.S. at 
433. Since the evidence demonstrates that 
no standards were developed or consistently 
applied to actual hires, labor force 
comparisons cannot be restricted by 
hypothetical skill levels. Indeed, the 
notion that the Beach Gang truck driver job 
is a "skilled" job conflicts with this 
Court's repeated statements, in the context

F.2d 835, 842 (10th Cir. 1981); Fisher v. Procter 
and Gamble Manufacturing Co.. 613 F.2d 527, 546 
(5th Cir. 1980), cert, denied. 449 U.S. 1115
(1981); Parson v. Kaiser Aluminum and Chemical 
Corporation. 575 F.2d 1374, 1385 (5th Cir. 1978), 
cert, denied sub, nom. Local 13000. U.S. 
Steelworkers of America v. Parson. 441 U.S. 968 
(1979); Rowe v. General Motors Corporation. 457 
F.2d 348, 359 (5th Cir. 1972).

23

249



of OTR trucking cases, that this is a skill 
widely possessed or easily obtained. See. 
e.g.. Franks and Teamsters.

It is hard to imagine that driving a 
truck around a cannery reguires more skill 
than driving an 18 wheel tractor-trailer 
across country. Similarly, it is hard to 
understand why Alaskan native Americans are 
deemed not sufficiently skilled to cook on 
board fishing boats, or to be deckhands. 
They lack the requisite "skill” level, 
according to petitioners, because they have 
previously been excluded from these jobs.

It is the defendant's burden to prove 
the necessity of special qualifications, if 
they are in doubt. EEOC v. Radiator 
Specialty, Co.. 610 F.2d 178, 185 n.8 (4th 
Cir. 1979); Davis v. Califano, 613 F.2d 
957, 964-65 (D.C. Cir. 1979); Harrell v. 
Northern Elec. Co., 672 F.2d 444, 448 (5th

24

250



Cir. 1982). Where the very subjectivity of 
the practice is challenged as contributing 
to, or creating, the discriminatory result, 
numerous courts have held that comparative 
labor force data may be limited only by 
reference to the "minimum objective 
qualifications" for the job. Davis v. 
Califano. 613 F.2d at 964. There were no 
minimum objective qualifications for the 
jobs at issue in this case because none had 
been identified and none was uniformly 
required.
Furthermore, prior experience and skill 

requirements would themselves be discrim­
inatory. Defining the labor pool according 
to discriminatory standards would be 
entirely inappropriate:

The circuitousness of this 
bootstrap argument becomes 
obvious when one recalls that it 
is [the] qualifications for 
flight officer that appellant 
claims are discriminatory.

25

251



Spurlock v. United Airlines, 475 F.2d 216,
218 (10th Cir. 1972) (emphasis in 
original).14/ in such a case, it would be 
unreasonable to require plaintiffs to rely 
on the very data that "may be biased 
against them." De Medina v. Reinhardt. 686 
F.2d 997, 1010 n.8 (D.C. Cir. 1982).

Lower federal courts are virtually un­
animous in imposing on employers the burden 
of proving the existence and/or necessity 
of special job qualifications or skills 
where the requirements are not uniformly 
imposed, where they are not clearly jus­
tifiable, or where they may themselves 
cause or contribute to discriminatory 
exclusions. See EEOC v. Rath Packing Co.. 
787 F.2d at 327, 336 ("Rath could not IV

IV in Spurlock, the court held that a prima 
facie case was established through evidence of the 
"minuscule" number of blacks employed, and that the 
employer bore the burden to justify the job qualifier

26

252



identify any criteria it used in selecting 
employees or any common qualifications or 
skills that its employees possessed"); 
Crawford v. Western Elec. Co.. 614 F.2d 
1300, 1315 (5th Cir. 1980) ("an employer 
may not utilize wholly subjective standards 
by which to judge its employees' qualifi­
cations and then plead lack of qualifi­
cation when its promotion procedure is 
challenged as discriminatory"); Kinsey v. 
First Regional Securities. 557 F.2d 830, 
837-38 (D.C. Cir. 1977) ("objective 
criteria" had not been uniformly applied; 
skill requirements should be scrutinized if 
the evidence showed systematic exclusion of 
blacks).

Petitioners further assume that skills 
and qualifications are unevenly distributed 
in the population on the basis of race or 
ethnicity. The "more logical assumption,"

27

253



however, is that skills of the sort 
involved in this casei^/ are evenly 
distributed. De Medina v. Reinhardt. 686 
F.2d at 1008 n.7; United States v. County 
of Fairfax. 629 F.2d at 939. The employer 
bears the "burden of producing evidence 
from which it is possible to evaluate the 
likelihood that the disproportionate impact 
was caused by unequal qualifications." De 
Medina, 686 F.2d at 1009 n.7 (quoting D. 
Baldus and I. Cole, Statistical Proof of 
Discrimination 194-195 (1980)). See also 
Catlett v. Missouri Highway & Transp.
Comm1n., 828 F.2d at 1266 (employers burden 
to show that statistical disparity results 
from women's lack of interest). *

Hie skill level, to the extent it exists, 
is of the experiential sort, commonly available to 
all groups. Unlike brain surgeons and rocket 
scientists, societal discrimination does not 
operate against non-whites in these categories, but 
rather fosters their participation.

28

254



In sum, the assertion that skills, 
qualifications, or interest level precludes 
the statistical comparisons undertaken by 
respondents is neither apparent from the 
nature of the jobs nor is it factually 
supported in the record. The burden was on 
the employers to prove the existence of 
minimum, objective qualifications actually 
applied during the relevant period to 
individuals actually hired so that 
respondents could, if necessary, adjust 
their statistical presentation. EEOC yiA. 
Radiator Specialty. 610 F.2d at 185-86. 
Absent this showing, respondents' 
statistical data cannot be rejected or 
discounted for failing to account for 
spurious or hypothetical skills and 
qualifications that are allegedly lacking

29

255



in the excluded class and that may
themselves foster discrimination.^/
B. The Data Comparing Non-White

Representation In Cannery and Non- 
Cannery Jobs, Along With Other 
Evidence, Demonstrates the Impact of 
the Employers' Practices
Petitioners and the United States

challenge the Ninth Circuit's reliance on
the patent racial stratification in this
workforce in finding a prima facie
case.!// To make this argument, they are

--/ Indeed, some plaintiffs and class members 
had college educations, JA 52, 56-60, 63-67, 71- 
73, 75-77, 85-86, but were apparently not deemed 
sufficiently qualified to work in the Beach gang, 
quality control, clerical, or other jobs.

^2/ The court may well have considered this 
evidence in the context of other evidence in the 
record, since nothing in the opinions below sug­
gests that the court relied solely on that 
evidence. Whether or not the Court of Appeals 
considered the totality of the evidence, 
respondents are, of course, entitled to assert any 
ground for affirmance of the judgment in their 
favor. Trans World Airlines. Inc, v. Thurston. 469 
U.S. Ill, 119 n.14 (1985). The totality of the 
evidence is clearly the appropriate standard to 
determine the existence of a prima facie case. 
Bazemore v. Friday. 478 U.S. at 400-01.

30
256



forced to rely on petitioners' own 
practices - the very subject of this 
lawsuit - to construct an argument that 
this evidence is irrelevant. In particu­
lar, petitioners rely on their practice to 
hire from outside the workforce to fill 
non-cannery jobs to justify the contention 
that comparisons to the external labor 
market are more relevant than internal 
comparisons. The government relies, 
ultimately, on the claim that use of Local 
37 as a referral source results in the 
"overrepresentation" of non-whites in 
cannery jobs, distorting any comparison 
between that group and non-cannery hires. 
Neither addresses the anomaly their 
arguments suggest: that the employers'
own challenged acts become the basis for 
defeating liability. Both ignore the fact 
that it was the employers' own decision to

31

257



hire for non-cannery jobs through offices 
in Washington and Oregon, and that the 
labor agreement with Local 37 gave manage­
ment the exclusive right to select new 
h i r e s . T h e  rehire preference compounded 
the harmful effects of this management 
prerogative.

The contention that non-whites are 
"overrepresented" is itself ironic: both
the government and petitioners claim that 
application of Griggs will "force" 
employers to adopt "quotas" to avoid 
potential liability. Both now seek the 
benefit of a "ceiling" quota. They argue

This fact distinguishes this case from 
General Bldg. Contractors Ass'n v. Pennsylvania, 
458 U.S. 375 (1982), in which the union was the 
exclusive referral source and the sole agent found 
liable for intentional discrimination under 42 
U.S.C. §1981. Compare Arizona Governing Comm, v. 
Norris. 463 U.S. 1073, 1089 & n.21 (1983) 
(employers gain no immunity by delegating tasks to 
third parties where "employers are ultimately 
responsible").

32

258



that their practices should be insulated 
from liability so long as their hiring is 
proportionate with the appropriate labor 
pool as they define it, whether or not the 
practices impact more heavily on non-whites 
than whites.12/

In any event, the record is devoid of 
any factual verification for the claim that 
non-whites are "overrepresented." In fact, 
the high proportion of non-whites may as 
easily reflect the fact that whites, having 
more options, are less willing to do this 
hard, seasonal work in remote locations. 
These factors, far from supporting the 
"overrepresentation" theory, suggest that 
the best comparison available is the actual 
workforce, a figure that avoids speculation 
about which elements of the labor force

12/ Ihis contention has already been rejected 
by this Court. Connecticut v. Teal. 457 U.S. 440
(1982).

33

259



might be available for or interested in 
this employment.

The other comparative data all suffer 
serious, if not fatal, flaws. The district 
court's own characterization of census
data-2^/ reveals why census figures are 
singularly inappropriate: they are largely
comprised of individuals in year-round, 
fixed location jobs near their place of 
residence; the canneries are located in 
remote, sparsely-populated regions and 
provide only seasonal employment. There is 
no apparent relevance of this data to this 
case, and the district court provided no 
rationale for its acceptance. The finding

5̂/ The district court noted that census data 
is "dominated by people who prefer full-year, 
fixed-location employment," but nonetheless found 
that "such data is nevertheless appropriate in 
defining the labor supplies for migrant seasonal 
work." Fdg.120, 34 CCH Empl. Prac. Digest 3̂4,437, 
p.33,829. Why such data is "appropriate" is not 
explained.

34

260



in this regard is simply devoid of any 
foundation and is clearly erroneous.

Especially in the absence of minimum 
objective qualifications, the apparent 
restriction of minorities and women to 
lower paying less desirable jobs has often 
been viewed by the courts as highly proba­
tive evidence of discrimination, and suffi­
cient to support a prima facie case. See 
Shidaker v. Tisch. 833 F.2d 627 (7th Cir.
1986), cert, denied. __  U.S. __ , 108
S.Ct. 2900 (1988); EEOC v. Radiator Spec­
ialty Co.. 610 F.2d at 181-82, Muller v. 
U.S. Steel Corp.. 509 F.2d 923. A 
"plaintiff in a Title VII suit need not 
prove discrimination with scientific 
certainty." Bazemore v. Friday. 478 U.S. 
at 400 (Brennan, J., concurring).

The evidence of racial stratification 
does not exist in a vacuum, and the

35

261



totality of the evidence clearly supports 
the finding of a prima facie case of 
discrimination id. at 401-02.

C. The Employers' Use of An 
Undifferentiated Hiring 
Procedure Makes the End Result 
Appropriate For a Measurement of Disparate Impact

Petitioners contend that the plain­
tiffs improperly relied on the "cumulative 
effects" of their employment practices and 
that, having failed to demonstrate the 
individual impact of each specific prac­
tice, their challenge must fail. Having 
created a multifactorial selection process, 
with subjective elements and unweighted 
components, the employers cannot assert 
that the system is immune from attack 
because it does not have separate, scored 
or weighted factors whose impact can be 
separately and independently ascertained.

36

262



In some cases, it might be entirely 
appropriate to require a plaintiff to 
demonstrate the discriminatory effect of 
identifiable criteria by which employees or 
applicants are selected. For example, 
police departments commonly subject new 
potential recruits to paper and pencil 
tests, medical examinations, physical 
ability tests, and background inquiries. 
Each one of these pre-employment hurdles is 
scored: a certain number of applicants are
eliminated, the remaining may be ranked.
The system contains both objective and 
subjective elements. Those who would 
challenge its validity may reasonably be 
expected to identify where the discrimina­
tion, if any, occurred, because in such a 
system the effects of each aspect of the 
screening process can be separately

37

263



ascertained. C f . Connecticut v. Teal, 457
U.S. 440.

But where the elements of the hiring 
process are not so clearly identifiable 
and applicants neither pass, fail nor score 
at any point until the end of the process 
when they are either hired or not, the only 
reasonable target for challenge is the 
result of the process. Even the United 
States recognizes that in such a situation 
the impact of each element in a multifacto­
rial process cannot be demonstrated.21/

II. TITLE VII REQUIRES MORE THAN A 
SHOWING OF BUSINESS RELATED 
PURPOSES TO DEFEND APPARENTLY 
DISCRIMINATORY EMPLOYMENT 
PRACTICES

21/ In its brief, at p.22, the government 
concedes that "if [multiple] factors combine to 
produce a single ultimate selection decision and it 
is not possible to challenge each one, that 
decision may be challenged (and defended) as a 
whole."

38

264



Flouting seventeen years of Title VII 
jurisprudence, and Congressional ratifica­
tion of that caselaw, the employers and the 
government in this case advocate abandon­
ment of the central tenet of Griggs: that
equal employment opportunity may not be 
offered "merely in the sense of the fabled 
offer of milk to the stork and the fox,
...[but] that the vessel in which the milk 
is proferred be one all seekers can use.
The Act proscribes ... practices that are 
fair in form, but discriminatory in 
operation. The touchstone is business 
necessity." 401 U.S. at 431 (emphasis
added).

Instead, the employers argue that 
they should be able to defend their prac­
tices simply by offering "reasonably clear 
and specific" reasons. Petitioner's Brief 
at 39; see also U.S. Brief at 27. In fact,

39

265



the government would dispense with Griggs 
entirely: "Nothing about disparate impact 
cases justifies a departure from the model 
for litigating disparate treatment cases." 
Id? see also Petitioner's Brief at 47 
("Indeed, the rigid formula of Griggs 
itself should be reexamined in this 
context").

Thus, the government suggests that 
employers should be able to defend dis­
criminatory practices for any reason other 
than "non-business," U.S. Brief at 24-25,
& n. 35. It would accept an employer's 
sincere but unsupported assertions.^/ It

^/ Indeed, the employer and United States 
characterize the defendant's burden as one of 
production despite decisions by this Court stating 
that the burden is one of proof. Albemarle. 422 
U.S. at 425 ("the company must meet the burden of 
proving that its tests (were) job related");
Dothard v. Rawlinson. 433 U.S. at 329 (the defen­
dants failed "to prove that the challenged require­
ments are job-related"). See also Connecticut v. 
Teal. 457 U.S. at 451 (employers must "demonstrate 
that the examination given was not an artificial,

40

266



was precisely this contention that this 
Court rejected over fifteen years ago when 
it held in Griggs that the employer's 
attempt to improve the overall quality of 
the workforce through educational and 
testing requirements, a "reasonably clear 
and specific" business related purpose, 
failed to provide an adequate defense.
Stare decisis precludes such a departure 
from established law, ratified by Congress.

A. Congress Has Ratified This Court's 
Requirement That Employers Must Demonstrate the Business Necessity of 
Practices Which Result in 
Discrimination.
In 1971 this Court expressly rejected 

Duke Power Company's defense that its 
educational and testing requirements, 
adopted in response to the increasing 
complexity of its business, served a

arbitrary, or unnecessary barrier, because it measured skills related to effective performance....")
41

267



"legitimate business purpose": the company
wanted to improve its efficiency and have 
some assurance that employees would be able 
to advance through the ranks.

In Griggs. the Court determined that 
the statutory exemption for professionally 
developed tests, 42 U.S.C. §2000e-2(h), did 
not shield Duke Power since the test had 
been "used to discriminate," as demon­
strated by its effect on blacks. Such 
tests, like other neutral practices with 
disproportionately adverse effects, could 
only be justified if the employer could 
show a "manifest relationship" to 
successful job performance or business 
necessity. Id. at 433.

The same year, Congress expressly 
endorsed the Griggs interpretation of Title 
VII, retaining the statutory language cited 
in Griggs. The House Report also ex-

42

268



plicitly quoted Griqqs -̂l/ and endorsed 
both the effects test and the holding that 
an employer's good faith could not defeat a 
finding of discrimination, if the employer 
failed to prove the existence of an 
overriding business necessity. H.R. 92-238, 
92d Cong., 1st Sess. 8 (1 9 7 1 ).24/
The Court has itself recently acknowledged 
that "Congress recognized and endorsed the 
disparate impact analysis employed by the 
Court in Griggs." Connecticut v. Teal. 457 
U.S. at 447 n. 8.

23/ see, e.q., H.R. 92-238, 92d Cong. 1st 
Sess. 8, 21 (1971) (specific reference to 
"business necessity" as "touchstone").

24/ congress further incorporated Griggs by 
providing that: "In any area where the new law
does not address itself, or in any areas where a 
specific contrary intention is not indicated... 
present case law as developed by the courts would 
continue to govern the applicability and construc­
tion of Title VII." 118 Cong. Rec. 7166, 7564
(1972).

4 3

269



Nor does legislative history provide 
any basis for relaxing Griggs1 requirements 
in cases involving subjective employment 
practices. The legislative history of the 
1972 amendments is replete with examples of 
racial stratification in upper level 
positions, positions where subjective 
criteria are most likely to be used.25/ 

Recognizing the increasingly subtle 
and complex nature of discrimination, the 
legislative history refers to "systems and 
effects." The House Report included within 
the ambit of prohibited "systems" practices 
relating to seniority, lines of progres­
sion, practices which perpetuate the

25/ For example, the concentration of 
minorities and women in the lowest paying and least 
desirable jobs was particularly noted: H.R. 92-
238, 92d Cong., 1st Sess. 23 (1971). 117 Cong.
Rec. 31960 (9/15/71) (remarks of Rep. Perkins); 117 
Cong. Rec. 32095 (9/16/71) (remarks of 
Rep. Faunteroy); 117 Cong. Rec. 32097 (9/16/71) 
(comments of Rep. Abzug); 117 Cong. Rec. 32101 
(9/16/71) (comments of Rep. Badillo).

44

270



present effects of earlier discrimination 
through various institutional devices, as 
well as testing and validation require­
ments. H.R. 92-238, 92d Cong. 1st Sess. 8
(1972) .

The rule of law enunciated in Griggs 
has been adopted by Congress on other 
occasions, as well, in defining prohibited 
discrimination. For example, it has been 
used to define conduct prohibited by the 
Voting Rights Act as well as the 
Rehabilitation Act.21/ In sum, Congress 
has on several occasions relied upon, and

2V Ihe legislative history of the Voting 
Rights Act Amendments of 1982, 97 P.L. 205, 96 Stat. 131 42 U.S.C. §1973 (1982), shows express 
reliance on Griggs: "the results test to be
codified in Section 2 is a well defined standard, 
first enunciated by the Supreme Court and follcwed 
in numerous lower federal court decisions". S.
Rep. No. 97-417, 97th Cong., 2d Sess. at 17, reprinted in 1982 U.S. Code Cong, and Admin.News 
193-94. Section 504 of the Rehabilitation Act, 29 
U.S.C. §794 (1982), has been similarly held to incorporate Griggs standards. Alexander v. Choate, 
469 U.S. 287 (1984).

45

271



endorsed, the Griggs decision, as 
interpreted in this and lower federal 
courts. An integral part of that decision 
is the insistence on business necessity as 
the standard by which practices, "fair in 
form but discriminatory in operation," must 
be justified.
B. Decisions of This Court Reveal That 

the Business Necessity Defense Is An 
Essential Element of Disparate Impact Analysis
Like Congress, this Court has 

consistently affirmed the rule of Griggs as 
originally enunciated, including its 
insistence on business necessity as the 
standard by which discriminatory practices 
must be judged. Thus, this Court has held 
that employment practices which "operate as 
'built-in headwinds'" must be validated and 
"shown, by professionally acceptable 
methods, to be 'predictive of or signifi­
cantly correlated with' important elements

46

272



of work behavior which comprise or are
relevant to the job...." Albemarle, 422
U.S. at 431 (quoting 29 C.F.R. §1607.4
(c)). The Court has specifically
emphasized Title VII's rigorous burden:

it is an insufficient response to 
demonstrate some rational basis for 
the challenged practices. It is 
necessary in addition, that they be 
'validated' . . . .  However this 
process proceeds, it involves more 
probing judicial review of, and less deference to. the seemingly reasonable 
acts of administrators."

Washington v. Davis. 426 U.S. 229, 247
(1976) (emphasis added) (distinguishing
constitutional cases) .

27/ power courts have consistently required 
that a defendant provide professional or empirical 
validation for their discriminatory practices. 
Boston Chapter. NAACP. Inc, v. Beecher, 504 F.2d 1017, 1022 (1st Cir. 1974), cert, denied, 421 U.S.  
910 (1975) (rejecting written test for firefighters); Grant v. Bethlehem Steel, 635 F.2d 
1007, 1015 (2d Cir. 1980), cert, denied, 452 U.S.  
940 (1981); Geller v. Markham. 635 F.2d 1027, 1032- 
1034 (2d Cir. 1980), cert, denied. 451 U.S.  945
(1981) (cost concerns rejected); Robinson v. 
Lorillard Corp.. 444 F.2d 791, 799 (4th Cir.), cert, dismissed, 404 U.S. 1006 (1971); Fisher v.

47

273



Allegations of good faith will not 
rebut a prima facie case of class-wide 
discrimination. Griggs, 401 U.S. at 431, 
and see Teamsters. 431 U.S. at 342 n.24.
Nor will a desire to hire only the best 
qualified applicants. Id. Instead, the 
employer has to show that the challenged 
practices are of "great importance" or of a 
"compelling nature." Connecticut v. Teal.

Procter and Gamble Manufacturing Co.. 613 F.2d 527, 
544-45 (5th Cir. 1980), cert, denied. 449 U.S. 1115 
(1981) (qualifying tests and experience); Rowe v. 
General Motors Company. 457 F.2d 348, 358 (5th Cir. 
1972) (rejecting "ebb and flow" in production 
level); Harless v. Duck. 619 F.2d 611, 616-617 (6th 
Cir.), cert, denied. 449 U.S. 872 (1980) (physical 
ability test and oral interview); Caviale v. State 
of Wisconsin Department of Health and Social Serv. 744 F.2d 1289, 1294-95 (7th Cir. 1984) (requirement 
that applicants be members of Career Executive 
Program); Firefighters Institute for Racial 
Equality v. City of St. Louis. 549 F.2d 506, 511 
(8th Cir.), cert, denied. 434 U.S. 819 (1977) 
(promotional exam); Hawkins v. Bounds. 752 F.2d 500 
(10th Cir. 1985) (discriminatory detailing); Walker 
v. Jefferson County Home. 726 F.2d 1554, 1559 (11th 
Cir. 1984); Haves v. Shelby Memorial Hospital. 726 
F.2d 1543, 1553 n.15 (11th Cir. 1984) (potential litigation costs).

48

274



457 U.S. at 451. 2 3 / Accord Nashville Gas
Co. V. Sattv, 434 U.S. 136, 143 (1977);
Dothard v. Rawlinson 433 U.S. at 331; 
General Electric Coro, v. Gilbert. 429 U.S. 
at 138 n.14 (1976); McDonnell Douglas Corp.
v. Green, 411 U.S. 792, 805 (1973).

Contrary to the government's and the 
petitioners' suggestion, the business 
necessity defense is an "affirmative 
defense," entirely distinct from the "much 
less burdensome riposte... applicable to

2 3 / see also Hawkins v. Bounds. 752 F.2d 500 
("great importance"); Grant v. Bethlehem Steel, 635 
F.2d 1007 ("genuine business need"); Kirby v.
Colony Furniture Co., 613 F.2d 696, 705 n.6 (8th 
Cir. 1980) ("compelling need"); Walker v. Jefferson 
County Home. 726 F.2d at 1559 ("rigorous standard"); Kirkland v. New York State Department 
of Correctional Services, 520 F.2d 420, 426 (2d 
Cir. 1975), cert, denied, 429 U.S. 823 (1976) 
("heavy burden"); Boston Chapter, NAACP v.Beecher. 504 F.2d at 1024 ("substantially related" 
by "convincing facts"; Robinson v. Lorillard, 444 
F.2d at 798 ("sufficiently compelling"); Williams 
v. Colorado Springs School District, 641 F.2d 835, 
842 (10th Cir. 1981) ("the practice must be essen­
tial, the purpose compelling").

49

275



individual disparate treatment cases.
Lewis v. Bloomsburq Mills. Inc.. 773 F.2d 
561, 571-72 (4th Cir. 1985). Articulation 
of a legitimate non-discriminatory reason 
"simply has no doctrinal relevance as a 
negating or justifying defense to 
statistically proven patterns and practices 
of discrimination ... [it is] 'about as 
relevant as a minuet is to a thermonuclear 
battle."' Id. at 572 n.19 (citation 
omitted.) Accord Nash v. Consolidated 
City of Jacksonville. 837 F.2d 1534, 1536 
(11th Cir. 1986).

If the position now advocated by the 
government and the employers were the law, 
Title VII would have done little to 
desegregate the workplace. For example, 
the employer in Dothard may have sincerely 
believed that height and weight correlated 
with effectiveness as a corrections

50

276



officer, but that belief was not factually 
supported. If a lower level of proof —  
merely assertion of a business purpose —  
had been accepted, women would still be 
virtually excluded from law enforcement 
jobs, many factory jobs, and other 
positions for which size has been wrongly 
assumed to be relevant.^/ See also Weeks 
v. Southern Bell Tel. & Tel. Co., 408 F.2d 
228, 236 (5th Cir. 1969) ("using these 
class stereotypes denies desirable

22/ on only one occasion has the Court even 
arguably excused the usual validation requirement. 
In New York City Transit Authority v. Beazer. 440 U.S. 568 (1979), the Court noted that plaintiffs' 
statistical data was "weak," possibly not even 
sufficient to establish a prima facie case. Id. at 
587. Under these circumstances, the Court found 
the restriction of methadone users from "safety 
sensitive" positions was job-related. Id., n.31.
Ihe ultimate holding was contained in a footnote: 
"Whether or not respondents' weak showing was 
sufficient to establish a prima facie case, it 
clearly fails to carry respondents' ultimate 
burden..." Id. This decision provides no basis 
for the conclusion that the Court has dispensed 
with the business necessity requirement in response 
to a prima facie case of discrimination.

51

277



positions to a great many women perfectly 
capable of performing the duties 
involved").

A simple explanation fails to rise to 
the level of the evidence of systematic 
discrimination inherent in a prima facie 
case of class-wide discrimination. Any 
employer can always articulate some 
business-related reason for its practices, 
but that does nothing to address the 
systematic impact of discriminatory 
practices. Title VII's focus is and should 
be on the eradication of practices that 
inhibit equal employment opportunity, 
regardless of motive.

2 0 / indeed, business related purposes have 
been advanced to justify even facially discrimi­
natory practices, and the asserted non-invidious 
motives were most likely genuine. E.g.. Los 
Angeles Dep't of Water and Power v. Manhart, 435 
U.S. 702 (1978). Sex discrimination cases reveal 
the irrelevance of bigotry or animus to a finding 
of discrimination. Invidious discrimination 
against women has historically been characterized

52

278



Statistical data may always be 
challenged as irrelevant, unreliable or 
inaccurate, because the plaintiff's prima 
facie case can always be attacked directly. 
Class disparate treatment cases best define 
this process: the defendant must produce
its own statistical analysis, or valid 
statistical critique, demonstrating why the 
statistics are incorrect or unworthy of 
credibility. Teamsters. 431 U.S. at 340.
In reviewing the statistical data in 
Bazemore v. Friday, the Court noted that 
the defendants failed to provide "evidence 
to show that there was in fact no 
disparity":

by benign motives. The Supreme Court has 
invalidated intentional discrimination even where 
the intent was" to favor [women], not to disfavor 
them." Wenqler v. Druggists Mutual Ins. Co.. 446 
U.S. 142, 150 (1980). See also Orr v. Orr. 440 
U.S. 268 (1979) (alimony statute which benefitted 
only women invalidated as unconstitutional 
intentional discrimination).

53

279



Respondents' strategy at trial 
was to declare simply that many 
factors go into making up an 
individual employee's salary? 
they made no attempt that we are 
aware of-statistical or 
otherwise-to demonstrate that 
when these factors were properly 
organized and accounted for there 
was no significant disparity 
between the salaries of blacks 
and whites.

478 U.S. at 403-04 n.14 (references 
omitted) (Brennan, J., concurring).
As Bazemore thus makes clear, affirmations 
of good intentions do not refute a 
statistical showing of discrimination - 
only facts suffice.^/

Failing refutation of the statistical 
evidence, an employer is free to prove the 
business necessity of practices with

"[HJypotheses or conjecture will not 
suffice" to rebut statistical evidence of 
discrimination. Cable v. Hot Springs school Dist.. 
682 F.2d 721, 730 (8th Cir. 1982) (citation 
omitted). Catlett v. Mission Highway & Transp. 
Comm.. 828 F.2d at 1266 (defendant may not rely on 
"mere conjecture or assertion" but must produce 
"direct evidence").

54

280



discriminatory impact.^/ Here, the 
employers offered little other than 
assertions of their own good faith and the 
business related purposes that were 
allegedly served.-2-2/

In the context of both disparate 
impact and class disparate treatment cases, 
this Court has squarely and consistently 
rejected the assertions of business-related 
purposes to rebut a discrimination claim 
that relied in whole or in part on class-

The job analysis prepared in this case is 
unavailing, since it is conceded that the 
qualifications had not been previously identified 
or required. Validation undertaken in preparation for litigation is always subject to particular 
scrutiny. Albemarle. 422 U.S. at 433 n.32. 
Moreover, the study was materially defective in 
that it did not attempt to correlate inportant 
elements of work behavior with the job 
qualifications identified. This is legally insufficient. Id. at 431-32.

3̂/ Hie district court accepted these mere 
"articulations" but the court failed to apply 
disparate inpact and these conclusions were thus 
governed by an incorrect legal standard.

55

281



wide statistical evidence of 
discrimination. ^4/
C. Cost Considerations Do Not Establish 

Business Necessity Or Excuse The 
Necessity To Prove It
The employers and the government 

contend that the potential costs of non­
discrimination should excuse discriminatory 
practices, in essence arguing that the 
assertion of a cost-based rationale should

34/ To the extent that petitioners and the 
government rely on the opinion of four members of 
the Court in Watson v. Fort Worth Bank & Trust Co.,
487 U.S. __, to support the conclusion that an
"articulation" defense is appropriate in this case, 
that reliance is misplaced. Watson involved a 
single individual claimant who asserted that blacks 
were disproportionately affected by the bank's 
subjective promotion practices. The Court held 
that she could proceed on such a theory. It had no 
occasion to address the burden on a defendant in 
response to evidence of class-wide discrimination. 
This Court's decisions have always distinguished 
between individual and class disparate treatment 
cases with regard to both the prima facie case and 
the defense. Compare Teamsters with Texas Dep't of 
COnrounitv Affairs v. Burdine. 450 U.S. 248 (1981). 
When evidence of classwide discrimination is 
present, even in an individual case, the Teamsters 
model is the more relevant one.

56

282



suspend the obligation to prove business 
necessity. But this is simply another 
rendition of their contention that 
articulation of a business related purpose 
should provide a defense. Here, the claim 
regarding costs is wholly speculative and 
unsupported. On this basis alone it should 
fail.

Morover, Congress has rejected the
notion of a cost-based defense. In 1978,
opponents of the Pregnancy Discrimination
Act complained that compliance would be too
costly. Representative Hawkins, sponsor of
the amendment, replied:

Eradicating invidious discrimination 
by definition costs money: It is
cheaper to pay all black workers less 
than all white workers, or all women 
less than all men. The fact that it 
would cost employers money did not 
prevent Congress from enacting the 
Equal Pay Act or title VII, and it 
should not prevent this Congress from 
making clear that title VII prohibits 
this form of sex discrimination as 
well.

57

283



Legislative History of the Pregnancy
Discrimination Act (1979) (committee print
prepared for the Senate Committee on Labor
and Human Resources) at 26. The Senate
Report concluded:

even a very high cost could not 
justify continuation of the policy of 
discrimination against pregnant women 
which has played such a major part in 
the pattern of sex discrimination in 
this country.

Id. at 48.
This Court has explicitly rejected a 

cost-based defense on three separate 
occasions. Los Angeles Dep't Water & Power 
v. Manhart. 435 U.S. 702, Newport Hews 
Shipbuilding & Dry Dock Co. v. EEOC. 462 
U.S. 669 (1983), and Arizona Governing 
Committee v. Norris. 463 U.S. 1073
(1983).35/ Even the Griggs Court must have

35/ See also Robinson v. Lorillard, 444 F.2d 
at 799-800 and n.8; Haves v. Shelby Mem. Hosp.. 726 
F.2d 1543, 1552 n.15 (11th Cir. 1984); Smallwood v.

58

284



recognized that validation and the 
alteration of unlawful practices would 
entail some costs, but mandated them 
anyway, and has repeatedly done so.
III. PRINCIPLES OF STARE DECISIS PRECLUDE

ABANDONING OR ALTERING THE BUSINESSNECESSITY DEFENSE.
The decisions noted above constitute 

an integral part of Title VII jurisprudence 
and are central to the disparate impact 
theory. Principles of stare decisis 
dictate adherence to Griggs. Albemarle. and 
other authoritative interpretations of 
Title VII. The Court has held that stare 
decisis "weighs heavily" and precludes a 
departure from precedent where Congress had 
the opportunity to reject the "Court's 
interpretation of its legislation" but

United Airlines. Inc.. 661 F.2d 303, 307 (4th Cir. 
1981), cert, denied. 456 U.S. 1007 (1982) 
Gathercole v. Global Assocs.. 545 F. Supp. 1280, 
1282 (N.D. Cal. 1982).

59

285



declined to do so. Illinois Brick Co. v.
Illinois, 431 U.S. 720, 736 (1977).
Accord NLRB v. International Longshoremen 
Association. AFL-CIO. 473 U.S. 61, 84 
(1985). In deciding whether to overrule an 
earlier case, the Court is required to 
determine whether that act "would be 
inconsistent with more recent expressions 
of Congressional intent." Patsv v. Florida 
Board of Regents. 457 U.S. 496, 501
(1982) ; Monell v. New York City Department 
of Social Services. 436 U.S. 658, 695
(1978). See also Guardians Association v. 
Civil Service Commission of the City of New 
York. 463 U.S. 582, 612 (1983) (O'Connor, 
concurring)(constrained to follow Court's 
prior interpretation of Title VI). Here, 
overruling Griggs. or any of its essential 
elements, particularly the business 
necessity requirement, would plainly be

60

286



"inconsistent" with "expressions of 
Congressional intent."

Congress is, of course, free to 
reverse a statutory construction by this 
Court that is inconsistent with its intent, 
and it has done so on several occasions, 
notably when the Court has narrowed the 
scope of civil rights laws.3-6/

In some of these prior instances, 
arguably the intent of Congress had not 
been as clear as it is here. Congress 
could hardly have made plainer its intent 
to adopt the entire constellation of

3V Ehg., General Electric v. Gilbert, 427 
U.S. 125 (1976) led to the Pregnancy Discrimination 
Act, 42 U.S.C. §2000e(k) (1982). Grove City 
College v. Bell. 456 U.S. 555 (1984) was overruled 
by the Civil Rights Restoration Act, Pub. L. 100- 
259 (1988). Mobile v. Bolden. 446 U.S. 55 (1980), 
was reversed by the Voting Rights Act Amendments of 
1982, 42 U.S.C. §1973, et seg. (1982). In Mobile 
v. Bolden a plurality of the Supreme Court broke 
with precedent and substantially increased the burden on plaintiffs in voting discrimination cases 
by requiring proof of discriminatory purpose.

61

287



holdings for which Griggs stands. Under 
these circumstances, adherence to stare 
decisis will not only preserve the 
integrity of the judicial process, by 
reinforcing judicial reliability and 
predictability, it will also ensure that 
Congressional action endorsing Griggs will 
be respected.

62

288



CONCLUSION
For the foregoing reasons, amici 

respectfully submit that the judgment of 
the Ninth Circuit in favor of respondents 
should be affirmed.

Respectfully submitted,

JOAN E. BERTINCounsel of Record 
KARY L. MOSS ISABELLE KATZ PINZLER 
JOHN A. POWELL American Civil Liberties 

Union Foundation 
132 West 43rd Street 
New York, New York 10036 
(212) 944-9800

November 5, 1988

63

289





No. 87-1387

In The

Sup rem e Court of tlje llm tc b  s t a le d
October Term, 1988

WARDS COVE PACKING COMPANY, INC., 
CASTLE & COOKE, INC.

Petitioners,
v.

FRANK ATONIO, et a l ,
Respondents.

On Writ Of Certiorari To The United States 
Court Of Appeals For The Ninth Circuit

BRIEF FOR THE AMERICAN SOCIETY FOR 
PERSONNEL ADMINISTRATION AS 

AMICUS CURIAE SUPPORTING PETITIONERS

Of Counsel:
BREED, ABBOTT & ‘ LAWRENCE Z. LORBER
MORGAN J. ROBERT KIRK

International Square 
1875 Eye Street, N.W.
Suite 1000
Washington, D.C. 20006 
(202) 466-1100

Attorneys for Amicus Curiae, 
American Society for Personnel 
Administration 
(‘Counsel of Record)

September 9, 1988



1

QUESTIONS PRESENTED

1. Should “disparate impact” plaintiffs bis permit­
ted to challenge facially neutral selection devices used 
to fill positions in one job category based only on 
statistics showing that plaintiffs are over-represented 
in a different job category?

2. In applying the disparate impact analysis, did 
the Ninth Circuit improperly alter the burdens of 
proof and engage in impermissible fact finding in dis­
regard of established precedent of this Court?

3. Did the Ninth Circuit commit error in allowing 
plaintiffs to challenge the cumulative effect of a wide 
range of alleged employment practices under the dis­
parate impact model?

292



11

TABLE OF CONTENTS

Page

Questions Presented ...................................................  i

Table of Authorities ..................................................  iv

Consent to Filing .........................................................  1

Interest o f  the Amicus Curiae ................................ 1

Statement of F acts ..................................................... 2

Summary o f  Argument ...........................................  4

Argument ..........................................................................  7
A. In the Wake of Watson, Employers Must Be

Permitted To Judge The Personal Qualities 
Of Individual Job Applicants On A Case-By- 
Case Basis Without Being Saddled With Un­
manageable Burdens .....................................  7

B. Justice O’Connor’s Formulation Of Disparate 
Impact Analysis In Watson Is Based On Prin­
ciples Essential To A Workable Application
Oi The Model ................................................ 10
1. Because Statistical Imbalances Alone Do

Not Support A Presumption Of Unlawful 
Discrimination, Evidence Of An Adverse 
Impact Must Be Linked To A Discrete 
Selection Criterion ................................... 10

2. The First and Third Questions Presented
Should Be Resolved In Favor of 
Petitioners ................................................  15

C. Employers Should Not Bear the Burden Of
Proving the Business Necessity of Nonstan- 
dardized Selection Criteria ............................  19
1. By Their Very Nature, Nonstandardized 

Selection Criteria Are Not Amenable to 
the Techniques By Which the Job Relat­
edness Of Standardized Criteria Is Proven 
Or Disproven ............................................  21

293



• • •111

2. The Plurality Opinion In Watson Outlines
Evidentiary Standards Which Recognize 
the Special Nature Of Legitimate, Non- 
standardized Criteria ...............................  24

3. The Second Question Presented Ought To
Be Resolved In Favor Of Petitioner ....  28

Conclusion .......................................................................  30

294



IV

TABLE OF AUTHORITIES

CASES: Page
AFSCME v. State of Washington, 770 F.2d 1401

(9th Circuit 1985) .............................................  12
Albemarle Paper Co. v. Moody, 422 U.S. 405

(1975) .................................................................  20,25
Anderson v. Bessemer City, 470 U.S. 564 (1985) .. 17
Atonio v. Wards Cove Packing Co., Inc., 34 E.P.D.

1 34,437 (W.D. Wash. 1983).................. 2,3,A,passim
Atonio v. Wards Cove Packing Co., Inc., 768 F.2d

1120 (9th Cir. 1985) .........................................  4
Atonio v. Wards Cove Packing Co., Inc., 810 F.2d

1477 (9th Cir. 1987).........................................  4,16
Atonio v. Wards Cove Packing Co., Inc., 827 F.2d

439 (9th Cir. 1987).....................................  4,16,28,29
Fumco Construction Corp. v. Waters, 438 U.S. 567

(1978)   15,25
Griggs v. Duke Power Co., 401 U.S. 424

(1971)   9,20,21,24
Johnson v. Transportation Agency, Santa Clara

County, California, 480 U.S.__ , 94 L.Ed.2d
615 (1987)..........................................................  7,19

New York City Transit Authority v. Beazer, 440
U.S. 568 (1979)..............................................  12,20,25

Sheet Metal Workers v. EEOC, 478 U.S----- , 106
S.Ct. 3019 (1986)..............................................  19

Spaulding v. University of Washington, 740 F.2d
686 (9th Cir. 1984)...........................................  15

Steelworkers v. Weber, 443 U.S. 193 (1979) ............ 14
Teamsters v. United States, 431 U.S. 324 (1977) .. 21
Texas Department of Community Affairs v. Bur-

dine, 450 U.S. 248 (1981) ................................ 14,20
Washington v. Davis, 426 U.S. 229 (1976) ............  24,25
Watson v. Fort Worth Bank & Trust, 56 U.S.L.W.

4922 (June 29, 1988), vacating 798 F.2d
791 (5th Cir. 1986)................................. 5,7,9,passim

295



V

Page
CONSTITUTION AND STATUTES:
Civil Rights Act of 1964, Title VII, 42 U.S.C. §

2000e et seq................................................................. 3,passim
Uniform Guidelines on Employee Selection Proce­

dures, 29 C.F.R. § 1607 (1978) ......................  21

MISCELLANEOUS:
Bartholet, Application of Title VII to Jobs in High

Places, 95 Harv. L. Rev. 945 (1982) .............. 28

Table of Authorities Continued

296



In The

Sup rem e Court of tlje ® niteb  s ta t e d
October Term, 1988

No. 87-1387

WARDS COVE PACKING COMPANY, INC., 
CASTLE & COOKE, INC.

Petitioners,

v.

FRANK ATONIO, et a l,

Respondents.

On Writ Of Certiorari To The United States Court Of 
Appeals For The Ninth Circuit

BRIEF FOR THE AMERICAN SOCIETY FOR 
PERSONNEL ADMINISTRATION AS 

AMICUS CURIAE SUPPORTING PETITIONERS

CONSENT TO FILING
This Amicus brief is filed pursuant to Supreme Court 

Rule 36.2, with the written consent of all parties. Letters 
of consent have been filed with the Clerk of this Court.

INTEREST OF THE AMICUS CURIAE
The American Society for Personnel Administration 

("ASPA”) is the world’s largest association of personnel

297



2

and human resources professionals, representing over 
40,000 members in business, government, and education 
dedicated to the advancement of personnel and human 
resource management. Accordingly, ASPA and its mem­
bers have a keen interest in the development and enforce­
ment of the myriad of laws and regulations which govern 
many aspects of employment.

As the major professional association of the human re­
sources profession, ASPA is vitally concerned with the 
orderly evolution of laws defining, in practical terms, the 
meaning of equal employment opportunity. ASPA has long 
recognized its special responsibility to support and en­
courage compliance with fundamental principles of equal 
employment opportunity in the administration of efficient, 
workable personnel management systems. ASPA believes 
that the present case provides an excellent opportunity for 
this Court to reaffirm the appropriate balance between the 
compatible goals of providing equal employment opportun­
ities and preserving the right of managers to make legit­
imate personnel judgments in the best interest of their 
organizations.

STATEMENT OF FACTS*

Respondents are persons of Chinese, Filipino, Japanese 
and Native American descent who have been employed in 
fish canning facilities owned by Petitioners in remote, 
widely separated areas of Alaska (1, 50). Most of the jobs 
at the canneries are seasonal and temporary and are filled 
by migrant workers (52, 119). Since summer salmon runs 
are very short and the fish are extremely perishable, it is 1

1 Unless otherwise noted, each of the facts set forth herein is taken 
from the district court’s findings of fact following a nonjury trial. Ato- 
nio v. Wards Cove Packing Co., Inc., 34 E.P.D. 1 34,437 (W.D. Wash. 
1983) (copy appended as Attachment 1 to Petition For Writ of Certiorari 
herein). References such as “(1, 50)” refer to the numbers of the district 
court’s findings of fact.

298



3

essential that the canneries operate at peak production (51, 
63). “The slightest mistake in calibrating can size or in 
retort (cooking] management, for example, could result in 
a threat of widespread botulism, a disease fatal to hu­
mans.2

There were two general categories of jobs at Petitioners' 
canneries. The first, referred to as “cannery” or “laborer” 
jobs, included production line positions. The second, re­
ferred to as “noncannery jobs,” included all other de­
partments (82). Most cannery worker jobs did not require 
employees who were literate or able to communicate ef­
fectively in the English language, and none required em­
ployees to be available prior to the short, summer salmon 
run (117). Most of the so-called “noncannery” jobs required 
both English literacy and early season availability (117). 
It is the “noncannery” jobs which are at issue in this 
lawsuit (82).

Respondents Title VII claim alleged unlawful discrimi­
nation on the basis of color in Petitioners' selection of 
employees for the at-issue, noncannery jobs. Cannery 
workers and laborers at Petitioners’ facilities were pre­
dominately nonwhite. In these jobs, nonwhites were over­
represented in comparison to the relevant labor supply 
(105, 107). In noncannery positions, the district court found 
that whites and nonwhites were employed in percentages 
which approximated their availability in the relevant labor 
supply (123). Nevertheless, Respondents contended that the 
difference in the percentage of nonwhites in cannery jobs 
versus noncannery jobs supported a finding of unlawful 
discrimination.

The district court disagreed. It found that many of the 
jobs at Petitioners’ facilities were covered by union con­
tracts, and that Local 37 of the I.L.W.U.—the membership 
of which was predominately Filipino—provided an over- *

*Atonio, supra at fn.l, 34 E.P.D. at 33,840.

299



4

supply of nonwhites for cannery worker positions (84, 103). 
It found that Petitioners received relatively few applica­
tions for noncannery positions from nonwhites (89). It 
found experience in cannery positions did not qualify em­
ployees for noncannery jobs, that there was no opportunity 
for on-the-job training for skilled, noncannery jobs, and 
that Petitioners did not promote from within but filled 
positions by rehiring past employees or hiring new em­
ployees from the external labor market (57, 95, 97). In 
short, the district court found that Petitioners' “cannery 
workers and laborers do not form a labor pool for other 
jobs at [Petitioners'] facilities” (110).

The district court's opinion initially was affirmed on 
appeal5 but later was reversed by the Ninth Circuit sitting 
en banc.* Upon rehearing, the original panel remanded the 
case to the district court with instructions to consider 
Respondents' evidence under the “disparate impact” model 
of employment discrimination.5

SUMMARY OF ARGUMENT
Led by the human resources profession, American em­

ployers are firmly committed to providing equal employ- *

* Atonio v. Wards Cove Packing Co., Inc., 768 F.2d 1120 (9th Cir. 
1985)

* Atonio v. Wards Cove Packing Co., Inc., 810 F.2d 1477 (9th Cir. 
1987) (en banc).

1 Atonio v. Wards Cove Packing Co., Inc., 827 F.2d 439 (9th Cir. 
1987). The Court should note that while the district court believed 
“disparate impact” analysis should not be applied to all aspects of 
Respondents’ claim, it took pains to state its opinion as to the "business 

• necessity” of certain employment practices if "disparate impact” were 
applied. The district court concluded Respondents had faded to prove 
a "disparate impact” prima facie case or Petitioners had demonstrated 
"business necessity” with respect to (a) requiring English language 
literacy for noncannery workers, (b) word-of-mouth recruitment among 
relatives, (c) the rehire preference, (d) housing workers by department, 
and (e) feeding workers according to ethnic preferences. See Atonio, 
supra at fn.l, 34 E.P.D. at 33,840-844.

300



5

ment opportunities to people regardless of characteristics 
like color or gender. But in order to work effectively to 
further the goals of Title VII, employers must have a clear, 
operational definition of applicable legal rules. And those 
rules must not be so unworkable that the only practical 
alternative for employers is to operationalize “equal op­
portunity” by proportionate hiring of blacks, women and 
members of other protected groups.

In the wake of this Court's opinion in Watson v. Fort 
Worth Bank & Trust, 56 U.S.L.W. 4922 (June 29, 1988), 
vacating 798 F.2d 791 (5th Cir. 1986), employers are very 
uncertain about the practical, operational meaning of “dis­
parate impact” theory as applied to review a series of 
individual, subjective personnel judgments. And they fear 
that the only manageable way to comply with this new 
rule will be to hire “by the numbers”—a result which 
would stand the purpose of Title VII on its head.

The Court should seize the opportunity offered by the 
present case to restate the rule of Watson in terms which 
provide employers a clear, operational definition of its re­
quirements. Most importantly, those requirements must 
take into account the practical realities of countless per­
sonnel judgments made by fair-minded employers every 
single day. The Court should adopt the evidentiary stand­
ards outlined in the plurality opinion in Watson and spell 
them out in much greater detail, both with respect to the 
quality of a proper prima facie case and the nature of the 
intermediate burden to be carried by employers.

“Disparate impact” plaintiffs ought to be required to 
prove that a specific selection criterion disqualified a sig­
nificantly disproportionate number of individuals because 
of their membership in a protected class. Where the chal­
lenge is directed at a series of subjective personnel judg­
ments, requiring identification of a specific, discrete 
selection criterion is a particularly important part of the

301



6

foundation for a meaningful analysis. Here, Respondents’ 
prima facie proof fell far short. Respondents failed to show 
they were excluded from the at-issue jobs—in fact, the 
record reflects no adverse impact at all. Instead, 
Respondents relied entirely on their over-representation in 
jobs not at issue. And Respondents failed to show that 
any specific selection device caused their over-represen­
tation in not at issue jobs or denied them the privilege of 
being over-represented in the jobs at issue. Instead, 
Respondents relied on a shotgun approach alleging that a 
legitimate difference in the percentage of nonwhites in at- 
issue and not at-issue jobs was the result of the “cumu­
lative effect” of a variety of policies and practices. But 
Respondents never proved that this “statistical stratifi­
cation” was caused by any particular selection device. Ac­
cordingly, the first and third questions presented should 
be resolved in favor of Petitioners.

Employers defending “disparate impact” challenges to 
their use of nonstandardized, subjective selection criteria 
should not bear an intermediate burden of proving the 
“business necessity” of those criteria. Unlike standardized 
selection devices, subjective judgments of important per­
sonal characteristics like loyalty or tact are not amenable 
to a priori testing to determine whether they will dis­
qualify disproportionate numbers of protected individuals 
and, if so, to determine their relationship to business goals. 
Unless an employer's burden of proof could be satisfied 
simply by stating the opinion that the job relatedness of 
qualities like loyalty and tact is self-evident, such a burden 
would be unmanageable and would force employers into 
the untenable realm of proportional hiring. The Court must 
reject any evidentiary burden likely to produce this result.

302



7

ARGUMENT
A. In the Wake of W a tso n , Employers Must Be Per­

mitted To Judge The Personal Qualities Of Individ­
ual Job Applicants On A Case-By-Case Basis Without 
Being Saddled With Unmanageable Burdens
It is a standard tenet of personnel administration 
that there is rarely a single, 'best qualified’ per­
son for a job. An effective personnel system will 
bring before the selecting official several fully- 
qualified candidates who each may possess dif­
ferent attributes which recommend them for se­
lection. Especially where the job is an 
unexceptional, middle-level craft position, without 
the need for unique work experience or educa­
tional attainment and for which several well-qual­
ified candidates are available, final determinations 
as to which candidate is ‘best qualified’ are at 
best subjective.

Johnson v. Transportation Agency, Santa Clara County,
California, 480 U.S___ _ 94 L.Ed.2d 615, 636 n.17 (1987)
(quoting Brief for ASPA as Amicus Curiae in support of 
Respondents).

As a corollary to the foregoing principle, personnel 
professionals also know that the business of selecting em­
ployees from a pool of well-qualified finalists is not a sim­
ple, mechanical process. Far from it. In order to identify 
individuals who will best serve the needs of their enter­
prises employers must approach the task with flexibility 
and creativity. The scarcity of perfect applicants means 
that each individual’s strengths must be discounted by his 
or her weaknesses. Imperfect applicants must be judged 
in relation to one another and measured against the em­
ployer’s reasonable definition of job requirements rather 
than measured against some external yardstick of perfec­
tion. At the same time, employers must be prepared to 
recognize and credit unique or outstanding qualities pre-

303



8

sented unexpectedly. While the process of final selection 
can be disciplined by advance thinking about the kind of 
individual who is likely to succeed, these prejudgments 
must often give way to new information, the qualities of 
available applicants and the pressures of time.

Indeed, in a “service economy" where intangible per­
sonal characteristics are often critical to the definition of 
a quality employee, a judge's selection of a law clerk is 
an excellent model for considering the practical dynamics 
of employee selection. Each candidate for a clerkship must 
possess certain minimum qualifications summarized by the 
fact that he or she has earned a degree from an accredited 
law school. Some judges may require one year or more of 
prior clerkship experience. But beyond these narrow, “ob­
jective" criteria, the judge’s selection decision must be 
based on a series of subjective judgments about a wide 
range of factors. What was the quality of the candidate's 
law school, course selections and academic performance? 
How valuable was the candidate’s prior clerkship experi­
ence? What is said by those recommending the candidate’s 
selection and how much weight should be attributed to 
their views? Perhaps most importantly, what does the se­
lecting judge see and hear when he or she looks into the 
eyes of a hopeful candidate during a final interview? Is 
there common sense, commitment and clear thinking or 
distant self-importance?

Well-qualified candidates almost always outnumber the 
positions available. Minimum qualifications are almost al­
ways satisfied. Choosing the clerk who will work most 
effectively with the selecting judge is of critical importance 
to the success of work in the upcoming term. What se­
lection criteria will identify the very best clerk? Years 
later, will the selection decision be deemed illegitimate 
according to standards of equal employment opportunity 
which can be applied only after the total number of clerks 
selected is large enough to support a judgment about the 
significance of an overall statistical imbalance?

304



9

This Court's holding in Watson v. Fort Worth Bank & 
Trust, supra, (hereinafter “Watson”), changed those stand­
ards in ways which appear to be fundamental but which 
have not yet been fully revealed. Prior to Watson, selecting 
officials knew that candidates must not be treated differ­
ently because they are black or female or members of other 
protected groups. Employers knew they could not screen 
out candidates who, for example, lacked certain educa­
tional attainment or experience if those criteria excluded 
a disproportionate number of protected individuals but did 
not bear “a manifest relationship to the employment in 
question.” Griggs v. Duke Power Co., 401 U.S. 424, 432 
(1971). But, prior to Watson, this Court had “consistently 
used conventional disparate treatment theory . . .  to review 
hiring and promotion decisions that were based on the 
exercise of personal judgment or the application of inher­
ently subjective criteria.” Watson, 56 U.S.L.W. at 4925. 
Accordingly, employers were confident of their freedom to 
judge the personal qualities of individual candidates as long 
as those judgments were not tainted by unlawful prejudice.

Since Watson, that freedom is in doubt. Watson held 
that, after a period of years, individual personnel judg­
ments may be deemed unlawful if, viewed collectively, 
members of a protected group were selected less often 
than others. To return to our clerkship model, if a judge’s 
hiring decisions over the years selected male candidates 
significantly less often than female candidates, rejected 
males would be entitled to relief in the absence of proof 
of the “business necessity” of the criteria which produced 
those results.

The Watson result troubles employers for two reasons. 
The first is the fact of great uncertainty about the stand­
ards which will govern a retrospective "disparate impact” 
analysis of a series of individual, subjective personnel judg­
ments. The second is the prospect that, once defined 
clearly, those standards will prove unmanageable unless 
employees are chosen “by the numbers,” thus avoiding

305



10

statistical imbalances that can trigger a “disparate impact” 
analysis.

The case before the Court requires a clear, operational 
definition of the new rule announced in Watson. Without 
one, employers face the chilling uncertainty of knowing 
that something called “disparate impact” theory may be 
applied to individual personnel judgments but not knowing 
whether they should act in accordance with the evidentiary 
standards set forth in the Court’s plurality opinion or the 
fundamentally different view expressed in the concurring 
opinion authored by Justice Blackmun. ASPA urges the 
Court to spell out the rule of Watson in terms which 
provide clear guidance and which preserve the right of 
employers to make legitimate, subjective personnel judg­
ments.
B. Justice O’Connor’s Formulation Of Disparate Impact 

Analysis In W atson  Is Based On Principles Essential 
To A Workable Application Of The Model

If the “disparate impact” approach may be used to chal­
lenge individual, subjective personnel judgments, ASPA 
urges the Court to adopt the evidentiary standards for 
such cases outlined in the plurality opinion in Watson. That 
opinion expresses important principles which, when prop­
erly applied, should permit a workable balance between 
legitimate employer discretion and the goal of equal em­
ployment opportunity. 1

1. Because Statistical Imbalances Alone Do Not Support 
A Presumption Of Unlawful Discrimination, Evidence Of 
An Adverse Impact Must Be Linked To A Discrete Se­
lection Criterion

Justice O’Connor's discussion of evidentiary standards 
in Watson began by recognizing a problem inherent in Title 
VII challenges based on statistical disparities.

It is completely unrealistic to assume that unlaw­
ful discrimination is the sole cause of people fail-

306



11

ing to gravitate to jobs and employers in accord 
with the laws of chance. It would be equally un­
realistic to suppose that employers can eliminate, 
or discover and explain, the myriad of innocent 
causes that may lead to statistical imbalances in 
the composition of their work forces.

Watson, 56 U.S.L.W. at 4926 (citation omitted) (emphasis 
added). Indeed, Title VII expressly disclaims any require­
ment that employers prefer members of protected groups 
in order to avoid or compensate for numerical imbalances. 
42 U.S.C. § 2000e-2(j).«

Since a statistical imbalance must be linked to a specific 
cause in order to have any meaning in a Title VII case, 
the first, crucial burden borne by a “disparate impact" 
plaintiff is “isolating and identifying the specific employ­
ment practices that are allegedly responsible for any ob­
served statistical disparities." Watson, 56 U.S.L.W. at 
4927. In other words, our hypothetical male clerkship can­
didate must not only show that disproportionately few 
males were chosen—a preliminary fact which itself is no 
evidence of unlawful discrimination—but must also identify 
specifically what his prospective judge did that caused the 
disqualification of more male clerks than was expected by 
chance. Unless the focus is narrowed in this way, no foun­
dation exists on which to base a proper “disparate impact" 
analysis.

At this very first step of the analysis the rule intended 
by the Court in Watson can become blurred—and the bur- *

* “A new subsection 703(j) is added to deal with the problem of racial 
balance among employees. The proponents of .. . have carefully stated 
on numerous occasions that Title VII does not require an employer to 
achieve any sort of racial balance in his work force by giving prefer­
ential treatment to any individual or group. Since doubts have persisted, 
subsection(j) is added to state this expressly.” 110 Cong. Rec. S 12723 
(daily ed. June 4, 1964), Statement of Senator Humphrey, reprinted in 
the EEOC's Legislative History of Title VII and XI of Civil Rights 
Act of 1964, at 3005.

307



12

den on employers can become unmanageable—in the ab­
sence of a clear definition of the category of selection 
devices which are proper targets of a disparate impact 
attack. The Court’s opinions in Watson discuss the impact 
of employment “tests," “requirements" and “criteria," but 
also speak in terms of selection “practices," “procedures," 
and “systems." The former set of terms describes discrete 
selection devices which can be identified specifically and 
analyzed individually. The latter set describes aggregations 
of variables, each of which may include some factors which 
have had an impact on the selection of certain employees 
and other factors which have not.

“Disparate impact" analysis should only be applied to 
review discrete selection criteria which can be identified 
and analyzed individually. “Disparate impact" plaintiffs 
must not be permitted to challenge multifactor selection 
“procedures" or “systems" as if they were a single, in­
divisible “cause" of a statistical imbalance. See, e.g., New 
York City Transit Authority v. Beazer, 440 U.S. 568 (1979); 
see also AFSCME v. State of Washington, 770 F.2d 1401, 
1405 (9th Cir. 1985) (Kennedy, J.) (“Disparate impact anal­
ysis is confined to cases that challenge a specific, clearly 
delineated employment practice applied at a single point 
in the job selection process.")

The reasons for imposing such a limitation are obvious. 
Permitting a "shotgun" attack against an employer’s over­
all selection system would render meaningless the require­
ment of showing a causal link between a particular 
selection device and an adverse impact. Because, as Justice 
O'Connor wisely observed, it would be “unrealistic to sup­
pose that employers can . .  . discover and explain, the myr­
iad of innocent causes" that may have led to such a result, 
plaintiffs must bear the burden of identifying the specific 
criterion they believe caused unjust discrimination. Watson, 
56 U.S.L.W. at 4926. For example, if our hypothetical 
male clerkship candidate could simply attack the collective 
impact of all the selection criteria used by the judge, it

308



13

would always be the case that somewhere among that 
universe of criteria would be the criterion which caused 
the adverse impact. Without specification of a discrete cri­
terion, proof of causation would consist of nothing more 
than the simple claim that the judge’s criteria caused the 
impact. Moreover, shotgun attacks would require employ­
ers to justify many—perhaps all—facets of their selection 
systems even if an adverse impact was the product of only 
one element of that system. No purpose would be served 
by requiring our hypothetical judge to demonstrate the 
“business necessity” of, for example, a preference for can­
didates from ivy league law schools or candidates with 
clinical experience if in reality it was the recommendations 
of law school professors that caused the selection of a 
disproportionate number of women.

Similarly, a “disparate impact” plaintiff should not be 
permitted simply to attack the general practice of making 
a final selection decision based on a personal interview, 
as if the interview itself was a selection criterion. Once 
again, the entire concept of a causal link between a specific 
selection device and some adverse impact would be ren­
dered meaningless by this approach. The selection criterion 
that produced the impact may or may not have been among 
those applied during the interview. If it is not, analysis 
of the legitimacy of interviews as a selection device is 
completely useless. And even if the operative criterion was 
applied during the interview, analysis of the “business ne­
cessity” of the criterion cannot begin until it is identified. 
Instead of permitting plaintiffs to attack employers’ judg­
ments in a vague and general way, all of the criteria 
applied during personal interviews should be identified—a 
single interrogatory will accomplish this result—and courts 
should then require plaintiffs to show which of those cri­
teria caused the numerical imbalance observed.

If an employer cannot identify the criteria it used to 
distinguish candidates, the “disparate impact” model of 
analysis should not be applied. This does not mean that

309



14

an employer with a standardless selection system would 
prevail. Rather, it simply means that the review of such 
an employer's selection decisions should proceed under the 
“disparate treatment” model.

Contrary to Justice Blackmun’s concern that the lack 
of specific selection criteria might “shield [an employer] 
from liability,” an employer’s inability to articulate the 
basis for its selection decisions would leave it vulnerable 
to a variety of attacks. Watson, 56 U.S.L.W. at 4931, n.10 
(Blackmun, J., concurring). Indeed, that inability may re­
quire a judgment against the employer if a “disparate 
treatment” plaintiff had made out a prima facie case of 
discrimination. Texas Department of Community Affairs v. 
Burdine, 450 U.S. 248, 254 (1981) (“If the trier of fact 
believes the plaintiffs evidence, and if the employer is 
silent in the face of the presumption [of unlawful discrim­
ination], the court must enter judgment for the plaintiff 
because no issue of fact remains in the case.”). If unable 
to identify its selection criteria, the employer may even 
be in a worse position than if, under the “disparate im­
pact” model, it bore the onerous “burden of establishing 
that the absence of specified criteria was necessary for 
the proper functioning of the business.” Watson, supra, 
at 4931 n.8. Thus, the appropriateness of the “disparate 
impact” approach does not turn on an unprincipled pre­
diction about whether plaintiffs or employers are more 
likely to prevail. Rather, the important, enduring principle 
is that Congress never intended Title VII to require em­
ployers to adopt any external set of hiring criteria, much 
less an idealized set of “objective criteria carefully tailored 
to measure relevant job qualifications.” Watson, 56 
U.S.L.W. at 4931 (Blackmun, J., concurring). See, Steel­
workers v. Weber, 443 U.S. 193, 207 (1979) (Congress did 
not intend to limit traditional business freedom, even with 
respect to certain race-conscious affirmative action).

This Court has recognized “[t]he dangers of embarking 
on a course . . .  where the court requires businesses to

310



15

adopt what it perceives to be the ‘best’ hiring proce­
dures. . . Fumco Construction Corp. v. Waters, 438 U.S. 
567, 578 (1978). Instead, Title VII enforcement should ac­
cept employers’ selection devices for what they are and 
apply standards of review appropriate to test the legiti­
macy of the device chosen by the employer. If discrete 
tests or criteria have been used which systematically ex­
cluded protected individuals, “disparate impact’’ analysis 
may be applied in accordance with the rule of Watson. If 
no such tests or criteria can be identified, “disparate im­
pact’’ analysis is not a useful tool and, instead of forcing 
employers to adopt judicially approved selection devices, 
their employment decisions should be reviewed under the 
“disparate treatment’’ approach.

2. The First and Third Questions Presented Should Be Re­
solved In Favor of Petitioners

A straightforward application of the foregoing principles 
requires that the first and third questions presented be 
answered in favor of Petitioners.

The third question presented should be addressed first 
because it goes to the heart of the problem—namely, that 
Respondents failed to show any causal link between a dis­
crete selection device and a significant, disparate impact 
against minority applicants. Respondents simply alleged 
that the over-representation of minority employees in the 
not at-issue, cannery worker jobs resulted from the cu­
mulative effect of a variety of employment practices.7 De­
spite the Ninth Circuit’s apparent recognition that the 
“disparate impact” model should not authorize a “wide 
ranging attack on the cumulative effect of a Company’s 
employment practices.’ Spaulding [v. University of Wash­
ington, 740 F.2d 686, 707 (9th Cir. 1984)]”, the court’s en 1

1 The insufficiency of this allegation, even if a causal link to a specific 
criterion were proven, is discussed immediately below. S e e , i n f r a , at 
pp. 17-19.

311



16

banc opinion proceeded to hold that “practices which cause 
adverse impact may be considered individually and collec­
tively” Atonio, supra at fn.4, at 1486 n.6 (emphasis added). 
On remand, the original panel applied this holding in a 
way which relieved Respondents of their proper eviden­
tiary burdens and placed unmanageable burdens on 
Petitioners.

The statistics show only racial stratification by 
job category. This is sufficient to raise an infer­
ence that some practice or combination of prac­
tices has caused the distribution of employees by 
race and to place the burden on the employer to 
justify the business necessity of the practices 
identified by the plaintiffs.

Atonio, supra at fn.5, at 444 (emphasis added).

The Ninth Circuit thus proceeded with its “disparate 
impact” analysis in the absence of proof of any causal link 
between an adverse impact and a specific selection device. 
This rule would mean that every employer with a dispro­
portionately high number of protected individuals in any 
job category would be obligated to justify every selection 
device—even those used for different job categories—which 
plaintiffs might allege is somehow related to that numer­
ical imbalance. The more employment practices plaintiffs 
indict, the more their employers must defend. And if the 
justification for any particular practice falls short, the em­
ployer would risk liability regardless of the actual adverse 
impact of that practice. In the present case, for example, 
Petitioners demonstrated the business necessity of their 
“rehire preference” to the satisfaction of both the district 
court and the Court of Appeals but nevertheless risk li­
ability on the basis of practices with doubtful causal con­
nections to the proportion of whites and nonwhites in 
various jobs—practices like referring to a fish butchering

312



17

machine by the name given it by its inventor (the "Iron 
Chink”).8 *

The courts below erred in permitting a challenge based 
on cumulative effects and without proof of causation. That 
error should be reversed. With respect to the first question 
presented, the error of the lower courts was even more 
extreme.

This Court’s plurality opinion in Watson restated the 
obvious point that a causal link must be established be­
tween a specific selection device and a significant impact 
which is adverse to protected individuals.

Once the employment practice at issue has been 
identified, causation must be proved; that is, the 
plaintiff must offer statistical evidence of a kind 
and degree sufficient to show that the practice 
in question has caused the exclusion of applicants 
for jobs or promotion because of their member­
ship in a protected group.

Watson, 56 U.S.L.W. at 4927 (emphasis added). Below, 
Respondents made no showing that any selection device 
caused the disproportionate exclusion of minorities from 
at-issue jobs. Indeed, the district court found that minor­
ities were not under-represented in the at-issue jobs and 
that “in some instances, nonwhites are overrepresented in 
the jobs taken on a department-by-department basis.” Ato- 
nio, supra at fn.l, 34 E.P.D. at 33,829 (finding of fact 
123) (emphasis added).8 Rather than demonstrating an ad­
verse impact with respect to at-issue jobs, Respondents’ 
entire case hung on their ability to treat the over-repre­
sentation of minorities in not at-issue jobs as the proper

• See, A t o n i o ,  s u p r a  at fn.l, 34 E P D  at 33,826 (finding of fact 65).

* This was one of numerous findings of fact not credited by the court 
below, a practice in direct conflict with this Court’s decision in A n d e r s o n  

v . B e s s e m e r  C i t y , 470 U.S. 564 (1985).

313



18

foundation for a “disparate impact" analysis of devices 
used to select employees for different jobs.

Respondents’ contorted approach makes a shambles of 
the theory on which they rely. “Disparate impact" theory 
demands some connection between specific employment de­
vices and the hiring decisions produced by those devices— 
otherwise the indispensable element of causation is non­
existent. The theory demands some demonstration of an 
impact that is adverse—otherwise the claim is properly one 
of reverse discrimination brought by a different set of 
plaintiffs. And the theory requires proof of an adverse 
impact with respect to jobs at issue—otherwise there would 
be no limit to the burden borne by employers. If adverse 
impact in one job category could be used to challenge the 
selection devices used for an entirely different job cate­
gory, a showing of adverse impact anywhere in a facility 
would require proof of the “business necessity" of selection 
devices used everywhere in the facility. This would mean 
that even if our hypothetical judge selected male and fe­
male clerks in perfect proportion to their availability, the 
judge still would have to demonstrate the “business ne­
cessity" of every criterion used to select law clerks if, for 
example, he or she had employed a disproportionately high 
number of female secretaries.

The only conceivable rationale which may have led the 
Ninth Circuit to rely on the overrepresentation of non­
whites in one job category to require proof of the “business 
necessity" of criteria used to select employees in a dif­
ferent category in which nonwhites were not underrepre­
sented—and it would have been a twisted, faulty rationale— 
would have been a rationale suggesting that the high per­
centage of nonwhite cannery workers defined the “ex­
pected" percentage of nonwhites among noncannery 
workers. If this was the rationale, it was flat wrong. The 
district court specifically found that the two categories of 
jobs required different sets of skills and qualifications and 
that cannery workers were not part of the available labor

314



19

pool for noncannery jobs.10 This Court has held consistently 
that legitimate expectations about minority representation 
in particular job categories depend on the availability of 
individuals with the qualifications for the jobs in question.

[AJnalysis of a more specialized labor pool nor­
mally is necessary in determining under-repre­
sentation in some positions. If a plan failed to 
take distinctions in qualifications into account in 
providing guidance for actual employment deci­
sions, it would dictate mere blind hiring by the 
numbers, for it would hold supervisors to 
“achievement of a particular percentage of mi­
nority employment or membership . . .  regardless 
of circumstances such as economic conditions or 
the number of qualified minority applicants . . . "
Sheet Metal Workers v. EEOC, 478 U .S.__ , 106
S.Ct. 3019 (1986) (O’Connor, J., concurring in 
part and dissenting in part).

Johnson v. Transportation Agency, supra, 94 L.Ed.2d at 
633.

As a matter of law, statistical evidence showing a nu­
merical imbalance among employees filling one category 
of jobs cannot support a “disparate impact” assault on 
devices used to select employees for a different category 
of jobs. Accordingly, the first question presented should 
be decided in favor of Petitioners.
C. E m ployers Should N ot B ear th e  Burden O f P roving  

th e  B u sin ess N ecess ity  o f  N onstandardized  S e lec tion  
C riteria

Prior to Watson, it was sometimes said that a distin­
guishing feature of “disparate treatment” and “disparate

'•A t o n i o , s u p r a  at fn.l, (findings of fact 117 and 110).
Once again, the failure to credit these findings cannot be justified. S e e ,  
A n d e r s o n  v . B e s s e m e r  C i t y ,  s u p r a  at fn.9.

315



20

impact" analysis was the nature of the intermediate bur­
den on the employer once a plaintiff made out a prima 
facie case. In a “disparate treatment" context, the em­
ployer's intermediate burden is not a burden of proof but 
rather a burden of “articulation"—to explain clearly the 
nondiscriminatory reasons for its actions. Burdine, supra, 
at 255 n.9 and 260. In “disparate impact" cases challeng­
ing standardized selection criteria, this Court has char­
acterized the employer's intermediate burden variously as 
one of “showing"11 or “demonstrating"12 or “establishing"13 
or “proving"14 that its criterion bears “a manifest rela­
tionship to the employment in question."15

Now that Watson has opened an entire new category 
of employer activity to “disparate impact" review, char­
acterizations of the employer's intermediate burden arising 
in other contexts should not be applied automatically to 
this new category of cases. As discussed below, because 
of a fundamental difference between evidence available to 
employers to justify standardized versus nonstandardized 
criteria, the employer's intermediate burden should not 
rise to the level of a burden of proof by a preponderance 
of the evidence. Where "disparate impact" theory is ap­
plied to nonstandardized, subjective personnel judgments, 
employers should bear an intermediate burden of produc­
tion similar to that in a “disparate treatment" case. In 
these new cases, the intermediate burden on employers 
should be to produce evidence of a “manifest relationship" 
between their nonstandardized criterion and a legitimate 
business need.

11 G r i g g s ,  s u p r a , at 432.

11 B e a z e r ,  s u p r a , at 587.

" I d . , at 587 n.31.

M A l b e m a r l e  P a p e r  C o . v .  M o o d y , 422 U.S. 405, 425 (1975). 

11 G r i g g s ,  s u p r a , at 425.

316



21

1. By Their Very Nature, Nonstandardized Selection Cri­
teria Are Not Amenable to the Techniques By Which 
the Job Relatedness Of Standardized Criteria Is Proven 
Or Disproven

In a “disparate treatment" case, the ultimate issue is 
whether a selecting official intended to discriminate against 
an individual based on his or her membership in a pro­
tected class. Teamsters v. United States, 431 U.S. 324, 335 
n.15 (1977). The intent to discriminate is, of course, some­
thing the decision maker knows about at the time an em­
ployment decision is made. Discriminatory intent can be 
recognized and abandoned by decision makers seeking to 
comply with the law. In short, the “disparate treatment" 
approach embodies standards which leave no doubt about 
what the law requires and afford employers the oppor­
tunity to judge their behavior at the time an employment 
decision is made and to conform it to the law.

Similarly, in a “disparate impact" case challenging a 
standardized selection device—one, as in Griggs, which sys­
tematically disqualified a disproportionate number of pro­
tected individuals—an employer is able to assess its position 
before the device is actually used. By definition, a stand­
ardized selection device is one which will apply precisely 
the same measure in precisely the same way to as many 
candidates as necessary. A standardized device, therefore, 
can be tested prior to its implementation to determine 
whether it will impact adversely members of a protected 
group. If such an impact is observed, it can also be tested 
to determine whether it is a reasonably good predictor of 
success on a job or is otherwise justified by “business 
necessity." Accordingly, employers can assess a standard­
ized device a priori and decide whether to implement or 
abandon it. Indeed, the federal government has regulated 
this process for many years by means of guidelines in­
structing employers on how to carry out this assessment. 
See, Uniform Guidelines on Employee Selection Proce­
dures, 29 C.F.R. $ 1607 (1978).

317



22

However, an employer judging important personal qual­
ities of individual applicants based on nonstandardized, 
subjective criteria is in a fundamentally different position. 
By definition, nonstandardized devices—for example, judg­
ments during an interview about a candidate’s loyalty or 
tact—do not apply precisely the same measure in precisely 
the same way time after time. No matter how detailed 
the guidelines, different interviewers will have somewhat 
different notions of the meaning of nonstandardized cri­
teria like loyalty and tact. -The words they use to test 
these attributes will vary in subtle ways. Their assessments 
will depend to some extent on the course and content of 
conversation during the interview. Judgments may vary 
depending on whether the conversation stumbles into com­
plex or controversial topics, whether the prior interviewee 
seemed wonderful or impossible, whether the interviewer 
is eager to meet a potential employee or is bored with a 
lengthy selection process, and a host of other uncontrolled 
variables.

Under these circumstances, it would impose an unman­
ageable burden on employers to require them to prove that 
a nonstandardized criterion like loyalty or tact was essen­
tial to good performance on the job. For example, even 
if our hypothetical judge in search of a law clerk had the 
time to “pre-test” his or her judgments of loyalty and tact 
on a group of one hundred law school graduates to de­
termine whether an adverse impact would result, that ex­
ercise would be of precious little value because the judge 
could not control the application of those nonstandardized 
criteria to the next one hundred candidates so as to be 
confident that the results would be the same.

Moreover, how would our judge prove the “business 
necessity” of qualities like loyalty or tact? Since judgments 
of nonstandardized criteria cannot be quantified with con­
fidence, it would be impossible to construct a meaningful 
historical record which compared a clerk’s loyalty or tact­
fulness “scores”to other scores rating the clerk’s job per-



23

formance.1* If the judge had consistently hired clerks based, 
in part, on an assessment of their loyalty and tact, the 
only way to prove the necessity of these criteria would be 
to abandon them and watch for demonstrably inferior per­
formance by those newly hired.17

The practical impossibility of constructing a meaningful 
proof of the job relatedness of nonstandardized selection 
criteria contrasts sharply with the practical necessity of 
proving the job relatedness of standardized tests. The only 
way for employers to demonstrate the job relatedness of 
paper and pencil tests is by means of some form of val­
idation study. Assuming the study is not itself defective, 
its results will constitute proof, at a stated level of con­
fidence, that the test is either related to the jobs in ques­
tion or that it is not. The all-or-nothing quality of the 
results of such studies shrinks to the vanishing point the

'•Constructing such a record would be made even more difficult by 
the need to create a standardized measure to rate the quality of a 
clerk’s performance. As the plurality pointed out in W a t s o n :

(SJuccess at many jobs in which such qualities [including loyalty 
and tact] are crucial cannot itself be measured directly. Opinions 
often differ when managers and supervisors are evaluated, and 
the same can be said for many jobs that involve close cooperation 
with one’s co-workers or complex and subtle tasks like the pro­
vision of professional services or personal counseling.

Watson, 66 U.S.L.W. at 4926.
it Even this approach is only a theoretical, impractical possibility be­

cause (l)criteria like loyalty, tact and so on would have to be abandoned 
only one at a time in order to test the effect of each on performance, 
and (2)such a series of tests could not be completed in time to respond 
to a ’’disparate impact” challenge, particularly if only a small number 
of clerks were hired each year.

It should be noted that each of the practical problems faced by our 
hypothetical judge would be multiplied dramatically in the context of 
a large employer with a large number of selecting officials, each doing 
his or her best to judge critical personal qualities of a variety of ap­

plicants.

319



24

difference between a burden of producing evidence and a 
burden of proof. Thus, in the case of standardized selection 
devices, it is understandable that courts have sometimes 
required “proof’ rather than “evidence” of job related­
ness.18

Reasonably available evidence of the job relatedness of 
subjective, nonstandardized criteria typically will be 
suggestive rather than dispositive. In many cases, to im­
pose a burden of proof of “business necessity” with respect 
to this category of selection criteria would be to outlaw 
them. In mandating equal employment opportunity, Con­
gress never intended to outlaw the use of business judg­
ment in hiring or impose unmanageable burdens on 
employers to justify judgments not tainted by an intent 
to discriminate against members of protected groups.

2. The Plurality Opinion In Watson  Outlines Evidentiary 
Standards Which Recognize the Special Nature Of Le- 
gitim ate, Nonstandardized Criteria

In light of the practical dilemmas that would be faced 
by employers forced to prove the business necessity of 
subjective personnel judgments, the plurality in Watson 
interpreted the “manifest relationship” test of Griggs in 
a way that was perfectly appropriate. The plurality opinion 
properly rejected the notion that the Griggs test “implied] 
that the ultimate burden of proof can be shifted to the 
defendant.” Watson, 56 U.S.L.W. at 4927. Instead, the 
plurality opinion limited the intermediate burden on em­
ployers in cases of this type to a burden of production.

'* A S P A  does not mean to cast doubt on the plurality's important 
observation in W a t s o n  that this Court has never required employers to 
"introduce formal ‘validation studies’ showing that particular criteria 
p r e d i c t  a c t u a l  o n - th e - jo b  p e r f o r m a n c e . ”  W a t s o n , 56 U.S.L.W. at 4928 
(emphasis added). As the plurality illustrated with examples including 
W a s h i n g t o n  v . D a v i s , 426 U.S. 229, 250 (1976), the ability of a test to 
predict actual on-the-job performance is not a necessary element of the 
Court's definition of either test validity or job relatedness. See also, 
I d . , at 256 (Stevens, J., concurring).

320



25

Thus, when a plaintiff has made out a prima facie 
case of disparate impact, and when the defendant 
has met its burden of producing evidence that its 
employment practices are based on legitimate 
business reasons, the plaintiff must show that 
other tests or selection devices, without a simi­
larly undesirable racial effect, would also serve 
the employer’s legitimate interest in efficient and 
trustworthy workmanship.

Id. (citing Albemarle Paper Co., supra). The plurality con­
cluded that imposing a greater intermediate burden on 
employers would be to require more—in the case of non- 
standardized criteria not amenable to objective proof—than 
the Court had required in other contexts. See New York 
City Transit Authority v. Beazer, supra (methadone users 
properly excluded from nonsafety-sensitive jobs based on 
simple articulation of rationale for personnel policy); Wash­
ington v. Davis, supra (written test justified by simple 
rationale that test which predicted success at police train­
ing academy was “manifestly related” to police work de­
spite absence of demonstrated link between test and actual 
performance as a police officer).

The plurality was also correct in recognizing that:
[i]n the context of subjective or discretionary em­
ployment decisions, the employer will often find 
it easier than in the case of standardized tests 
to produce evidence of a ‘manifest relationship 
to the employment in question.’ It is self-evident 
that many jobs, for example those involving man­
agerial responsibilities, require personal qualities 
that have never been considered amenable to 
standardized testing.

Watson, 56 U.S.L.W. at 4928 (quoting Fumco, supra, at 
578). This conclusion simply recognized that, unlike com­
plicated paper-and-pencil examinations, the job relatedness

321



26

of certain nonstandardized criteria is apparent on their 
face. Compare the process for analyzing the job relatedness 
of a standardized written exam and a nonstandardized cri­
terion like tactfulness. There is no way to draw an im­
mediate conclusion about the job relatedness of a ten page 
test booklet containing dozens of questions. First, one must 
ask what knowledge did those questions seek? Was the 
format bilingual or did it automatically exclude non-English 
speaking people of color? What was the relationship, if 
any, between the subject matter of each question or the 
totality of the questions and the job at issue?

By contrast, there is nothing complicated or indirect 
about judging the “manifest relationship'’ between a per­
sonal quality like tactfulness and any position in which an 
employee is obligated to work with other people. This is 
a matter that can be judged as soon as the criterion and 
the position are identified. While some may quibble about 
just how important it is to employ a person with tact 
rather than a person who is rude, the relationship between 
the criterion and the work is self-evident.

Justice Blackmun’s response to the plurality opinion on 
this point is misguided. Justice Blackmun wrote:

It would make no sense to establish a general 
rule whereby an employer could more easily es­
tablish business necessity for an employment 
practice, which left the assessment of a list of 
general character qualities to the hirer’s discre­
tion, than for a practice consisting of the eval­
uation of various objective criteria carefully 
tailored to measure relevant job qualifications.
Such a rule would encourage employers to aban­
don attempts to construct selection mechanisms 
subject to neutral application for the shelter of 
vague generalities.

Watson, 56 U.S.L.W. at 4931 (footnote omitted).

322



27

In fact, it does make sense that an employer who chooses 
to implement a relatively complicated, standardized paper- 
and-pencil test may have some difficulty explaining the 
relationship between that test and good performance. It 
does make sense that the Duke Power Company had more 
difficulty showing the business necessity of its written 
qualifying exam than it would have had explaining why it 
wanted employees with common sense or ambition or any 
other personal quality the value of which is self-evident. 
Some employers may choose to shoulder a heavier burden 
because use of a standardized device has special value in 
their particular circumstances—for example, as a rough 
screen for large batches of applicants too numerous to 
interview. The plurality opinion in Watson simply recog­
nized that relatively sophisticated, standardized selection 
devices may require analyses of job relatedness that are 
more sophisticated than those required for common sense, 
subjective criteria.

Justice Blackmun's complaint also seems to ignore some 
compelling realities of employee selection. As a practical 
matter, employers are not able to choose freely between 
selection devices which are "objective” and "neutral” and 
those which are "subjective” and "discretionary.” Certain 
personal qualities "have never been considered amenable 
to standardized testing.” Watson, 56 U.S.L.W. at 4928. If 
employers are to assess these qualities—and, of course, 
they must—they must not be saddled with unmanageable 
risks. Justice Blackmun’s approach failed to address in a 
practical way how employers would manage an interme­
diate burden of proof of the "business necessity” of non- 
standardized criteria. He cited an amicus brief filed by the 
American Psychological Association in support of Ms. Wat­
son suggesting that such criteria are amenable to "psy­
chometric scrutiny” but did not explain how such scrutiny 
would work and did not recognize the great expense of 
such a program if, indeed, it is workable at all. Justice 
Blackmun’s approach did not deal with the likelihood that

323



28

employers would be forced to avoid the expense and bur­
den of psychometrics by simply hiring “by the numbers.” 
In fact, Justice Blackmun relied on Professor Bartholet’s 
discussion of the feasibility of validating nonstandardized 
assessments, a discussion in which Professor Bartholet rec­
ognized that “quota or racially proportionate hiring” may 
be the result and, indeed, concluded that racially propor­
tionate hiring “seems an appropriate solution.” Bartholet, 
Application of Title VII to Jobs in High Places, 95 Harv. 
L. Rev. 945, 1026-7 (1982).19

3. The Second Question Presented Ought To Be Resolved 
In Favor of Petitioners

On remand from the decision en banc, the Ninth Circuit 
panel cited Respondents’ allegation that “the lack of ob­
jective job qualifications and the consequent hiring on the 
basis of subjective evaluations has an adverse impact on 
nonwhites in the canning industry.” Atonio, supra at fn. 
5, at 446. The panel’s discussion of this claim, however, 
is somewhat confusing in that it appears to direct the 
district court to “analyze whether these qualifications were 
actually applied in a nondiscriminatory manner.” Id. This

‘•We believe that the minority opinion in Watson was particularly 
misguided in suggesting that the “business necessity” of a selection 
device may be disproven by evidence that, in a particular case, it “failed 
in fact to screen for the qualities identified as central to successful job 
performance.” Watson, 56 U.S.L.W. at 4930 n.6. The opinion noted 
that one of Ms. Watson’s competitors, Mr. Kevin Brown, performed 
poorly after he was selected for the position sought by Watson. Such 
anecdotal evidence should carry no weight in judging the legitimacy of 
a selection device. The legitimacy of a college-degree requirement, for 
example, should be unaffected by the fact that a particular college 
graduate failed in a job after being selected over someone without a 
college degree. Standing alone, an individual performance says nothing 
about whether the selection device was legitimate or effective. The 
rejected nondegree candidate may have failed in the job much more 
quickly or seriously. An effective selection device promises to be suc­
cessful in the long run-it does not guarantee successful performance 
by each and every employee selected.

324



29

appears to be an analysis appropriate to a “disparate treat­
ment" case. Yet, the panel’s discussion of subjective cri­
teria closed with the statement “ [finally, and most 
importantly, the court must make findings as to the job­
relatedness of the criteria actually applied.” Id.

Despite this apparent confusion, one thing is clear. If 
the challenged practice of using subjective selection criteria 
is analyzed below according to the theory of “disparate 
impact,” the Ninth Circuit has held that Petitioners must 
"prove the job relatedness or business necessity of the 
practice.” Id., at 442. In fact, Petitioners should never 
have to address the issue of business necessity because, 
as discussed in Section B above, Respondents have ne­
glected their prima facie burden of proving a causal link 
between a specific subjective criterion and a significant 
adverse impact.

Nevertheless, if this matter is remanded for any purpose 
which may implicate the matter of Petitioners’ interme­
diate burden, this Court ought to issue clear instructions. 
If the subject of Respondents’ “disparate impact” chal­
lenge is a nonstandardized, subjective selection criterion, 
Petitioners’ intermediate burden should be to produce evi­
dence of a “manifest relationship” between that criterion 
and a legitimate business need.

325



30

CONCLUSION

For all of the foregoing reasons, the judgment of the 
court below should be reversed.

Of Counsel: 
Breed, Abbott & 

Morgan

Respectfully submitted,

•Lawrence Z. Lorber 
J. Robert Kirk 
International Square 
1875 Eye Street, N.W.
Suite 1000
Washington, D. C. 20006 
(202) 466-1100

Attorneys for Amicus Curiae, 
American Society for Personnel 
Administration
(•Counsel of Record)

September 9, 1988

326



No. 87-1387

In The
Supreme Court of the United States

October Term, 1987

-------------- o--------------

WARDS COVE PACKING COMPANY, INC., 
CASTLE & COOKE, INC.,

Petitioners,
v.

FRANK ATONIO, et al.,
Respondents.

--------------- o---------------

ON WRIT OF CERTIORARI TO 
THE UNITED STATES COURT OF APPEALS 

FOR THE NINTH CIRCUIT

• ----------------------------o ---------------------------

BRIEF FOR THE CENTER FOR CIVIL RIGHTS 
AS AMICUS CURIAE IN SUPPORT OF PETITIONERS

o

ARTHUR H. ABEL 
FAEGRE & BENSON 
2200 Norwest Center 
90 South Seventh Street 
Minneapolis, MN 55402 
(612) 336-3000

* Counsel of Record

CLINT BOLICK*
JERALD L. HILL 
MARK J. BREDEMEIER 
LANDMARK LEGAL FOUNDATION 
CENTER FOR CIVIL RIGHTS 
107 Second Street, N.E.
Washington, D.C. 20002 
(202) 546-6045

327



1

TABLE OF AUTHORITIES ............................................ ii
INTEREST OF AMICUS CURIAE ...............................  1
SUMMARY OF ARGUMENT ................................ ........  2
ARGUMENT ...........................................................................  3

THE ANALYTICAL FRAMEWORK FOR AD­
V ER SE IMPACT THAT THE PLURALITY AR­
TICULATED IN W A T S O N  v. F O R T  W O R T H  
B A N K  A N D  T R U S T  SHOULD BE APPLIED TO 
BAR R ESPO N D EN TS’ BROAD, ILL-DEFINED  
CHALLENGES TO PET ITIO N E R S’ BU SIN ESS  
DECISIONS .........................................................................  3

A. The Plurality’s Opinion in W atson  Properly
Recognized that the Rationale Underlying 
Title V II ’s Evidentiary Burden is the Same 
for All C a ses.................................   5

B. The P lurality’s Opinion in W atson  Properly
Adapted the Traditional Adverse Impact 
Theory to the Circumstances of Subjective 
Decisionm aking..................................................    12

C. The Ninth Circuit’s Decision Reversing the 
Trial Court’s Judgment for Petitioner’s Con­
flicts with the Plurality’s Opinion in W atson  16

CONCLUSION ............................ ...... ............................       20

T A B L E  O F  C O N T E N T S
P a g e

328



11

T A B L E  O F  A U T H O R I T I E S
P a g e

Cases
Albemarle ra fter  Co. v. Moody, 422 U.S. 405 (1975)...4,10,

11,15
Anderson v. L iberty  Lobby, hie,., 477 U.S. 242 (1986) 8
Atonio v. W ards Cove Packing Co., 827 F.2d 439 

(9th Cir. 1987) ........................................ ...........16,17,18
Atonio v. W ards Cove Packing Co., 810 F.2d 1477

(9th Cir. 1987) (en banc) ........................................... 6
Atonio v. W ards Cove Packing Co., 34 Einpl. Prac.

Dec. (CCH) 33,821 (W.I). Wash. 1983) ...................  19
Board o f Trustees o f Keene S ta te  College v.

Sweeney, 439 U.S. 24 (1978) ........................... .........  7
Connecticut v. Teal, 457 U.S. 440 (1982) .....................4,15
Dothard v. Rawlinson, 433 U.S. 321 (1977) ................. 4,15
Furnco Construction Co. v. W aters, 438 U.S. 567

(1978) .. ........ ......... ... -.................................. -...5,8,10,14
Griggs v. Duke Power Co., 401 U.S. 424 (1971)....... Passim

Hazelwood School Dist. v. United S ta tes, 433 U.S.
299 (1977) .....................................................-......... - 9,15

McDonnell Douglas Cory. v. Green, 411 U.S. 792
(1973) .. ....... ...... ......... ......... -....................... — ......5,7,8

New York C ity Transit A uthority  v. Beazer, 440 
U.S. 568 (1979) .................... -........................-........4, 7,11

Team sters v. United S ta tes, 431 U.S. 324 (1977)...... . 7,9
Texas D epartment o f Com m unity A ffa irs  v. Bur- 

dine, 450 U.S. 248 (1981) .................- -..... -...7, 8, 9,10
United S ta tes Postal Service v. Ailcens, 460 U.S.

711 (1983) .. ....................................... ............. — ...-  6

329



Ill

United Steelw orkers o f Am erica v. Weber, 443 U.S.
193 (1979) .............................................................................  13

W ashington v. Davis, 426 U.S. 229 (1976) ...................... 4,11
W atson v. Fort W orth Dank and Trust, — U.S. —,

108 S.Ct. 2777 (1988) ....................................................Passim

Other A uthorities 

R ules of E vidence
Fed. R. Evid. 3 0 1 ......................................................... ......... - 10

Legislative M aterials
110 Cong. Rec. 13,076-79 (1964) (discussion be­

tween Sen. Ervin and Sen. Cooper) ........... ........ ...._... 13

110 Cong. Rec. 13,080 (1964) (remarks of Sen. 
Humphrey) ................................ -....................... ............. ...5, 13

Treatises and A rticles

F. James & G. Hazard, Civil Procedure (2d ed.
1977)  ..........................................................-.......— ...  7

Lerner, E m ploym ent D iscrimination: A dverse Im ­
pact, Validity, and E quality  1979 Sup. Ct. Rev. 1 7 .... 12

Lerner, W ashington v. Davis: Q uantity, Quality 
and E quality  in Em ploym ent Testing, 1976 Sup.
Ct. Rev. 263 ...................-.................... ...........-...........— ...  6

B. Schlei & P. Grossman, E m ploym ent D iscrim ina­
tion Law  (2d ed. 1983) .................................... ................. 11

9 J. Wigmore, Evidence (3d ed. 1940) ....... ............ ........  10

T A B L E  O F  A U T H O R I T I E S — C o n tin u ed
P a g e

330



III. AT THE REBUTTAL STAGE OF A DISPA­
RATE IMPACT CASE, AN EMPLOYER NEED 
ONLY SHOW THAT ITS SELECTION DE­
VICES ARE REASONABLE IN LIGHT OF 
THE JOB AT ISSUE AND THE NATURE OF

iv

TABLE OF CONTENTS— Continued
Page

THE BUSINESS.....................................................  24

CONCLUSION .....................................................................  28

331



No. 87-1387

--------------- o----------------

In The
Supreme Court of the United States

October Term, 1987

------------- o-------------

WARDS COVE PACKINO COMPANY, INC., 
CASTLE & COOKE, INC.,

v.
Petitioners,

FRANK ATONIO, et al.,

Respondents.
--------------- o--------------

ON WRIT OF CERTIORARI TO 
THE UNITED STATES COURT OF APPEALS 

FOR THE NINTH CIRCUIT

--------------- o--------------

BRIEF FOR THE CENTER FOR CIVIL RIGHTS 
AS AMICUS CURIAE IN SUPPORT OF PETITIONERS

------------- o------------

INTEREST OF AMICUS CURIAE

The Landmark Legal Foundation Center for Civil
Rights is a public interest law center dedicated to pro­
moting the core principles of civil rights: equality under 
law and fundamental individual liberties. A vital aspect

1

332



2

of this mission is defending the integrity of civil rights 
laws in order to give meaning to the precious popular 
consensus expressed in those laws.

------------- o-------------

SUM M ARY OF ARGUM ENT

In W atson v. F ort W orth Bank and Trust, — IJ.R. —, 
108 S.Ct. 2777 (1988), this Court extended the statistical 
principles behind the adverse impact theory to reach sub­
jective employment practices under Title VII. That ex­
tension, as carefully delineated in Justice O’Connor’s 
plurality opinion, was entirely consistent with the basic 
principles established in prior Title VII cases, which his­
torically defined just two types of analysis: the adverse 
impact and disparate treatment tests.

The vital contribution of the W atson  plurality was to 
harmonize these two theories of proof, which previously 
were evolving in the lower courts in analytically incon­
sistent and sometimes contradictory ways. The W atson  
plurality demonstrates that just as there is but one ob­
jective in Title VII cases—to identify discriminatory em­
ployment practices—so is there a single coherent method 
of analysis, of which adverse impact and disparate treat­
ment are distinct but overlapping variants.

The adverse impact theory merely describes a method 
of prim a facie analysis based on the use of statistics. A l­
though much of the dictum in Griggs v. Duke Power Co., 
401 U.S. 424 (1971), is limited to the unique context of 
that case, its prima facie standards are transferable to

333



3

any case susceptible of proof by statistical inference. But 
because vague statistical challenges based on subjective 
decisionmaking have the capacity to “chill” an employer’s 
nondiscriininatory personnel judgments, this Court should 
adopt the plurality’s standards in W atson  and apply them 
to this and other such cases.

--------------- o--------------—

ARGUMENT

THE ANALYTICAL FRAMEWORK FOR AD­
VERSE IMPACT THAT THE PLURALITY AR­
TICULATED IN WATSON v. FORT WORTH 
BANK AND TRUST SHOULD BE APPLIED TO 
BAR RESPONDENTS’ BROAD, ILL-DEFINED 
CHALLENGES TO PETITIONERS’ BUSINESS 
DECISIONS
In W atson v. F ort W orth  Bank, and T ru s t , — U.S. — , 

108 S.Ct. 2777 (1988), this Court held for the first time 
that the adverse impact theory for proving Title VII dis­
crimination theoretically reached employment decisions 
based on subjective criteria. However, cognizant of the 
potential “chilling effect” that such an extension might 
have on legitimate business practices, the plurality care­
fully circumscribed that theory in order to keep it “within 
its proper bounds.” Id. at 2788. The plurality’s close ex­
amination of the theoretical foundations of Title V II’s 
evidentiary burdens was rationally conceived and should 
be applied here—for the respondents advance sweeping, 
ill-defined claims of subjective discrimination, and take 
precisely the shotgun approach to litigation that W atson's 
careful analysis was intended to proscribe.

334



4

Indeed, it was this spectre of freewheeling litigation 
practice that the Bank in W atson  raised, warning that a 
wholesale extension of the traditional adverse impact the­
ory (outlined in Griggs v. Duke Power Co., 401 U.S. 424 
(1971)) would engender an onslaught of nebuously framed 
civil rights claims, against which employers would find it 
impossible to defend without surreptiously adopting 
schemes for “preferential treatment.” See W atson, 108
S.Ct. at 2786.1

This litigation problem was slight when the adverse 
impact theory was confined to the traditional Griggs-type 
scenario, where seemingly arbitrary (but facially neutral) 
objective “measuring devices” were involved.2 But mind-

Indeed, one commentator complains that adverse impact 
has been applied indiscriminately “to cases arising out of vastly 
different factual contexts, making the burden of proving dis­
criminatory effects weightless, and the [employer's! burden . . . 
onerous, at times impossible." Lerner, Washington v. Davis: 
Quantity, Quality and Equality in Employment Testing, 1976 
Sup. Ct. Rev. 263, 267.

Griggs involved the seemingly arbitrary use of standard­
ized employment tests that were administered to all employees 
equally but which had a substantially adverse impact on the 
passage rate of blacks. In Chief justice Burger's oft-quoted 
phrase, the facially neutral and otherwise objective tests were 
illegal under Title VII because they acted like “built-in head­
winds" against minority groups and were “unrelated to meas­
uring job capability." 401 U.S. at 432. 5ee also Albemarle 
Paper Co. v. Moody, 422 U.S. 405 (1975) (written aptitude tests 
and diploma requirements); Washington v. Davis, 426 U.S. 229 
(1976) (written test of verbal skills); Dothard v. Rawlinson, 433 
U.S. 321 (1977) (height and weight requirements); New York 
City Transit Authority v. Beazer, 440 U.S. 568 (1979) (rule against 
employing drug addicts); Connecticut v. Teal, 457 U.S. 440
(1982) (written examination).

During the period that these decisions were written, it is 
clear that the Court did not intend the Griggs analysis to apply

(Continued on following page)

335



5

ful that a mechanical extension of Griggs into the sphere 
of subjective decisionmaking would “lead in practice to 
perverse results” that were antithetical to Title V IPs goal 
of employment opportunities based on qualifications,3 the 
plurality in W atson  carefully harmonized the traditional 
Title VII analyses of adverse impact and disparate treat­
ment (outlined in McDonnell Douglas Corp. v. Green, 411 
U.S. 792 (1973)).

A. The Plurality’s Opinion in Watson Properly 
Recognized that the Rationale Underlying 
Title VII’s Evidentiary Burden is the Same 
for All Cases.

1. One of the breakthrough of the plurality’s opinion 
in W atson  was a more careful explication of the eviden­
tiary considerations appropriate to proving discrimination 
through adverse impact. The prior absence of a unified 
adverse impact framework was particularly vexing in the 
area of subjective decisionmaking. Although the Court 
had gone far toward analyzing such cases in the past, see, 
e.g., Furnco Construction Corp. v. W aters, 438 U.S. 567 
(1978), it never previously considered how in this context 
the use of statistical data from the adverse impact theory 
should fit into the evidentiary scheme.

(Continued from previous page)
fully outside of the narrow context of cases where employers 
allowed test results, or other "fixed measures," to control their 
personnel decisions. See, e.g., Griggs, 401 U.S. at 433, 436; 
McDonnell Douglas Corp. v. Green, 411 U.S. 792, 802 n.14
(1973); Dolhard, 433 U.S. at 340 (Rehnquist, J., concurring).

To quote Senator Humphrey, "what [Title VIII does . 
is simply to make it an illegal practice to use race as a factor 
in denying employment. It provides that men and women shall 
be employed on the basis of their qualifications . . . "  110 
Cong. Rec. 13,088 (1964).

336



6

This problem has been a subject of enormous contro­
versy through the years, and one that constantly has 
plagued both courts and commentators.4 But we believe 
that the analysis of Watson’s plurality does much to elim­
inate that confusion and to stake out the neutral principles 
on which future litigants may rely.

2. It now seems clear that, whatever the chosen meth­
od of proof, the Court views the “ultimate determination 
of factual liability” as truly the same for all civil rights 
cases. United States Postal Service v. Aikens, 4G0 U.S. 
711, 718 (1983) (Blackmun, «T., concurring). At bottom, 
the plaintiff must adduce sufficient evidence to imply and 
ultimately prove that a particular employment practice 
discriminates on the basis of race, color, gender, national 
origin, or religious preference. See Aikens, 460 U.S. at 715; 
Watson, 108 S.Ct. at 2790.

4 5ee, e.g., Atonio v. Wards Cove Packing Co., 810 F.2d 
1477, 1480-81 & n.1 (9th Cir. 1987) (en banc) (discussing the 
differing views throughout the Circuits); id. at 1491-92 & n.4 
(five judges concurring) (similar discussion). One commentator, 
particularly frustrated with this state of affairs, has written:

The trouble with [the “traditional"! categories of [Title 
Vll's analysisl is that few cases with multiple plaintiffs fit 
neatly or exclusively into one category or the other. Most 
cases can be placed in either, and cases are now won or 
lost, depending upon the pigeonhole in which they are 
placed. The whole process begins to bear a disquieting 
resemblance to the bad old writ-of-action days when clev­
erness in juggling legal forms counted more heavily than 
the substance of the cases. This unfortunate impression 
is reinforced by the fact that the Court itself has begun to 
juggle the categories in arbitrary ways in order to get re­
sults it wants in particular cases.

Lerner, Employment Discrimination: Adverse Impact, Validity, 
and Equality, 1979 Sup. Ct. Rev. 17, 29-30 (footnote omitted).

337



7

Thus, in the first stage of all Title VII cases, a plain­
tiff who seeks to prove discrimination through an indirect 
showing must tender enough evidence to create a legal in­
ference of discrimination. See Teamsters v. United States, 
431 U.S. 324, 358 (1977). Although the precise formula 
often will vary according to the facts of each case,5 the 
plaintiff must come forth with sufficient proof from which 
a reasonable fact-finder can infer causation. See New York 
City Transit Authority v. Beazer, 440 U.S. 568, 584 (1979).

If properly supported, this inference will attain the 
status of a legal presumption and will shift to the defen­
dant the common law burden of rebuttal. Texas Depart­
ment of Community Affairs v. Burdine, 450 U.S. 248, 254 
& n.7 (1981). See also F. James & 0. Hazard, Civil Pro­
cedure §7.9, p. 225 (2d ed. 1977). At this “second stage,” 
the employer must produce just enough admissible evidence 
to meet the presumption and create a “genuine issue of 
fact” as to whether an employment prastice is based on 
legitimate factors, Burdine, 450 U.S. at 254; is “reason­
ably related to the achievement of some legitimate goal,” 
Furnco, 438 U.S. at 578; or otherwise has “a manifest re­
lationship to the employment in question,” Griggs, 401 U.S. 
at 432.6

Compare McDonnell Douglas Corp. v. Green, 411 U.S. at 
802-03 with Teamsters v. United States, 431 U.S. 324, 340 (1977) 
(both cases stating that the nature of the evidentiary burdens, 
including the use of statistics, will depend on the particular 
facts involved).
6 Of course, the employer need not actually convince the 
court that it acted for these reasons. See Board of Trustees of 
Keene State College v. Sweeney, 439 U.S. 24, 25 (1978). The 
evidentiary burden merely is designed to rebut the presumption 
of unlawful conduct and, thereby, "focus the issues" for the 
plaintiff's ultimate burden of proof. See, eg., Burdine, 450 
U.S. at 253.

338



8

3. The coherence and appeal of this neutral approach 
is obvious. But only recently has its efficacy become ap­
parent. For a long time both courts and commentators 
were constrained by the outlines of Griggs and McDonnell 
Douglas, whose specific analytical guidance would not eas­
ily accommodate the potentially vast scope of subjective 
discrimination. As discussed below, these cases present 
variations within what should be one analytical continuum. 
However, because they were treated separately instead of 
together, Title VII’s modes of analysis wrongly came to 
be viewed through a bipolar lens. This problem was ex­
acerbated by the fact that the nature of the employer’s 
evidence is controlled almost entirely by the evidence it 
seeks to rebut.

For example, in the traditional disparate treatment 
case, this Court has characterized the plaintiff’s initial 
burden as simply to show that he was qualified, but re­
jected, for a job that someone similarly situated but out­
side of the protected Title VII class later received. This 
circumstantial showing raises an inference of unlawful 
discrimination. And, “if the employer remains silent in 
the face of the presumption, the court must enter judgment 
for the plaintiff because no issues of fact remain in the 
case.” Bur dine, 450 U.S. at 254 (footnote omittted).7 The 
employer’s burden on rebuttal, then, is to “raise[] a gen­
uine issue of fact,” id. at 254, by “articulating] some legiti­
mate, non-discriminatory reason” on which its subjective 
personnel decision was based. Furnco, 438 U.S. at 578.

7 The evidentiary test for creating a genuine issue of dis­
puted fact is discussed generally in Anderson v. Liberty Lobby, 
Inc., 477 U.S. 242, 254-55 (1986).

339



9

Turning to the traditional adverse impact case, one 
view is that the plaintiff must meet a higher initial burden 
than in the disparate treatment case, which the defendant 
then must “disprove.” See Watson, 108 S.Ct. at 2792 
(Blackmun, J., concurring). But there is no apparent rea­
son for such a rule, except that it is an outgrowth of the 
unique facts in Griggs. The better view, analytically, and 
one suggested by Watson’s plurality is that a plaintiff must 
make the same initial showing in both cases: to tender 
enough evidence from which a court reasonably may infer 
illegal discrimination “under [the] circumstances.” Bur- 
dine, 450 U.S. at 253; cf. id. at 254 n.7. If this seems more 
difficult in the impact case, it merely is because statistical 
proof is so open to misuse that the law will not permit its 
admission unless (1) the proper foundation is made and 
(2) the impact is sufficiently “significant” that an infer­
ence of causation is reasonable. See, e.g., Hazelwood 
School Dist. v. United States, 433 U.S. 299, 310-13 (1977); 
Teamsters v. United States, 431 U.S. at 339-40.

Once this showing has been made and a presumption 
raised, the employer’s evidentiary burden on rebuttal is 
identical to that required in a traditional disparate treat­
ment case.8 In short, the employer must justify its con­
duct by showing that its business practice “is reasonably 
related to the achievement of some legitimate goal,” Furn-

8 See Burdine, 450 U.S. at 254 n.7 ("(On the Title VII con­
text we use 'prima facie case' . . .  to denote . . . only the es­
tablishment of a legally mandatory, rebuttable presumption 
. .  ..") (emphasis added).

340



10

co, 438 U.S. fit 578,9 or otherwise is the product of some 
“business necessity,” Griggs, 401 U.S. at 431. See W atson, 
108 S.Ct. at 2790; cf. Albemarle Paper Co. v. Moody, 422 
U.S. 405, 425 (1975) (a traditional impact case relying on 
McDonnell Douglas, 411 U.S. at 802, to describe this 
“burden”).10

In the usual case, meeting this obligation is fairly 
simple, because the targeted decisionmaking practice often 
will have a facially reasonable relationship to the job in 
question. Sec W atson, 108 S.Ot. at 2791. However, in 
cases like Griggs, where employers substitute arbitrary

9 The employer need not assume a burden of proof or a 
duty to persuade; rather, it merely must carry a burden of pro­
duction— i.e., “of going forward with evidence . . .  to meet the 
presumption." See Fed. R. Evid. 301. See generally 9 ). Wigmore, 
Evidence § 2491 (3d ed. 1940).
10 Three Justices in Watson would cast the burdens some­
what differently. They would hold that in the traditional dis­
parate treatment case the employer merely must "produce" re­
buttal evidence, but that in a traditional adverse impact case it 
must "prove" a business justification. Watson, 108 S.Ct. at 2792. 
The problem with this analysis is that it is based on loose lan­
guage, not cogent logic. In cases that pre-date Watson, this 
Court regrettably has used words like "proof" and "prove" to 
define the second stage of a Title VII inquiry, when the con­
text of those cases reveals that the Court did not intend to give 
those terms their full technical sway. See Burdine, 450 U.S. at 
254 n.7 (suggesting precisely this point "in the Title VII con­
text") (emphasis added).

For example, in Furnco Construction Corp. v. Waters, 438 
U.S. 567 (1978), the Court said that the employer's burden "is 
merely that of proving that he based his employment decision 
on a legitimate consideration, and not an illegitimate one such as 
race." Id. at 577 (emphasis added). But three sentences later, 
the Court explained that this so-called "proof" only needed to 
"dispel the adverse inference." Therefore, "the employer need 
only articulate some legitimate, nondiscriminatory reason" for 
the decision. Id. at 578 (emphasis added) (citation omitted).

341



11

and seemingly unnecessary “ employment tests” for their 
business judgment, it may in fact be more difficult for the 
employer to defend itself. This difficulty is not because 
the employer must “ disprove” discrimination or other­
wise meet a higher evidentiary standard. Rather, it is the 
natural consequence of justifying the rigid use of an em­
ployment test that appears unrelated to the job in ques­
tion. Cf. Griggs, 401 U.S. at 431 (employer must show 
that its employment practice is “ manifestly related to job 
performance” ).11

11 Much literature exists describing the notion of test "vali­
dation," to which the Court in dictum gave a passing nod in 
Albemarle Paper Co. v. Moody, 422 U.S. 405, 426-29 & n.23 
(1975). See generally B. Schlei & P. Grossman, Employment Dis­
crimination Law ch. 4 (2d ed. 1983). However, the Watson 
plurality specifically observed that, under this Court's prior hold­
ings, "employers are not required . . .  to introduce formal 'val­
idation studies' showing that particular criteria predict actual 
on-the-job performance." Watson, 108 S.Ct. at 2790 (citing 
New York City Transit Authority v. Beazer, 440 U.S. 568, 587
(1979) (flat rule against employing drug addicts upheld because 
the Court considered it obvious that "legitimate employment 
goals of safety and efficiency" were served), and Washington 
v. Davis, 426 U.S. 229, 250 (1976) (written test upheld because 
it was related to success at the police academy, "wholly aside 
from [the test's! possible relationship to actual performance as 
a police officer")).

This interpretation has been hailed as a rational legal ap­
proach by one lawyer-psychologist, who observes that:

All recognized scientific validation methods require 
the use of elaborate, formal procedures which are difficult, 
time-consuming, and costly. . . .  [In making their employ­
ment decisions, what most employers) have relied upon 
instead is what psychometricians call "face validity."

Face validity is . . .  a modern name for the basic, cen­
turies-old standard of Anglo-American law— reasonableness 
— and business and factory managers are hardly the only

(Continued on following page)

342



12

B. The Plurality’s Opinion in Watson Properly 
Adapted the Traditional Adverse Impact 
Theory to the Circumstances of Subjective 
Decisionmaking1.

1. The traditional adverse impact test, as articulated 
in Griqqs, was designed to discourage the use of “ artifi­
cial, arbitrary, and unnecessary barriers to employment” 
that bad an illegally discriminatory impact under Title 
VII. Grip qs, 401 U.S. at 431. Moreover, Griqqs was con­
cerned mostly with curbing the use of “ testing or measur­
ing procedures,” where employers gave such devices “ con­
trolling force” in the workplace. Id. at 436. In such cases, 
where employers abdicate their judgment to seemingly ar­
bitrary measuring devices, ordinary deference to employer 
judgments does not necessarily attach. See id.

However, Griqq.s does not supplant an employer’s 
right to make qualitative business judgments in the work­
place. To the contrary, this Court recognized in Griqqs 
that Title VII “ expressly protects the employer’s right to 
insist that any prospective applicant . . . must meet the

(Continued from previous page)
ones who rely upon it in selecting people for jobs. Face 
validity or reasonableness is what courts, legislatures, and 
the professions also rely upon when they insist that a law 
degree is required for the practice of law, a psychology 
degree for the practice of psychology, or training in educa­
tion for the practice of teaching. These reauirements have 
never been validated. They probably could not be vali­
dated. Face validity has simply been accepted and enforced 
on the basis of its inherent plausibility for jobs enumerated 
and for a myriad of other jobs for skilled workers, profes­
sional or nonprofessional, white collar or blue.

Lerner, Employment Discrimination: Adverse Impact, Validity, 
and Equality, 1979 Sup. Ct. Rev. 17, 18-19 (footnotes omitted) 
(emphasis added).

343



13

applicable job qualifications” that the employer selects. 
401 U.S. at 434 (quoting 110 Cong. Rec. 7247 (1964) (mem­
orandum of Sen. Case and Sen. Clark)). Moreover, this 
Court subsequently stated that:

Title VII could not have been enacted into law with­
out substantial support from legislators in both Houses 
who traditionally resisted federal regulation of pri­
vate business. Those legislators demanded as a price • 
for their support that ‘‘management prerogatives . . . 
be left undisturbed to the greatest extent possible.” 
H.R. Rep. No. 914, 88th Cong., 1st Sess., pt. 2, p. 29 
(1963).

United Steelworkers of America v. Weber, 443 TJ.S. 193, 
206 (1979).

Thus, while Title VII was intended ‘‘to make it an 
illegal practice to use race as a factor in denying employ­
ment,” it was not intended to encroach on the employer’s 
right to manage its work force. See 110 Cong. Reo. 13,076- 
79 (1964) (discussion between Sen. Ervin and Sen. 
Cooper); 110 Cong. Rec. 13,080 (1964) (remarks of Sen. 
Humphrey).

2. The traditional adverse impact test outlined in 
Griggs simply does not reach purely subjective business 
judgments. However, there is no analytical proscription 
against using Griggs’ statistical proof methods to chal­
lenge such judgments as discriminatory. After all, the 
neutral principles that drive Title V II’s factual inquiry 
permit the use of any evidence from which a court reason­
ably may infer illegal discrimination ‘‘under the circum­
stances” of the case.

In the usual instance, subjective personnel judgments 
are particularly amenable to a disparate treatment analy-

344



14

sis. See, e.q., Furnco Construction Corp. v. Wafers, supra. 
In some cases, however, it may he possible to make out 
an adverse impact, claim, particularly where multiple plain­
tiffs challenge a specific subjective decisionmaking prac­
tice. In such cases, it is natural to seek initial guidance 
from Griqqs and adopt its prima facie standard as a model 
for evaluating the plaintiffs’ statistical evidence.

3. This is precisely what this Court did in Watson. 
when seven Justices agreed that a plai?itiff who seeks to 
challenge an employer’s subjective business judgment in 
making personnel decisions must do more than merely 
show that there are “ statistical disparities in the employ­
er’s work force.” 108 S.Ct. at 2788 (plurality opinion); 
id. at 2792 & n.2 (Blackmun, J., concurring). Instead, the 
plaintiff must:

“ isolat[e] and identify[]” each “ specific employ­
ment practice” that allegedly is “ responsible for any 
observed statistical disparities” in work force com­
position; and “ offer statistical evidence of a kind and 
degree sufficient to Bhow that the practice in question 
has caused” the alleged harm “ because of fthe plain­
tiff’s] membership in a protected group.”

Id. at 2788-89 (plurality opinion).

Of course, the quantum of proof at this stnge is not 
absolute. But as with all circumstantial evidence based on 
statistics, the “ statistical disparities must be sufficiently 
substantial that they raise an inference of causation.” 108
S.Ct. at 2789. In short, the disparity must suggest a 
“ statistical significance,” such that it is unlikely to have

345



15

occurred merely by chance and, therefore, if unexplained, 
may reasonably imply discrimination.12

In rebuttal, the employer must “ produce evidence that 
its employment practices are based on legitimate business 
reasons.” Watson, 108 S.Ct. at 2790.13 Usually this should 
only require the employer to identify a facially plausible 
business reason for its judgment. Cf. Dothard v. Rawlitt- 
son, 433 U.S. at 340 (Rehnquist, J., concurring) (the em­
ployer in an impact case must “ articulate the asserted job- 
related reasons underlying the [practice]” ). See note 11 
supra and accompanying text. It then falls to the plain­
tiff to prove a Title VII violation by showing that there 
exist other, less discriminatory decisionmaking practices 
that fulfill the employer’s “ business goals” equally as 
well, and at no greater cost or burden than the challenged 
practice. See Watson, 108 S.Ct. at 2790.

4. These carefully prescribed factors properly strike 
the balance mid-way along the continuum between what

12 See, e.g., Griggs, 401 U.S. at 426 (the employment test had 
to "operate to disqualify Negroes at a substantially higher rate 
than white applicants") (emphasis added); Albemarle, 422 U.S. 
at 425 (plaintiffs were required to show "that the tests in ques­
tion select[edl applicants . . .  in a racial pattern significantly 
different from that of the pool of applicants") (emphasis add­
ed); Dothard, 433 U.S. at 329 ("plaintiff need only show" that 
fixed standards "selected] applicants for hire in a significantly 
discriminatory pattern") (emphasis added); Teal, 457 U.S. at 
446 ("significantly discriminatory impact") (emphasis added).
13 Of course, before proceeding with this evidentiary stage, 
the employer may challenge the statistical premise of the prima 
facie case, and undermine any statistical inferences of causation. 
See Dothard v. Rawlinson, 433 U.S. 321, 331 (1977); Hazelwood 
School District v. United States, 433 U.S. 299, 309-312 (1977).

346



1G

traditionally has been called an adverse impact case (based 
on standardized tests or other fixed criteria) and a dis­
parate treatment case (based on impermissible subjective 
criteria). In this middle ground, where amorphous quali­
ties of subjective judgment, come into play, it is especially 
important for this Court to guide both lower courts and 
litigants in the legal standards necessary to apply a Wat­
son-type analysis. For these reasons, we believe that the 
Court should adopt as its holding the plurality’s opinion in 
Watson and apply that analysis to the present case.

0. The Ninth Circuit’s Decision Reversing the 
Trial Court’s Judgment for Petitioners Con­
flicts with the Plurality’s Opinion in Watson.

Turning to the current case, respondents attempted at 
trial to show that one or more of about sixteen challenged 
employment practices, either separately or together, vio­
lated Title VII.14 After multiple appeals, the Ninth Cir­
cuit selected several practices as potential subjects for an 
adverse impact challenge: word-of-mouth recruitment,
nepotism, separate hiring channels, housing messing and 
race labeling. Atonio v. Wards Cove Pacldnq Co., 827 
F.2d 439, 444-49 (9th Cir. 1987). Such broad, nebulous 
claims simply should not he allowed. In order to foster 
meritorious litigation and present an orderly case for trial, 
challenges to subjective decisionmaking must focus on the 
causative effects of “ isolated” decisionmaking practices. 
Otherwise, “ the only cost-effective means of avoiding ex­
pensive litigation” will he for employers to adopt illegal

Petition for Certiorari at 5-6 & n.3.

347



17

and pernicious “ quotes and preferential treatment [poli­
cies].” Watson, 108 S.Ct. at 2788.

1. Statistics. Respondents are “ unskilled” cannery 
workers. The Ninth Circuit found that they produced 
prima facie evidence of discrimination simply by tender­
ing numerical data of segregation in the workplace, vis-a- 
vis “ skilled” positions. Atonio, 827 F.2d at 444. The 
Court of Appeals relieved the respondents of any burden 
to prove that there actually existed minority individuals 
qualified for the skilled positions they challenged. In­
stead, it held:

The statistics show only racial stratification by job 
category. This is sufficient to raise an inference that 
some practice or combination of practices has caused 
the distribution by race . . . .

827 F.2d at 444 (emphasis added).

The Ninth Circuit’s holding does not comport with 
the corresponding test under Watson. First, evidence of 
“ mere disparities in the employer’s work force” will not 
establish a prima facie case. Watson, 108 S.Ct. at 2788; 
id. at 2797 & n.2 (Blackmun, J., concurring). Second, re 
spondents’ failure to “ isolate” and “ identify” the par­
ticular decisionmaking practice that caused the disparity 
is fatal to their case. Id. at 2788. They simply cannot 
allege a claim of adverse impact until they first identify a 
specific decisionmaking practice that, causes discrimina­
tion. Finally, the Court of Appeals improperly relieved 
respondents of the obligation to show that minority in­
dividuals actually were qualified for the skilled jobs at 
issue.

348



18

2. Specific Practices. The Court of Appeals also 
reviewed petitioners’ other claims of discrimination. How­
ever, respondents’ failure to isolate specific objectionable 
decisionmaking practices, or produce statistical evidence 
sufficiently probative of causation based on race, renders 
their entire claim insufficient. At minimum, this Court 
should vacate the Court of Appeals’ decision and remand 
for further findings in accordance with the plurality’s 
analysis in Watson.

I l l  so doing, this Court should stress the need for re­
spondents to make a substantial statistical showing of dis­
crimination ns to each challenged employment, practice. 
Moreover, petitioners have articulated legitimate business 
reasons for separate hiring channels (union versus non­
union hiring),15 informal recruitment (personal knowl­
edge and hiring of skilled workers by application only),16 
separate messing facilities (culinary preference and union 
restrictions),17 and separate housing facilities (seasonal re­
quirements and workshift harmony).18 Consequently, if 
respondents do establish a prima facie case, this Court 
should stress that to prevail they also must identify spe­
cific alternative practices that (1) fulfill the same business 
functions as the challenged practices, but that are neither 
(2) more costly or troublesome for the employer to imple-

15 Atonio, 34 Empl. Prac. Dec. (CCH) 33,821, 33,827-28 (W.D. 
Wash. 1983) (findings 85-90, 94, 101-103).
16 Id. at 33,827-28 (findings 87-89, 94); id. at 33,830 (find­
ings 124-128).
17 Id. at 33,836 (findings 143-147); id. at 33,844 (applying 
adverse impact test).
18 Id. at 33,836 (findings 149A-149C); id at 33,844 (applying 
adverse impact test).

349



19

mont, nor (3) needlessly intrusive of workplace manage­
ment.

3. Nepotism . Finally, we raise a special concern 
about nepotism. Single acts of nepotism are unlikely to be 
illegal. Cf. DeCintio v. W estchester County Medical Cen­
ter , 807 F.2d 304 (2d Cir. 1986), cert, denied, —  U.S. —, 
108 S.Ct. 89 (1987) (rejecting Title V II claim where 
woman nurse was hired by her paramour, despite the 
presence of other qualified applicants). However, a nepo­
tism policy or practice may be discriminatory where it is 
sufficiently pervasive. In such cases, the Ninth Circuit 
properly was concerned that if members of a predominant 
racial group hire only their own relatives, then ‘‘the prac­
tice necessarily has an adverse impact.” Atonio, 827 F.2d 
at 445.

But in the current case, the trial court’s finding of 
no discrimination should be sustained. The trial court 
found that ‘‘the [respondent’s] nepotism figures failed to 
differentiate those persons who became related through 
marriage a fter  starting work at the canneries.” Atonio, 
34 Empl. Prac. Dec. (OCII) 33,821, 33,840 (W.D. Wash.
1983) (emphasis in original). Moreover, the court found 
that ‘‘ [rjelatives of whites and particularly  nonwhites 
appear in high incidence at the canneries.”  Id. (empha­
sis added). Given these findings, it is difficult to see how 
respondents could possibly prove adverse impact because 
of race.

In sum, the evidence submitted by respondents at trial 
was insufficient to prove their adverse impact claim under 
any of this Court’s prior holdings, and particularly under

350



20

the standards applicable to subjective decisionmaking that 
were articulated by the W atson  plurality.

o

CONCLUSION

For the reasons expressed above, we believe that this 
Court should (1) reaffirm the plurality’s suggestion in 
W atson  that there exists a single analytical approach to 
deciding Title VIT cases; (2) adopt the plurality’s opinion 
as the proper mode of applying statistical evidence to sub­
jective decisionmaking practices; and (3) vacate the de­
cision below.

Respectfully submitted,

ARTHUR H. ABEL 
FAEGRE & BENSON 
2200 Norwest Center 
90 South Seventh Street 
Minneapolis, MN 55402 
(612) 336-3000

* Counsel of Record

CLINT BOLICK*
JERALD L. HILL 
MARK J. BREDEMEIER 
LANDMARK LEGAL FOUNDATION 
CENTER FOR CIVIL RIGHTS 
107 Second Street, N.E.
Washington, D.C. 20002 
(202) 546-6045

351





No. 87-1387

In The

g ’UftmttP Court of %  Ittttrft f&atra
October Term, 1988

Wards Cove Packing Company, Inc., 
Castle & Cooke, Inc.,

Petitioners,
v.

Frank Atonio, et al,
Respondents.

On Writ of Certiorari to the United States Court of Appeals
for the Ninth Circuit

BRIEF FOR THE CHAMBER OF COMMERCE 
OF THE UNITED STATES OF AMERICA 

AS AMICUS CURIAE SUPPORTING PETITIONERS

Of Counsel:
Stephen A. Bokat 
Mona C. Zeiberg 
National Chamber 

Litigation Center, Inc. 
1615 H Street, N.W. 
Washington, D.C. 20062 
(202) 463-5337

Glen D. Nager
(Counsel of Record)

Andrew M. Kramer 
David A. Copus 
Patricia A. Dunn  
Jones, Day, Reavis & Pogue 
1450 G Street, N.W. 
Washington, D.C. 20005-5701 
(202) 879-3939 
Attorneys for the

Chamber of Commerce of 
the United States of America

353



QUESTIONS PRESENTED

1. Whether a plaintiff-class may state a cause of ac­
tion under Title VII of the Civil Rights Act of 1964, 42 
U.S.C. (& Supp. Il l )  §§ 2000e et scq., based on the cum­
ulative effect of a wide range of non-racially motivated 
employment practices.

2. Whether proof that non-white persons are more 
heavily represented in one level of an employer’s work 
force than in another level of that work force establishes 
as a matter of law that the employer’s selection and em­
ployment practices have had a disparate impact on non­
white persons.

3. Whether the court below improperly shifted the 
burden of proof and/or applied an incorrect standard of 
proof in holding that the selection and employment prac­
tices challenged in this case were not sufficiently justified 
so as to rebut any prima facie case of disparate impact 
made against them.

( i )

354



INTEREST OF AMICUS................................................ 1

STATEMENT .................................................................... 2

SUMMARY OF ARGUMENT........................................ 6

ARGUMENT ......................................................................  9

I. A TITLE VII PLAINTIFF-CLASS MAY NOT
CHALLENGE THE CUMULATIVE EFFECT 
OF A WIDE RANGE OF SELECTION AND 
EMPLOYMENT PRACTICES UNDER THE 
DISPARATE IMPACT ANALYSIS .................  9
A. This Court Has Approved The Application

Of Disparate Impact Theory Only In Cases 
Where A Specific Employment Practice Is 
Itself Shown To Cause A Significantly Dis­
parate Exclusion Of Individuals In A Pro­
tected Group.....................................................  19

B. Extending The Disparate Impact Analysis
To Challenges To The Cumulative Effect Of 
Multiple Employment Practices Would Pro­
duce Results That Are At Odds With The 
Balance Struck By Congress In Title VII 12

C. The Decisions Of The Courts Of Appeals 
That Have Extended The Disparate Impact 
Analysis To Challenges To The Cumulative 
Effect Of Multiple Employment Practices
Are Based On Improper Concerns..................  16

II. INTERNAL WORK FORCE STATISTICS
CANNOT DEMONSTRATE THAT MINORI­
TIES HAVE BEEN DISPROPORTIONATELY 
EXCLUDED FROM JOBS UNLESS THE EM­
PLOYER HAS A POLICY OF PROMOTING 
FROM WITHIN....................................................  19

TA BL E OF CO NTENTS
Page

TA BL E OF A U T H O R IT IE S ................................................... v

(iii)

355



iv

III. a t  t h e  r e b u t t a l  sta g e  o f  a  d is pa ­
r a t e  IMPACT CASE, AN EMPLOYER NEED 
ONLY SHOW THAT ITS SELECTION DE­
VICES ARE REASONABLE IN LIGHT OF 
THE JOB AT ISSUE AND THE NATURE OF 
THE BUSINESS............................................... 24

CONCLUSION ........................... or

T A B L E  O F C O N T E N T S — C ontinued
Page

356



V

TABLE OF AUTHORITIES
Cases Page

AFSCME v. Washington, 770 F.2d 1401 (9th Cir.
1985) ........................................................................  12

Albemarle Paper Co. v. Moody, 422 U.S. 405
(1975) ............................................................ 12, 15, 25, 27

Anderson v. Bessemer City, 470 U.S. 564 (1985) .... 23
Bauer v. Bailar, 647 F.2d 1037 (10th Cir. 1981).....  17
Clark v. Chrysler Corp., 673 F.2d 921 (7th Cir.),

cert, denied, 459 U.S. 873 (1982)......................... 23
Connecticut v. Teal, 457 U.S. 440 (1982)................12,14,17
Coser v. Moore, 739 F.2d 746 (2d Cir. 1984)..........  23
Dothard v. Rawlinson, 433 U.S. 321 (1977)............  12,20
EEOC v. Federal Reserve Bank, 698 F.2d 633 (4th 

Cir. 1983), rev’d, sub nom. Cooper v. Federal Re­
serve Bank, 467 U.S. 867 (1984)............................  22

Espinoza v. Farali Mfg. Co., 414 U.S. 86 (1973).....  19
Furnco Constr. Corp. v. Waters, 438 U.S. 567

(1978)........................................................... ...........  24,27
General Elec. Co. v. Gilbert, 429 U.S. 125 (1976) .... 19
Green v. USX Corp., 843 F.2d 1511 (3d Cir. 1988), 

petition for cert, filed, 57 U.S.L.W. 3123 (U.S.
July 23, 1988) (No. 88-141) .................................  16, 18

Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1985).... 16,18
Griggs v. Duke Power Co., 401 U.S. 424 (1971).....passim
Harbison-Walker Refractories v. Brieck, No. 87-

271 (U.S. cert, granted, March 21, 1988)............ 2
Hazelwood School Dist. v. United States, 433 U.S.

299 (1977) ...........................................................   18,20
Hilton v. Wyman-Gordon Co., 624 F.2d 379 (1st

Cir. 1980) ...............................................................  23
International Bhd. of Teamsters v. United, States,

431 U.S. 324 (1977).....................  18,20,25
Johnson v. Transportation Agency, 107 S. Ct. 1442

(1987)  11,14,15,23
Johnson v. Uncle Ben’s, Inc., 628 F.2d 419 (5th Cir.

1980), vacated, 451 U.S. 902 (1981).....................  21
Local 28, Sheet Metal Workers’ Int’l Ass’n v.

EEOC, 478 U.S. 421 (1986) ...........................   15
Los Angeles, Dep’t of Water & Power v. Manhart,

435 U.S. 702 (1978) .............................................  23

357



Mohasco Corp. v. Silver, 447 U.S. 807 (1980)......  19
NAACP v. Medical Center, Inc., 657 F.2d 1322 (3d

Cir. 1981)........................................................... 25
New York City Transit Auth. v. Beazer, 440 U.S.

568 (1979) .........................................................passim
Bouncy v. Prudential Ins. Co., 668 F.2d 795 (5th

Cir. 1982) .......... ...............................................  13,18
Rivera v. City of Wichita Falls, 665 F.2d 531 (5th

Cir. 1982)..........................................................  22
Robinson v. Polaroid Corp., 732 F.2d 1010 (1st Cir.

1984)   13
Segar v. Smith, 738 F.2d 1249 (D.C. Cir. 1984), 

cert, denied sub nom. Meese v. Segar, 471 U.S.
1115 (1985).......................................................  16

Ste. Marie v. Eastern R.R. Ass’n, 650 F.2d 395 (2d
Cir. 1981) ..........................................................  21-22

Stewart v. General Motors Corp., 542 F.2d 445 
(7th Cir. 1976), cert, denied, 433 U.S. 919
(1977)   17

Texas Dep’t of Community Affairs v. Burdine, 450
U.S. 248 (1981) ................... 11,17,25

U.S. Postal Serv. Bd. of Governors v. Aikens, 460
U.S. 711 (1983) ................................................. 17

United Steelworkers v. Weber, 443 U.S. 193
(1979) ............................................................ 11,14,15

Washington v. Davis, 426 U.S. 229 (1976) ............  12
Watson v. Fort Worth Bank & Trust, 108 S. Ct.

2777 (1988)........................................................passim
Statutes and Regulations

42 U.S.C. § 1981................................................  3
Title VII of the Civil Rights Act of 1964, as 

amended, ("Title VII”), 42 U.S.C. §§ 2000e et 
seq.......................................................................  3

Title VII § 703 (a), 42 U.S.C. § 2000e-2(a)..... 6,10
Title VII § 703(a) (2), 42 U.S.C. § 2000e-2(a)

(2) ..............................................................  7,11
Title VII § 703(j ), 42 U.S.C. § 2000e-2(j)...... passim

vi

T A B L E  O F A U T H O R IT IE S — C ontinued
P a g e

358



vii

Pa Re
Uniform Guidelines on Employee Selection Proce­

dures (1978), 29 C.F.R. § 1607 ........................... 19
29C.F.R. § 1607.16Q.......................................... 19
29 C.F.R. § 16Q7.3A............................................ 19

Miscellaneous
B. Schlei and P. Grossman, Employment Discrimi­

nation Law (1983) .................................................  17
Baldus and Cole, Statistical Proof of Discrimina­

tion § 4.11 ...............................................................  20
Campbell, Regression Analysis in Title VII Cases: 

Minimum Standards, Comparable Worth, and 
Other Issues Where Law and Statistics Meet, 36
Stan. L. Rev. 1299 (1984)...................................... 16,18

H.R. Rep. No. 914, 88th Cong., 1st Sess., pt. 2
(1963) ...................................................................... 15

Lerner, E7nployment Discrimination: Adverse Im­
pact, Validity and Equality, 1979 Sup. Ct. Rev.
17 .............................................................................  26

Maltz, Title VII and Upper Level Employment—A 
Response To Professor Bartholet, 77 Nw. U.L.
Rev. 776 (1983) .....................................................  14

TABLE OF AUTHORITIES— Continued

359



In  T he

!$uprr«tP (Em td uf tljp luitpii l̂ tatPB
October T erm , 1988

No. 87-1387

W ards Cove Packing Com pany , Inc ., 
Castle & Cooke, Inc .,

Petitioners,
v.

F rank  A tonio, et a l,
Respondents.

On Writ of Certiorari to the United States Court of Appeals
for the Ninth Circuit

BRIEF FOR THE CHAMBER OF COMMERCE 
OF THE UNITED STATES OF AMERICA 

AS AMICUS CURIAE SUPPORTING PETITIONERS

INTEREST OF THE AMICUS *
The Chamber of Commerce of the United States of 

America, a nonprofit corporation organized and existing 
under the laws of the District of Columbia, is the largest 
federation of business, trade, and professional organiza-

* Counsel for both parties have consented to the filing of this 
amicus brief. Their consents have been filed with the Clerk of 
this Court.

360



2

tions in the United States. It represents the interests of 
over 180,000 corporations, partnerships, and proprietor­
ships, as well as state and local chambers of commerce 
and trade associations. Many of the Chambers’ members 
use multi-component selection and decision-making proc­
esses. Thus, the resolution of the questions presented in 
this case—involving whether disparate impact theory ap­
plies to challenges to the cumulative effect of multiple 
selection and employment practices; whether a disparity 
in the percentages of minorities employed in different job 
categories is a sufficient basis for establishing a prima 
facie disparate impact case; and whether and to what 
extent an employer must prove that a racial workforce 
disparity is justified by business necessity—is of signifi­
cant interest to the Chamber and its members. In sim­
ilar circumstances, the Chamber has filed amicus briefs 
with this Court. See, e.g., Harbison-Walker Refractories 
v. Brieck, No. 87-271 (U.S. cert, granted March 21, 
1988); Griggs v. Duke Power Co., 401 U.S. 424 (1971).

STATEMENT

1. Petitioners operate five salmon canneries in remote 
and widely-separated areas of Alaska. Pet. App. 111:2-3. 
Petitioners begin operations each year in May or June, a 
few weeks before the anticipated salmon runs, with a 
period known as the “preseason.” Id. at 111:4-5. During 
this preseason, petitioners bring in employees to assemble 
equipment, repair any winter damage to the facilities, 
and prepare the canneries for the onset of the canning 
season. Id. at III:5. The individuals who staff the can­
ning lines during the season—the “cannery” workers— 
arrive toward the end of the pre-season. Ibid. The can­
nery workers remain as long as the salmon runs last, and 
depart when the canning is completed. Id. at 111:5-6. 
The canneries lie vacant for the rest of the year. Id. at 
III:3.

361



3
Most of the jobs in the canneries are seasonal and 

petitioners must reconstitute their work forces each year. 
Pet. App. Ill :8. Petitioners hire the cannery workers, 
who are the lowest paid members of the summer work­
force, principally from native villages in Alaska and from 
the dispatcher of a primarily Filipino union local in 
Seattle, Washington. Id. at III: 11. Petitioners hire the 
more highly-paid “non-cannery” workers—e.g., machin­
ists and engineers who maintain the canning equipment; 
quality control personnel who conduct government- 
required inspections and recordkeeping; boat crews that 
operate transport equipment; and a variety of support 
personnel—from a multi-state region encompassing 
Alaska, the Pacific Northwest, and California. Id. at 
1:36, III:7. Petitioners select the non-cannery employees 
from among off-season applicants, word-of-mouth recruits, 
and “rehires” who worked at the canneries during prior 
seasons. Id. at 111:11. They transport nearly all of these 
employees to and from the canneries each year, and house 
and feed them while they are there. Id. at III :8.

2. Respondents, a class of non-white employees at the 
canneries, brought this action under Title VII of the 
Civil Rights Act of 1964, 42 U.S.C. §§ 2000e et seq., and 
42 U.S.C. § 1981, claiming that petitioners had discrim­
inated against them because of their race. Pet. App. 
Ill :2, 9. Specifically, respondents alleged that petitioners 
had intentionally discriminated against them by using 
certain employment and selection practices, including sep­
arate hiring channels, word-of-mouth recruiting, nepo­
tism, rehire preferences, language skill requirements, sub­
jective job qualifications, and segregated housing and 
messing facilities. Id. at 111:9-12. Respondents further 
alleged that these practices had an unlawful disparate 
impact on their opportunity to obtain the higher-paying 
non-cannery worker jobs. Id. at III :9.

During a lengthy non-jury trial, respondents supported 
their claims by showing that approximately 48 percent

362



4

of the individuals employed in the Alaska salmon canning 
industry since 1970 were non-white and that these non­
white persons had principally been employed as cannery 
workers. Pet. App. 1:35-36, 42. Respondents also showed 
that petitioners had not posted job vacancies in non­
cannery positions or promoted cannery workers to non­
cannery positions {id. at 1:28-29, 33-34, 39) ; and that 
petitioners had frequently hired relatives of existing white 
employees {id. at 1:104-05). Finally, respondents ar­
gued that petitioners had followed race-labeling practices 
and maintained racially-segregated housing and messing 
facilities {id. at 1:76-84).

In rebuttal, petitioners demonstrated that, while cen­
sus data indicated that the potential applicant pool for 
petitioners’ facilities was only 10 percent non-white (with­
out regard to place or position of current employment, 
skills, or pre-season availability), non-whites had been 
employed in 21 percent of the non-cannery positions. Pet. 
App. 1:35-37; Pet. 4. Petitioners explained that the can­
ning industry attracted applicants—for cannery and 
non-cannery positions—from a multi-state region, prin­
cipally because of the high wages that were guaranteed 
to workers. Pet. App. 1:41. Petitioners further explained 
that non-whites were more heavily represented in cannery 
worker positions than in either non-cannery jobs or the 
potential applicant pool, both because non-whites were 
concentrated in the communities surrounding the can­
neries and in the union from which petitioners obtained 
many of their cannery workers, and because the short 
and intense canning season generally precluded mid­
season training and promoting of cannery workers and 
required resort to the external labor market. Id. at 
1:18-19, 32, 36-38, 41-43, 45-46. Finally, petitioners 
showed that their housing and messing practices were 
structured to accommodate workers’ preferences, the 
workers’ arrival times and departmental assignments, the 
cost of providing such benefits, and the demands of the

363



5

employees’ collective bargaining representatives. Id. at 
1:81-84, 126-29.

The district court entered judgment in favor of peti­
tioners. Pet. App. 1:1-130. It held that petitioners’ sub­
jective decision-making criteria could not be challenged 
under the disparate impact theory. Id. at 1:102. It also 
held that respondents had failed to prove that petitioners’ 
language skill requirements and alleged nepotism policy 
had an unlawful disparate impact on non-whites. Id. at 
1:102-105. It then determined that, viewing all of the 
practices together, respondents had failed to establish 
disparate treatment. Id. at 1:106, 119. The court found, 
inter alia, that any employee could apply for any job 
at the canneries (id. at 1:33); that respondents’ statis­
tics were not probative of discrimination in the non­
cannery jobs requiring skills, experience, or availability 
(id. at 1:113-14); that the over-representation of non­
whites in the cannery positions was attributable to non- 
discriminatory factors, i.e., the undue concentration of 
non-whites in the local communities and in the referrals 
from the union dispatcher (id. at 1:109-11); and that, 
while respondents’ evidence as a whole “raised a mar­
ginal inference of discriminatory treatment” (id. at 
1:119), petitioners had successfully rebutted that infer­
ence with relevant statistics and other evidence showing 
that their practices were motivated by legitimate busi­
ness considerations. Id. at 1:35-43, 110-14, 119-22, 124- 
29.

3. A panel of the Ninth Circuit affirmed the judgment 
of the district court. Pet. App. 111:1-56. The en banc 
court subsequently vacated that judgment, however, and 
held that petitioners’ subjective employment practices 
could be challenged under the disparate impact theory. 
Id. at V:l-75. On remand from the en banc court, the 
panel then vacated the judgment of the district court and 
remanded for further proceedings. Id. at VI: 1-44.

364



6

The panel did not disturb the district court’s conclusion 
that intentional race discrimination had not been estab­
lished. Pet. App. VI: 16. But the panel found that a 
prima facie case of disparate impact against non-whites 
had been demonstrated. Pet. App. VI: 13-19. The panel 
noted that respondents had both introduced statistics 
showing “racial stratification by job category” and “iden­
tified certain practices which cause fd | that impact.” Id. 
at VI:18, 19. The panel found that, in combination, such 
evidence was “sufficient to raise an inference that some 
practice or combination of practices has caused the dis­
tribution of employees by race . . . .” Id. at VI: 18.

Having so held, the panel turned to the particular 
practices at issue to determine whether each was “linked 
causally with the demonstrated adverse impact” and, if 
so, whether it was justified by business necessity. Pet. 
App. VI: 19-39. The panel found that each practice had 
an “obvious” or “necessar[]y” or “clear” link to the 
racial disparity in the work force. Id. at VI:21, 28, 36. 
The panel then either rejected the justifications that pe­
titioners had offered for their practices—with the excep­
tion of the language skills and rehire policies—or re­
manded for further development of the facts supporting 
those justifications. See id. at VI:21, 25-27, 28, 30-32, 
37-39.

SUMMARY OF ARGUMENT

A. Congress carefully accommodated competing objec­
tives when it enacted Title VII in 1964. It sought in 
§ 703(a) of the statute to achieve equality of employment 
opportunities by removing arbitrary and unjustified bar­
riers to the employment of members of minority groups. 
But, as § 703(j) of the statute makes clear, it did so in­
tending not to disturb traditional management preroga­
tives or to require employers to engage in preferential 
treatment of minorities or work force balancing. Recog­
nizing this accommodation, this Court, in Griggs v. Duke 
Power Co., 401 U.S. 424 (1971), held that a violation of

365



7

§ 703(a) (2) may be established without a finding of il­
licit motivation where an employment practice dispropor­
tionately excludes individuals in a protected group and is 
not justified by legitimate business needs. In its subse­
quent decisions, the Court has approved the disparate 
impact theory only where these two limiting conditions 
have been met—i.e., where the plaintiff has established 
that a particular employment practice is itself the cause 
of a significant racial disparity and where the practice 
is not justified by business necessity.

Extending the disparate impact analysis to challenges 
to the cumulative effect of multiple employment practices 
would force employers seeking to avoid Title VII liabili­
ties to take actions that are at odds with the balance 
struck by Congress in Title VII and recognized by the 
Court in Griggs and its progeny. To do so would force 
employers, at the rebuttal stage of a disparate impact 
case, either to identify the practice, if any, that caused 
the disparity and demonstrate that that practice is justi­
fied by business necessity or to show that each component 
of the selection process, regardless of its individual im­
pact, is so justified. But this shifting of evidentiary bur­
dens would be inconsistent with this Court’s statements 
that the plaintiff, not the employer, bears the burden of 
producing evidence that the challenged practice has 
caused the alleged statistical disparity and that proof of 
a mere work force imbalance will not suffice. Alterna­
tively, of course, employers could abandon or modify their 
multiple selection and employment practices in an effort 
to avoid such challenges. But forcing employers to re­
structure their business practices would be inconsistent 
with Congress’s intent that Title VII not be interpreted 
to allow undue governmental intervention into private 
business decisions. Finally, employers could superimpose 
numerical quotas on their selection and employment proc­
esses to ensure that they achieve a racially-balanced 
work force. But, again, this result would be inconsistent

366



8

with Congress’ intention that employers not be required 
to use quotas to avoid Title VII liabilities.

The concerns expressed by some courts of appeals— (a) 
that plaintiffs cannot identify and prove the effects as­
sociated with the various selection and employment 
practices used by an employer, and (b) that several 
components which individually have no adverse impact 
may “interact” to cause a racially-imbalanced work 
force—do not justify the extension of disparate impact 
theory to the cumulative effect of multiple employment 
practices. Plaintiffs can use multiple regression analyses 
—i.e., statistical analyses that produce estimates of 
weights for each variable in a multi-factor process, thus 
indicating the effect that each variable has on an outcome 
—to identify and isolate the causes of racial work force 
disparities; plaintiffs can obtain information about an 
employer’s selection and employment practices through 
the liberal rules of discovery and access to the Equal 
Employment Opportunity Commission’s (“EEOC” ) in­
vestigatory files; and the fact that no single component 
of a multiple component process has an adverse effect on 
minorities establishes that any disparity associated with 
the overall process is a result of either lawful factors or 
disparate treatment, neither of which justifies applica­
tion of disparate impact analysis.

B. Respondents’ internal work force statistics are in­
sufficient as a matter of law to establish a prima facie 
case of disproportionate racial impact. While statistical 
evidence may take a variety of forms, it must, at a mini­
mum, establish a reasonable proxy for the relevant ap­
plicant pool so that, by comparison to the pool of em­
ployees actually hired, reasonable conclusions about rates 
of selection and rejection can be drawn. Respondents’ 
statistics—which focus on an internal work force imbal­
ance and the concentration of non-white persons in the 
canneries’ lowest paying jobs—do not establish such a 
proxy. Petitioners receive applications from persons both

367



9

within and without the work force, and respondents’ sta­
tistics thus measure only a subset of the potential appli­
cant pool. Reasonable conclusions about rates of selection 
and rejection cannot and should not be drawn from such 
obviously incomplete and under-inclusive data.

C. The court below misunderstood the nature of the 
rebuttal burden in a disparate impact case. By requiring 
petitioners to prove by a preponderance of the evidence 
that the challenged practices were justified by business 
necessity, the court below improperly relieved the plain­
tiff of its ultimate burden of persuasion in a Title VII 
case, equated a prima facie showing with a factual find­
ing of discrimination, and in effect held that a practice 
producing an adverse impact violates Title VII even 
though it may be justifiable. Moreover, in applying a 
standard of business necessity that requires employers to 
demonstrate more than that their practices are reason­
ably related to the requirements of their business, the 
court below erroneously rejected the substantial business 
justifications that petitioners proffered in defense of their 
selection and employment practices.

ARGUMENT
I. A TITLE VII PLAINTIFF-CLASS MAY NOT CHAL­

LENGE THE CUMULATIVE EFFECT OF A WIDE 
RANGE OF SELECTION AND EMPLOYMENT 
PRACTICES UNDER THE DISPARATE IMPACT 
ANALYSIS

The court below held that respondents had successfully 
established a prima facie case of race discrimination pro­
hibited by Title VII. The court did not question the dis­
trict court’s finding that respondents failed to demon­
strate intentional race discrimination. But the court con­
cluded that respondents had established a prima facie 
case of disparate impact with evidence (1) that petition­
ers’ work force is racially stratified and (2) that certain 
selection and employment practices are “obviously”, “nec­

368



10

essarily”, and “clearly” linked to that overall racial 
work force imbalance. This conclusion—i.e., that, without 
regard to the issue of motive or the significance of the 
disparity caused by any particular practice, plaintiffs in 
a Title VII suit may state a cause of action merely by 
identifying employment or selection practices that are col­
lectively linked to a racially-imbalanced work force— 
constitutes an unwarranted extension of the disparate 
impact theory and should be rejected by this Court.

A. This Court Has Approved The Application Of Dis­
parate Impact Theory Only In Cases Where A 
Specific Employment Practice Is Itself Shown To 
Cause A Significantly Disparate Exclusion Of Indi­
viduals In A Protected Group

This Court has said, and the language of § 703(a) of 
Title VII makes clear,1 that Congress’ basic objective in 
enacting Title VII was “to achieve equality of employ­
ment opportunities and [to] remove barriers that have 
operated in the past to favor an identifiable group of 
white employees over other employees.” Griggs v. Duke 
Power Co., 401 U.S. at 429-30. This Court has also rec­
ognized, however, that Congress had additional, compet­
ing objectives in mind when it enacted Title VII; spe­
cifically, the Court has recognized that, in § 703(j) of the * 1 2

1 Section 703(a) of the statute (42 U.S.C. § 2000e-2(a)) provides 
that:

It shall be an unlawful employment practice for an employer—
(1) to fail or refuse to hire or to discharge any individual, 

or otherwise to discriminate against any individual with respect 
to his compensation, terms, conditions, or privileges of employ­
ment, because of such individual’s race, color, religion, sex, or 
national origin; or

(2) to limit, segregate, or classify his employees or appli­
cants for employment in any way which would deprive or tend 
to deprive any individual of employment opportunities or other­
wise adversely affect his status as an employee, because of such 
individual’s race, color, religion, sex, or national origin.

369



11

statute,2 Congress expressed its concern that Title VII 
not be interpreted unduly to interfere with management 
discretion or to require employers to grant preferential 
treatment to minorities. See Johnson v. Transportation 
Agency, 107 S. Ct. 1442, 1450-51 n.7 (1987); Texas 
Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 
259 (1981); United Steelworkers v. Weber, 443 U.S. 193, 
206 (1979). It is against the background of these com­
peting statutory provisions and congressional objectives 
that the Court has shaped the contours of the disparate 
impact theory.

The Court first approved the use of disparate impact 
theory as a means of establishing unlawful employment 
discrimination in Griggs. At issue in Griggs were writ­
ten aptitude tests and a high school diploma requirement 
that the employer had adopted for the purpose of improv­
ing the general quality of its work force. Reversing a con­
trary holding of the court of appeals, this Court held that, 
in appropriate circumstances, a violation of § 703(a) (2) 
of the statute may be established without a finding of 
illicit motivation. 401 U.S. at 429-430. The Court ac­
knowledged that “the Act does not command that any 
person be hired simply because . . .  he is member of a 
minority group.” Id. at 430-31. But the Court concluded 
that no such preference is required, and, indeed, an un­
lawful preference for members of the majority group is 
eliminated, by prohibiting employment practices which 
disproportionately exclude individuals in a protected 
group and which have no “demonstrable relationship to 
successful performance of the jobs for which [they are] 
used.” Id. at 431. Because the high school diploma re­
quirement and written aptitude tests at issue each had 
its own significant exclusionary effect on blacks, and

2 Section 703(j) of the statute provides that “fnjothing contained 
in [Title VII] shall be interpreted to require any employer . . .  to 
grant preferential treatment to any individual or to any group 
because of the race . . .  of such individual or group on account of an 
imbalance” in the employer's work force. 42 U.S.C. § 2000e-2(j).

370



12

because neither selection criteria had a manifest relation­
ship to the requirements of the jobs for which each was 
used, the Court held that a violation of Title VII had 
been established. Id. at 430 n.6, 431-32, 436.

In its subsequent decisions, the Court has approved the 
disparate impact theory only where these two limiting 
conditions have been met—i.e., where a specific employ­
ment practice has itself caused a significantly dispropor­
tionate exclusion of individuals in a protected group and 
where that practice is not manifestly related to legitimate 
business needs. See, e.g., Connecticut v. Teal, 457 U.S. 
440 (1982) (written examination) ; New York City 
Transit Auth. v. Beazer, 440 U.S. 568 (1979) (prohibi­
tion on employment of methadone users) ; Dothard v. 
Rawlinson, 433 U.S. 321 (1977) (height and weight re­
quirements); Washington v. Davis, 426 U.S. 229 (1976) 
(written test); Albemarle Paper Co, v. Moody, 422 U.S. 
405 (1975) (employment test). Indeed, a plurality of the 
Justices stated just last Term that these two limitations 
are irreducible requirements for establishing a disparate 
impact violation. See Watson v. Fort Worth Bank & 
Trust, 108 S. Ct. 2777, 2788-91 (1988); see also id. at 
2792 and n.2 (concurring opinion); AFSCME v. Wash­
ington, 770 F.2d 1401, 1405 (9th Cir. 1985) (Kennedy, 
J.) (“Disparate impact analysis is confined to cases that 
challenge a specific, clearly delineated employment prac­
tice applied at a single point in the job selection 
process” ).

B. Extending The Disparate Impact Analysis To Chal­
lenges To The Cumulative Effect Of Multiple Em­
ployment Practices Would Produce Results That 
Are A t Odds With The Balance Struck By Congress 
In T itle VII

As at least a plurality in Watson and several court 
appeals have recognized, it would be improper to extend 
the disparate impact analysis to challenges to the cumula­
tive effect of multiple employment practices. See Watson

371



15

(1983). These “legislators demanded as a price for their 
support that ‘management prerogatives, and union free­
doms . . .  be left undisturbed to the greatest extent pos­
sible.* ’’ United Steelworkers v. Weber, 443 U.S. at 206, 
quoting H.R. Rep. No. 914, 88th Cong., 1st Sess., pt. 2, 
p. 29 (1963). Thus, even assuming that an employer 
could simplify its selection processes or separate its job 
tasks, which often would not be the case, requiring the 
employer to do so would represent precisely the type of 
federal intervention into private business that the key 
legislators would not accept.

For similar reasons, an employer plainly cannot be put 
in the position of having to adopt surreptitious quota 
systems in order to avoid Title VII liabilities. This Court 
has, of course, held that Title VII permits employers to 
engage in limited forms of voluntary affirmative action. 
See Johnson v. Transportation Agency, 107 S. Ct. at 
1450-51. But, as noted above, the Court has also recog­
nized that § 703(j) was added to Title VII to ensure 
that the statute would not be interpreted to “require em­
ployers or labor unions to use racial quotas or to grant 
preferential treatment to racial minorities in order to 
avoid being charged with unlawful discrimination.” Local 
28, Sheet Metal Workers Int'l Ass’n v. EEOC, 478 U.S. 
421, 453 (1986). The congressional record is replete with 
comments “that employers would not be required to in­
stitute preferential quotas to avoid Title VII liability.” 
United Steelworkers v. Weber, 443 U.S. at 207 n.7. A 
rule of law that “leave[s] the employer little choice . . . 
but to engage in a subjective quota system of employment 
selection” would thus be “far from the intent of Title 
VII.” Albemarle Paper Co. v. Moody, 422 U.S. at 449 
(Blackmun, J., concurring).

374



16

C. The Decisions Of The Courts Of Appeals That Have 
Extended The Disparate Impact Analysis To Chal­
lenges To The Cumulative Effect Of Multiple 
Employment Practices Are Eased On Improper 
Concerns

The courts of appeals that have permitted plaintiffs to 
challenge the cumulative effect of a wide range of em­
ployment practices under the disparate impact theory 
have been concerned (a) that plaintiffs do not have suffi­
cient ability or information to isolate the particular prac­
tice, if any, that has actually caused a work force imbal­
ance, and (b) that imbalances attributable to the inter­
action of several practices will escape judicial scrutiny if 
such challenges are not allowed. See Green v. USX Corp., 
843 F.2d 1511, 1522-25 (3d Cir. 1988), petition for cert, 
filed, 57 U.S.L.W. 3123 (U.S. July 23, 1988) (No. 88- 
141); Griffin v. Carlin, 755 F.2d 1516, 1525 (11th Cir.
1985) ; Segar v. Smith, 738 F.2d 1249, 1271-1272 (D.C. 
Cir. 1984), cert, denied sub nom. Meese v. Segar, 471 
U.S. 1115 (1985). Neither concern justifies the legal rule 
that these courts have applied.

The concern that plaintiffs cannot isolate the particular 
practice or practices, if any, actually causing a work force 
imbalance slights both the tools available to plaintiffs in 
Title VII cases and the burden of proof that rests with 
plaintiffs. Plaintiffs in Title VII cases can employ mul­
tiple regression analyses—i.e., statistical analyses that 
produce estimates of weights for each variable in a multi­
factor process, thus indicating the effect that each vari­
able has on an outcome—to identify and isolate the effects 
attributable to the various employment practices used by 
an employer. See Campbell, Regression Analysis in Title 
VII Cases: Minimum Standards, Comparable Worth, and 
Other Issues Where Lain and Statistics Meet, 36 Stan. L. 
Rev. 1299 (1984). Moreover, information concerning the 
effects of the employer’s employment practices is readily 
available to plaintiffs through the liberal rules of dis­

375



17

covery and through access to the EEOC’s investigatory 
files; thus, just as a plaintiff has sufficient means for ob­
taining the information necessary to establish that elu­
sive concept of discriminatory “motive,” the plaintiff has 
sufficient means for obtaining the information necessary 
to establish the more tangible concept of discriminatory 
“effect.” See Texas Dep’t of Community Affairs v. Bur- 
dine, 450 U.S. 248, 258 (1981); see also U.S. Postal Sew. 
Bd. of Governors v. Aikens, 460 U.S. 711, 716-717
(1983). Finally, while there may be instances in which 
multiple regression analysis does not provide a clear 
answer, or in which sufficient information is not avail­
able, this Court has made clear that plaintiffs bear the 
burden of proof in impact cases and, a fortiori, that 
plaintiffs bear the risk of loss associated with uncertainty 
or unavailability of proof about causation. See Connecti­
cut v. Teal, 457 U.S. at 446; see also Watson v. Fort 
Worth Bank and Trust, 108 S. Ct. at 2790 (plurality 
opinion) .4 * &

The concern that an employer may devise a scheme un­
der which several components of a selection process, none 
of which individually causes a disparate impact, “inter­
act” to produce a work force imbalance is equally un­
founded. An employer who, without intending to dis­
advantage members of the minority group, devises 
a system of employment practices in which no single prac­

4 Of course, while plaintiffs may sometimes be unable to carry 
their burdens of proof under the disparate impact theory, they are 
much more likely, in such circumstances, to be able to carry their 
burdens under the disparate treatment theory. Courts applying 
disparate treatment theory have been most likely to iind illicit motive 
where a plaintiff has shown that the employer’s selection process 
produced immeasurable results, relied on immeasurable judgments, 
was not well documented, and resulted in a gross work force dis­
parity. See Bauer v. Bailar, 647 F.2d 1037, 1045 (10th Cir. 1981); 
Stewart v. General Motors Corp., 542 F.2d 445, 450-451 (7th Cir.
1976), cert, denied, 433 U.S. 919 (1977); sec generally B. Scldei
& P. Grossman, Employment Discrimination Law, 191-205 (1983).

376



18

tice itself causes a disproportionate exclusion of minori­
ties simply has not violated Title VII. In such a case, the 
bottom line disparity is attributable to an aggregation of 
plainly lawful factors—e.g., applicant drop-out or em­
ployee self-selection, facially neutral practices without ad­
verse impact, and/or chance—and Title VII cannot rea­
sonably be interpreted to prohibit employers from engag­
ing in a combination of lawful acts. See Watson v. Fort 
Worth Bank & Trust, 108 S. Ct. at 2787; Bouncy v. Pru­
dential Ins. Co., 668 F.2d at 801-02; Campbell, supra, 36 
Stan. L. Rev. at 1318. Cases such as Green v. USX 
Corp., supra, and Griffin v. Carlin, supra, provide abso­
lutely no reasoning to support their contrary and un­
founded, assertions.

The Chamber does not mean to suggest that anything 
in this Court’s cases or the policies of Title VII would 
prohibit a Title VII plaintiff, in an appropriate case, 
from using the cumulative effect of an employer’s 
decision-making process as proof of a Title VII violation. 
In appropriate circumstances, a significant imbalance in 
a work force, supported by probative statistical analyses, 
may fairly lead to an inference of intentional dis­
crimination. See International Bhd. of Teamsters v. 
United States, 431 U.S. 324, 335-340 (1977); Ilazelivood 
School Dist. v. United States, 433 U.S. 299, 306-313
(1977). But, in approving the use of statistical imbal­
ances to establish intentional discrimination in such cir­
cumstances, the Court has stressed (Teamsters, 431 U.S. 
at 339-340 n.20) that:

the statistical evidence [cannot be] offered or used 
to support an erroneous theory that Title VII re­
quires an employer’s work force to be racially bal­
anced. Statistics showing racial or ethnic imbalance 
are probative in a case such as this one only because 
such imbalance is often a telltale sign of purposeful 
discrimination; absent explanation, it is ordinarily 
to be expected that nondiscriminatory hiring prac­
tices will in time result in a work force more or less

377



19

representative of the racial and ethnic composition of 
the population in the community from which em­
ployees are hired. Evidence of longlasting and gross 
disparity between the composition of a work force 
and that of the general population thus may be sig­
nificant even though § 703(j) makes clear that Title 
VII imposes no requirement that a work force mir­
ror the general population.

Implicit in this reasoning is the quite correct conclusion 
that § 703 (j ) bars the imposition of liability in non­
intent cases merely because of the cumulative effect of 
an employer’s overall employment practices. For, if a 
finding of intent is not required, and a showing of racial 
disproportion in the bottom line is, without more, suffi­
cient to prove a prima facie violation of Title VII, the 
very purpose of § 703(j)—to preclude the imposition of; 
liability merely because the employer has a racial im­
balance in its work force—would be defeated.0

II. INTERNAL WORK FORCE STATISTICS CANNOT 
DEMONSTRATE THAT MINORITIES HAVE BEEN  
DISPROPORTIONATELY EXCLUDED FROM JOBS 
UNLESS THE EMPLOYER HAS A POLICY OF PRO­
MOTING FROM WITHIN

Even assuming that the cumulative effect of petitioners 
employment practices can be challenged under a disparate 
impact theory, the court below erred in concluding that 5

5 The Chamber recognizes that the Uniform Guidelines on Em­
ployee Selection Procedures (1978), 29 C.F.R. § 1607, define a 
"selection procedure” to include "[a]ny measure, combination of 
measures, or procedure used as a basis for any employment deci­
sion” (29 C.F.R. § 1607.16Q) and subject all such selection proce­
dures to disparate impact analysis (29 C.F.R. § 1607.3A). But, to 
the extent the Guidelines approve the application of disparate im­
pact theory to the cumulative effect of multiple practices, they are 
inconsistent with § 703(j) and, therefore, not deserving of deference 
from this Court. See Espinoza v. Farah Mfg.  Co., 414 U.S. 86, 94-95 
(1973); General Elec. Co. v. Gilbert, 429 U.S. 125, 140-46 (1976); 
Mohasco Corp. v. Silver, 447 U.S. 807, 825 (1980).

378



20

respondents had successfully established that they were 
disproportionately excluded from non-cannery jobs. In­
ternal work force statistics, such as those relied upon by 
the court below, cannot demonstrate that minorities have 
been disproportionately excluded from jobs unless the em­
ployer has a policy of promoting from within, which 
petitioners do not.

It is well-settled that a plaintiff can establish a prima 
facie case of disparate impact based on statistical evi­
dence showing that an employment practice has had a 
disproportionate exclusionary effect on individuals in a 
protected minority group. See Griggs v. Duke Power Co., 
401 U.S. at 430 and n.6; New York City Transit Auth. 
v. Beazer, 440 U.S. 568, 584 (1979). To do so, the plain­
tiff must proffer statistics effectively measuring the effect 
that a challenged selection or employment process has 
had on applicants or employees and show that any meas­
ured disparity is “sufficiently substantial” to establish a 
prima facie case of discrimination. See Watson v. Fort 
Worth Bank & Trust, 108 S. Ct. at 2788-89. Statistics, 
of course, “come in infinite variety.” International Bhd. 
of Teamsters v. United States, 431 U.S. at 340; see 
generally Baldus and Cole, Statistical Proof of Discrim­
ination §4.11 (at 106-11). But, whichever kind of sta­
tistics are used, the resulting data must establish a rea­
sonable proxy for the relevant potential applicant pool; 
otherwise, reasonable conclusions about the rates of ap­
plicant selection and rejection cannot be drawn. See 
Hazelwood School Dist. v. United States, 433 U.S. at 
310-12; Dothard v. Rawlinson, 433 U.S. at 348 (White, 
J., dissenting).

This Court has accordingly rejected statistical proffers 
that distort the potential applicant pool available to the 
employer. In Hazelwood, for example, the Court found 
that including a school district that maintained a teaching 
staff that was 50 percent black “in the relevant market area 
fmightl distort!] the comparison.” 433 U.S. at 310-11.

379



21

Similarly, in Beazer, the Court held that the exclusion of 
methadone users in private treatment programs from the 
available pool improperly skewed the final statistical anal­
ysis. 440 U.S. at 585-86. In short, where a statistical 
proffer has improperly included or excluded particular 
groups of individuals from the potential applicant pool, 
the Court has been unwilling to find that a prima facie 
discriminatory rate of selection or rejection has been 
proved.

The statistics upon which the court below relied are 
likewise distorted. Respondents offered no applicant flow 
statistics. Moreover, under the comparative statistics 
they offered, the pool of cannery workers was treated as 
the relevant applicant pool for non-cannery worker jobs. 
But petitioners receive applications for non-cannery work 
from persons residing in Alaska, the Pacific Northwest, 
and California. In short, the members of the cannery 
worker pool at most constitute only a subset of the group 
of persons who reasonably can and do apply for the non- 
cannerv worker jobs. Reasonable conclusions about the 
rates of selection and rejection of non-whites simply can­
not be drawn from such incomplete and under-inclusive 
data; in these circumstances, internal work force data 
show nothing about the percentages of minorities that an 
employer can reasonably be expected to hire in particular 
jobs. Accordingly, the court below was wrong in finding 
that non-whites had been disproportionately excluded 
from non-cannery worker jobs.

This is not to say that an internal work force compari­
son may never be relevant in a disparate impact case. 
Such a comparison may be relevant where an employer 
promotes only from within." But, here, as the district

fl Even in these circumstances, of course, the internal work force 
data must be adjusted to account for the minimum qualifications 
required by the positions in issue. See Johnson v. Uncle Ben’s, Inc., 
628 F.2d 419, 425 (5th Cir. 1980), vacated on other grounds, 451 
U.S. 902 (1981); Ste. Marie v. Eastern R.R. Ass’n, 650 F.2d 395,

380



22

court found and the court of appeals did not dispute, 
petitioners accept applications from persons both within 
and without the work force. Moreover, as the district court 
also found, it was plainly reasonable for the petitioners 
to do so: Not only do petitioners’ non-cannery worker 
jobs often require skills, training and pre-season avail­
ability that the general cannery worker does not have, 
but the short and intense canning season generally pre­
cludes mid-season training and promoting of cannery 
workers and, rather, requires resort to the external labor 
market. Pet. App. 1:33-36, 40-41, 46-47. Indeed, be­
cause the high wages that petitioners guarantee make 
employment in the canneries attractive to persons in a 
multi-state region, the demands of equal opportunity law 
may well require petitioners to give equal consideration 
to applications received from outside the work force. In 
short, it is clear that the pool of cannery workers is not 
a reasonable proxy for the relevant potential applicant 
pool for non-cannery worker jobs, much less the only rea­
sonable proxy, as the court below implicitly held.* 7

Allowing a prima facie disparate impact case to be 
established simply by proof that an employer has an im­
balanced work force would place such an employer be­
tween Scylla and Charybdis. On the one hand, the em­
ployer would be subject to disparate impact claims from 
the members of the minority group that are concentrated 
at one level of its work force—here, for example, the 
Filipino and Native Alaskan cannery workers. On the

400-01 (2d Cir. 1981); EEOC v. Federal Reserve. Bank, 698 F.2d 
633, 669-60 (4th Cir. 1983), rev’d on other grounds sub nom. Cooper 
v. Federal Reserve Bank, 467 U.S. 867 (1984); Rivera v. City of 
Wichita Falls, 665 F.2d 631, 540-43 (5th Cir. 1982). Respondents 
did not attempt to make such adjustments in this ease.

7 Not only did the court below accept respondents’ plainly flawed 
statistical proffer, it ignored the district court’s conclusion that 
petitioners’ statistical proffer-show ing that, even without regard to 
cjualifications, the potential applicant pool in the states from which 
petitioners have received applications is only ten percent non-white,

381



23

other hand, were the employer to refuse to consider 
applications from persons outside the work force, 
it would be subject to disparate impact claims 
by members of minority groups (and, perhaps, whites) 
who would thereby be deprived of job opportunities— 
here, for example, the Hispanics residing in California 
who have reasonably applied for employment with peti­
tioners. This Court has said that Title VII, and espe­
cially the disparate impact theory, should not be inter­
preted to impose such conflicting legal obligations on an 
employer. See Los Angeles Dep't of Water & Power v. 
Manhart, 435 U.S. 702, 711 & n.20 (1978); Johnson v. 
Transportation Agency, 107 S. Ct. at 1451-52."

Rather, the statute should be interpreted to allow dis­
parate impact claims only where they are based on rea­
sonable proxies for the potential applicant pool. The only 
such proxy identified in this case was the one proffered 
by petitioners. Petitioners showed that they received ap­
plications from persons residing in Alaska, the Pacific * 8

while the non-cannery workers are 21 percent non-white—was the 
more convincing of the two. In ignoring this factual finding of the 
district court, the decision below conflicts with this Court’s decision 
in Anderson v. Bessemer City, 470 U.S. 564 (1985), as well ns with 
the decisions of other courts of appeals, which have found that 
external labor market data effectively rebuts internal work force 
comparisons. See, e.g., Hilton v. Wyman-Gordon Co., 624 F.2d 379, 
382 (1st Cir. 1980); Clark v. Chrysler Corp., 673 F.2d 921, 929 (7th 
Cir.), cert, denied, 459 U.S. 873 (1982); Cose.r v. Moore, 739 F.2d 
746, 752 (2d Cir. 1984).

8 Moreover, one effect of holding that a prima facie disparate 
impact case is established by evidence of a mere work force imbal­
ance would be to discourage employers from engaging in voluntary 
affirmative action. Any affirmative action that increases the per­
centage of minorities in some but not all job classifications might 
create the imbalance necessary for a disparate impact suit and, 
accordingly, employers would have great reason not to engage in 
affirmative action (as opposed to mere quota hiring) at all. The 
Court has said that Title VII should not be interpreted to create 
such disincentives against voluntary affirmative action. See Johnson 
v. Transportation Agency, 107 S. Ct. at 1450-51.

382



24

Northwest, and California. They further demonstrated 
through census data that, even without regard to the 
qualifications required for the non-cannery positions, the 
potential applicant pool in these states is only ten percent 
non-white; and that, by comparison, non-whites have 
filled 21 percent of their non-cannery positions and thus 
are over-represented in those positions. Only one conclu­
sion follows: that petitioners’ selection and employment 
practices have had no cumulative adverse impact on non­
whites.

III. AT THE REBUTTAL STAGE OF A DISPARATE IM­
PACT CASE, AN EMPLOYER NEED ONLY SHOW 
THAT ITS SELECTION DEVICES ARE REASON­
ABLE IN LIGHT OF THE JOB AT ISSUE AND THE  
NATURE OF THE BU SIN ESS

Having wrongly concluded that respondents established 
a prima facie case of discrimination, the court below ex­
acerbated its error by concluding that, with two excep­
tions, petitioners had failed to meet their burden of show­
ing that their employment practices were justified by 
business necessity. The court below not only placed too 
heavy a burden on petitioners—to “prove the job related­
ness or business necessity of the practice” giving rise to 
the disparity (Pet. App. VI:5; emphasis added)—but it 
ignored substantial evidence that petitioners’ practices 
were in fact so justified.

It is well-settled that a plaintiff in a Title VII case 
bears the “ultimate burden of proving a violation of 
Title VII.” New York Transit Anth. v. Beazer, 440 U.S. 
at 587 n.31. It is equally well-settled that a “prima facie 
showing is not the equivalent of a factual finding of dis­
crimination.” Fvrnco Const. Corp. v. Waters, 438 U.S. 
567, 579 (1978). Thus, while the employer in a prima 
facie disparate impact case—like the employer in a prima 
facie disparate treatment case—has a rebuttal burden, 
that burden is not one of persuasion; it is a burden of 
production. A violation of the statute is established—

383



25

satisfying the plaintiff’s burden of persuasion—only when 
an unjustified practice has been shown disproportionately 
to exclude minorities, and that conclusion can be drawn 
only after the assessment of business necessity has been 
made. See Griggs v. Duke Power Co., 401 U.S. at 431. 
Thus, requiring the employer to "prove” business neces­
sity at the rebuttal stage—as the court below did—ef­
fectively converts the plaintiff’s prima facie showing into 
an ultimate finding of discrimination. Neither the statu­
tory language nor this Court’s cases justify such a re­
quirement.®

Nor is there any proper justification for the overly 
demanding standard that the court below applied in re­
jecting petitioners’ explanations for their employment 
practices. To be sure, this Court has described the re­
buttal burden in a disparate impact case as focusing on 
the "business necessity” for the challenged practice 
(Griggs v. Duke Power Co., 401 U.S. at 431) and, on 
occasion, has suggested that, in particular circumstances, 
this burden may necessitate a formal validation study 
(see, e.g., Albemarle Paper Co. v. Moodg, 422 U.S. at

® To be sure, in Burdine, this Court "recognized that the factual 
issues, and therefore the character of the evidence presented, differ 
when the plaintiff claims that a facially neutral employment policy 
has a discriminatory impact on protected classes.” 450 U.S. at 252 
n.6. But, although the "character of evidence presented” may differ, 
this does not mean that a plaintiff’s burden of proof—to prove that 
he was a victim of discrimination— is any less in a disparate impact 
case. See Watson, 108 S. Ct. at 2785 (“Nor do we think it is 
appropriate to hold a defendant liable for unintentional discrimina­
tion on the basis of less evidence than is required to prove inten­
tional discrimination”); Beazer, 440 U.S. at 587 n.31. Indeed, it 
would be "illogical to impose a heavier burden on a defendant in a 
case where a neutral policy results in disparate impact than in one 
where the charge is unlawful animus” ( NAACP v. Medical Center, 
Inc., 657 F.2d 1322, 1335 (3d Cir. 1981)), since 'Tulndoubtedly 
disparate treatment was the most obvious evil Congress had in 
mind when it enacted Title VII” (International Bhd. of Teamsters 
v. United States, 431 U.S. at 335 n.15).

384



26

431). But the Court has also held that an employer may 
satisfy his rebuttal burden with evidence that the selec­
tion process serves the “legitimate employment goals of 
safety and efficiency” (New York Transit Auth. v. 
Beazer, 440 U.S. at 587 n.31), and that such evidence 
need not take the form of a validation study (see ibid.). 
On the contrary, as a plurality of the Justices recently 
reiterated in Watson, a disparate impact claim is rebutted 
when the evidence shows that “employment practices are 
based on legitimate business reasons” (108 S. Ct. at 
2790) ; that is, a disparate impact claim is rebutted by 
evidence “that thef] selection devices—test or nontest—are 
justified in light of the nature of the job and its relation 
to the overall enterprise. Face validity, otherwise known 
as reasonableness, should suffice.” Lerner, Employment 
Discrimination: Adverse Impact, Validity and Equality, 
1979 Sup. Ct. Rev. 17, 39.

Petitioners plainly established the “face validity”—or 
reasonableness—of each of the practices challenged in 
this case. Petitioners showed, and the district court 
found, that they did not engage in nepotism at all. Pet, 
App. 1:103-05. Petitioners also showed that the subjec­
tive job qualifications applied by petitioners were neces­
sary for safe and effective performance of the non­
cannery worker jobs. Id. at 1:35-36, 40-41, 45-47, 107- 
114. Petitioners further showed that job openings were 
not posted because petitioners received more applications 
than they had openings, because they received applica­
tions from a multi-state region, and because cannery 
workers could apply in the off-season—just like everyone 
else—for non-cannery worker jobs. Id. at 1:28-34. Fi­
nally, petitioners showed that petitioners’ race-labeling 
practices had no effect on non-white employees’ job oppor­
tunities and, furthermore, that petitioners’ housing and 
messing practices were structured to accommodate work­
ers’ preferences, the arrival time and departure of 
workers, the costs of providing such benefits, and the

385



27

union’s demands. Id. at 1:123-29. These explanations 
were entirely reasonable in light of the jobs in issue and 
the nature of petitioners’ business, and the court below 
erred in holding that petitioners had failed to meet their 
rebuttal burden.

Of course, had respondents offered evidence that peti­
tioners could have accomplished their legitimate business 
goals and still avoided a disparate impact, the courts 
would have had to consider it. See Watson v. Fort Worth 
Bank & Trust, 108 S. Ct. at 2790 (“the plaintiff must 
‘show that other tests or selection devices, without a sim­
ilarly undesirable racial effect, would also serve the em­
ployer’s legitimate interest in efficient and trustworthy 
workmanship’ ” (quoting Albemarle Paper Co. v. Moody, 
422 U.S. at 425)). But respondents did not do so. They 
simply argued that petitioners could have increased the 
percentage of non-whites in their non-cannery jobs by, 
for example, training cannery workers and promoting 
them to non-cannery positions. This argument ignores, 
of course, the canneries’ legitimate business reasons for 
not implementing such practices—i.e., the short and in­
tense canning season, the infrequency of mid-season job 
vacancies, and the cost of providing such training. See 
Watson v. Fort Worth Bank & Trust, 108 S. Ct. at 2790 
(“ [fjactors such as the cost or other burdens of proposed 
alternative selection devices are relevant . . . ”). But, more 
importantly, the argument rests on a misperception that 
Title VII requires employers to maximize their hiring of 
minority applicants.

This Court has made abundantly clear that Title VII 
“does not impose a duty to adopt a hiring procedure that 
maximizes hiring of minority employees” and that em­
ployers need not “pursuefl the course which would both 
enable |theml to achieve (their 1 own business goal (si 
and allow (them! to consider the most employment appli­
cations.” Furneo Const. Corp. v. Waters, 428 U.S. at

386



28

577-78 (emphasis in original). Any other conclusion 
would only invite courts to “required businesses to adopt 
what [they] perceive[] to be the ‘best’ hiring procedures” 
{id. at 578) and, as the plurality in Watson reiterated, 
“ ‘[cjourts are generally less competent than employers 
to restructure business practices, and unless mandated to 
do so by Congress they should not attempt it’ ” (108 S. 
Ct. at 2791 (citation omitted)).

For the reasons set forth above, the judgment of the 
court below should be reversed.

CONCLUSION

Respectfully submitted,

Glen D. N ager

Litigation Center, Inc. 
1615 H Street, N.W. 
Washington, D.C. 20062 
(202) 463-5337 
September, 1988

Of Counsel:
Stephen A. Bokat 
Mona C. Zeiberg 
N ational Chamber

(Counsel of Record) 
A ndrew M. Kramer 
David A. Copus 
Patricia A. Du n n  
Jones, Day, Reavis & Pogue 
1450 G Street, N.W. 
Washington, D.C. 20005-5701 
(202) 879-3939 
Attorneys for the

Chamber of Commerce of 
the United States of America

387





No. 87-1387

In  T he

guprmp (flmtrt of %  Inifpi'i §tutrn
October T erm , 1988

W ards Cove P acking  Com pany , Inc ., 
Castle & Cooke, Inc .,

Petitioners,
v .

F rank  A tonio, et al,
Respondents.

On Writ of Certiorari to the United States Court of Appeals
for the Ninth Circuit

BRIEF AMICUS CURIAE OF THE 
EQUAL EMPLOYMENT ADVISORY COUNCIL 

IN SUPPORT OF THE PETITIONERS

Robert E. W illiams 
Douglas S. M cDowell 
Edward E. Potter, P.C.* 
M cGuiness & W illiams 
1015 Fifteenth Street, N.W. 
Suite 1200
Washington, D.C. 20005 
(202) 789-8600

Attorneys for the Amicus Curiae 
Equal Employment Advisory 
Council

* Counsel of Record

389



T A B L E  O F C O N T E N T S
P age

INTEREST OF THE AMICUS CURIAE.......................  1

STATEMENT OF THE C A SE .........................................  4

SUMMARY OF ARGUMENT............................................  8

ARGUMENT........................................................................... 9

T A B L E  O F A U T H O R IT IE S .......................................................  iii

I. THE DISPARATE IMPACT THEORY OF TI­
TLE VII MAY BE USED ONLY WHERE A 
SPECIFIC, FACIALLY NEUTRAL EMPLOY­
MENT PRACTICE OR CRITERION APPLY­
ING TO A GROUP OF EMPLOYEES IS 
SHOWN TO CAUSE AN ADVERSE IMPACT 
ON A PROTECTED CLASS OF EMPLOYEES.. 9

A. This Court has Applied the Disparate Impact
Theory Only in Cases Challenging Specific, 
Facially Neutral Employment Practices or 
Criteria When the Plaintiff Has Shown a 
Direct Link Between Those Practices or 
Criteria and Adverse Impact on a Group of 
Employees.............................................................  10

B. Numerous Well Reasoned Decisions of Courts
of Appeals Have Held That The Disparate 
Impact Theory Is Limited To Claims Involv­
ing The Application Of Specific, Facially 
Neutral Employment Practices Or Criteria, 
Because Only In Such Cases Is It Possible 
To Demonstrate A Causal Connection Be­
tween A Particular Practice Or Criterion 
And Adverse Impact On Protected Employ­
ees...........................................................................  15

390



11

II. REQUIRING A DEFENDANT TO PROVE 
THE BUSINESS NECESSITY OF ITS PRAC­
TICES AFTER A PRIM A FACIE  CASE HAS 
BEEN ESTABLISHED—THAT IS, SHIFT­
ING THE BURDEN OF PERSUASION TO 
THE EMPLOYER—INTERFERES WITH 
TRADITIONAL MANAGEMENT PREROGA­
TIVES AND, IN EFFECT, MEANS THAT 
EMPLOYERS MUST ADOPT EITHER 
“BEST” HIRING PRACTICES OR QUOTAS, 
CONTRARY TO THIS COURT’S HOLDINGS

T A B L E  O F C O N T E N T S — C ontinued
P age

IN W ATSON, FURNCO  AND BU RD INE .......  20

CONCLUSION........................................................................ 28

391



Cases
T A B L E  O F A U T H O R IT IE S

P age

Albemarle Paper Co. v. Moody, 422 U.S. 405
(1975) ...............................................................  11,27

American Fed. of State, County, and Municipal
Employees v. State of Washington, 770 F.2d
1401 (9th Cir. 1985) ................................................ 13

Atonio v. Wards Cove Packing Co., Inc., 768 F.2d 
1120 (9th Cir. 1985), cert, denied, 108 S.Ct. 1293
(9th Cir. 1987).................................................. 6, 12,17,18

Atonio v. Wards Cove Packing Co., Inc., 810 F.2d 
1477, cert, denied, 108 S.Ct. 1293 (9th Cir.
1987) ................................................................. 3, 6, 7, 16, 19

Atonio v. Wards Cove Packing Co., Inc., 827 F.2d
439 (9th Cir. 1987) ................ ................................  7

Bazemore v. Friday, 106 S.Ct. 3000 (1986)............. 3
Blake v. City of Los Angeles, 595 F.2d 1367 (9th

Cir. 1979), cert, denied, 446 U.S. 928 (1980).....  27
Carroll v. Sears, Roebuck & Co., 708 F.2d 183 (5th

Cir. 1983)............................................................... 16,18, 19
Chrisner v. Complete Auto Transit, Inc., 645 F.2d

1251 (6th Cir. 1981)................................................. 26
Connecticut v. Teal, 457 U.S. 440 (1982).... 3, 8,12,13, 26
Dothard v. Rawlinson, 433 U.S. 321 (1977)............. 11
EEOC v. Kimbrough Investment Co., 703 F.2d 98

(5th Cir. 1983)...........................................................  26
Furnco Construction Corp. v. Waters, 438 U.S. 567

(1978).................................................... 3,11,12, 23, 24,25
General Electric Co. v. Gilbert, 429 U.S. 125

(1976) ..........................................................................  11
Green v. USX Corporation, 843 F.2d 1511 (3d Cir.

1988) , petition for cert, pending, (88-141).....  3
Griffin v. Board of Regents of Regency Universi­

ties, 795 F.2d 1281 (7th Cir. 1986)......................  18
Griffin v. Carlin, 755 F.2d 1516 (11th Cir. 1985).... 16
Griggs v. Duke Power Co., 401 U.S. 424 (1971).... 10,11,

16, 20, 24
Hazelwood School District v. United States, 433 

U.S. 299 (1977) 14



Hill v. Western Electric Co., Inc., 596 F.2d 99 (4th
Cir. 1979) ...................................................................  15

Hilton v. Wyman-Gordon Co., 624 F.2d 379 (1st
Cir. 1980) ...................................................................  14

International Brotherhood of Teamsters v. United
States, 431 U.S. 324 (1977) ............................... 2, 3, 9, 14

Kirby v. Colony Furniture Co., Inc., 613 F.2d 696
(8th Cir. 1980)........................................................... 2°. 27

Eatinos Unidos de Chelsea En Accion v. Secretai y 
of Housing and Urban Development, 799 F.2d
774 (1st Cir. 1986).................................................. 16

Lee v. Washington County Bd. of Ednc., 625 F.2d
1235 (5th Cir. 1980).............................................- l'r>

Maddox v. Claytor, 764 F.2d 1539 (11th Cir,
1985) ............................................................................ 8» 16

Mazus v. Department of Transportation of Penn­
sylvania, 629 F.2d 870 (3d Cir. 1980), cert, de­
nied, 4 49 U.S. 1126 (1981).......................................  14

Nashville Gas Co. v. Satty, 434. U.S. 136 (1977).... 11
New York City Transit Authority v. Beazer, 440

U.S. 568 (1979).................................................... H» 14. 16
Pack v. Energy Research and Development Adm in­

istration, 566 F.2d 1111 (9th Cir. 1977)............. 15
Pettway v. American Cast Iron Pipe Co., 494 F.2d 

211 (5th Cir. 1974), cert, denied, 467 U.S. 1243
(1984)    27

Pope v. City of Hickory, North Carolina, 679 F.2d
20 (4th Cir. 1982)....................................................  17

Pouncy v. Prudential Insurance Co., 668 F.2d 795
(5th Cir. 1982)..............................................2. *R. I7. i8

Robinson v. Polaroid Corp., 732 F.2d 1010 (1st Cir.
1984)..............................................-............................. 16' 1R

Spaulding v. University of Washington, 740 F.2d 
686, (9th Cir.), cert, denied, 469 U.S. 1036
(1984)   I5' 1°

Ste. Marie v. Eastern Railroad Association, 650
F.2d 395 (2d Cir. 1981)..........................................  14

iv

T A B L E  O F A U T H O R IT IE S — C ontinued
P age

393



V

T A B L E  O F A U T H O R IT IE S — C ontinued
P a g e

Talley v. United States Postal Service, 720 F.2d 
505 (8th Cir. 1983), cert, denied, 466 U.S. 952
(1984)   16

Texas Department of Community Affairs v. Bur-
dine, 450 U.S. 248 (1981) .......................................  3, 25

United A ir Lines, Inc. v. Evans, 431 U.S. 553
(1971) .......................................................................... 3

United States v. Bethlehem Steel Corp., 446 F.2d
652 (2d Cir. 1971)....................................................  26

United States Postal Service Board of Governors
v. Aikens, 460 U.S. 711 (1983).............................. 3

United Steelivorkers v. Weber, 443 U.S. 193
(1979)    24

Watson v. Fort Worth Bank Trust, 108 S.Ct.
2777 (1988).......................... 2, 3, 6, 8, 9, 12, 13, 19, 23, 25

Statutes
Civil Rights Act of 1964, Title VII, as amended,

42 U.S.C. 2000e, et seq............................................
Section 7 0 3 (a )(2 ); 42 U.S.C. § 2000e-2(a)

(2) ................................
Civil Rights Act of 1866, 42 U.S.C. § 1981...............

Rides and Regulations
Uniform Guidelines on Employee Selection Proce­

dures, 29 C.F.R. §§ 1607 et seq.....................20, 22, 26, 27

Congressional History
122 Cong. Rec. 22950 (daily ed. July, 1976).......
H.R. Rep. No. 914, 88th Cong., 1st Sess., pt. 2, p.

29 (1963) U.S. Code Cong. & Admin. News 1964, 
at 2391..........................................................................

Miscellaneous
Balleu, Courts, Psychologists, and the EEOC’s Uni­

form  Guidelines: An Analysis of Recent Trends 
Affecting Testing as a Means of Employee Selec­
tion, 36 Emory L.J. 203 (1987).............................  22

22

24

2,5

10
5

394



Vi

Bureau of National Affairs, Recruiting and Selec­
tion Procedures, PF Survey No. 146 (May
1988) ...........................................................................  21

Daily Lab. Rep. (BNA) D-14 (Dec. 5, 1978).......  22
Gwartney, Ashner, Haworth, Haworth, Statistics, 

the Law and Title VII:  A n  Economist’s View,
54 Notre Dame L. Rev. 633 (1979) ......................  21, 22

National Research Council, Ability Tests: Uses,
Consequences and Controversies (1982)............. 22

Potter, Employee Selection: Legal and Practical 
Altemvatives to Compliance and Litigation
(1986) .........................................................................  22

Rutherglen, Disparate Impact Under Title VII:
A n Objective Theory of Discrimination, 73 Va.
L. Rev. 1297 (1987) .................................................. 25

Schlei and Grossman, Employment Discrimination
Law, at 1329 (1983)................................................  26

T A B L E  O F A U T H O R IT IE S — C ontinued
P age

395



I n  T h e

S’ltpmttP (flmul of tiff g ’latpo
October T erm , 1988

No. 87-1387

W ards Cove P a ck in g  Co m pa n y , I n c ., 
Castle  & Cooke, In c .,

Petitioners,v.
F r a n k  A tonio , et al.,

Respondents.

On Writ of Certiorari to the United States Court of Appeals
for the Ninth Circuit

BRIEF AMICUS CURIAE OF THE 
EQUAL EMPLOYMENT ADVISORY COUNCIL 

IN SUPPORT OF THE PETITIONERS

T h e  E q u a l E m p lo y m e n t A d v is o r y  C o u n cil (E E A C )  r e ­
s p e c t fu lly  su b m its  th is  b r ie f  a s  a m ic u s  c u r ia e  in  su p p o r t  
o f  th e  P e t it io n e r s . T h e  p a r t ie s ’ w r it te n  c o n se n ts  h a v e  
b een  filed  w ith  th e  C lerk  o f  th is  C ou rt.

INTEREST OF THE AMICUS CURIAE

T h e  E E A C  is  a n  a s so c ia t io n  o f  e m p lo y er s  o r g a n iz e d  to  
p ro m o te  so u n d , p r a c tic a l a p p ro a c h e s  to  eq u a l e m p lo y m en t  
o p p o r tu n ity  a n d  a ffirm a tiv e  a c tio n . I t s  m em b ersh ip  com ­
p r is e s  a  b road  s e g m e n t  o f  th e  em p lo y er  c o m m u n ity  in  
th e  U n ite d  S ta te s , in c lu d in g  a b o u t 19 0  la r g e  em p lo y ers  
an d  tr a d e  an d  in d u s tr y  a s s o c ia t io n s  lo ca ted  th r o u g h o u t  
th e  c o u n tr y . I t s  g o v e r n in g  bod y is  a  B oa rd  o f  D ir e c to r s

396



2

com posed  o f  e x p e r ts  in  eq u a l e m p lo y m en t o p p o r tu n ity . 
T h eir  co m b in ed  e x p e r ie n c e  g iv e s  th e  C ou n cil an  in -d ep th  
u n d e r s ta n d in g  o f  th e  p r a c tic a l, a s  w e ll a s  leg a l a sp e c ts  o f  
equal e m p lo y m e n t p o lic ie s  an d  r eq u ire m en ts . T h e  m em ­
b ers o f  th e  C ou n cil a r e  f irm ly  co m m itted  to  th e  p r in c ip le s  
o f n o n d isc r im in a tio n  an d  eq u a l e m p lo y m en t o p p o r tu n ity .

A s  em p lo y er s , E E A C ’s m em b ers a re  su b je c t  to  T it le  
V II o f  th e  C iv il R ig h ts  A c t  o f  19G4, as amended, 42  
U .S .C . § 2 0 0 0 e  et seq., th e  s ta tu te  a t  is su e  in  th is  ca se , a s  
w ell a s  o th e r  eq u a l e m p lo y m en t s ta tu te s  an d  r e g u la t io n s . 
L a st ter m , in  Watson v. Fort Worth Bank & Trust, 108
S .C t. 2 7 7 7  ( 1 9 8 8 ) ,  th is  C o u rt ex p a n d ed  th e  a p p lic a b ility  
o f th e  d is p a r a te  im p a c t th e o ry  to  su b je c t iv e  c r ite r ia  and  
p r a c tic e s  in v o lv in g  an  in d iv id u a l p la in t if f . A  p lu r a lity  
o f th e  C o u rt s u g g e s te d  a n  a n a ly t ic a l fr a m e w o r k  fo r  a p ­
p ly in g  th e  d is p a r a te  im p a c t th e o ry  in  su ch  c a se s . T h is  
ca se  p r e se n ts  th e  f ir s t  o p p o r tu n ity  fo r  th e  fu l l  C o u rt to  
a p p ly  an d  c la r ify  th e  Watson d e c is io n  in  th e  c o n te x t  o f  a  
c la ss  a c t io n  u n d e r  T it le  V I I . A s  a r e su lt , E E A C ’s m em ­
b ers a r e  v ita l ly  in te r e s te d  in  th e  is su e s  b e fo re  th e  C ou rt  
in  th is  c a se , w h ic h  co n cern  th e  p ro p er  u se  o f  s ta t is t ic s  
and  th e  b u rd en s  o f  p r o o f  u n d er  th e  d isp a r a te  im p a c t  
th eo ry  w h en  th e  em p lo y er  se le c ts  sk illed  an d  u n sk illed  
em p lo y ees  fro m  d iffe r e n t  lab or  m a r k e ts  u s in g  m u lt ip le , 
su b je c t iv e  a n d  o b je c tiv e  c r ite r ia  an d  p r a c tic e s .

In  International Brotherhood of Teamsters v. United 
States, 4 3 1  U .S . 3 2 4 , 3 3 5  n .1 5  ( 1 9 7 7 ) ,  th e  C o u rt s ta te d  
th a t, in  a p p r o p r ia te  c ir c u m sta n c e s , e ith e r  a  d isp a r a te  
tr e a tm e n t  or d isp a r a te  im p a c t th e o ry  m a y  be a p p lied  to  a 
p a r tic u la r  s e t  o f  fa c ts .  W ith  r e sp ec t to  th e  u se  o f  s t a t i s ­
t ic s  in  c la s s  a c t io n s  to  e s ta b lish  a  p r im a  fa c ie  ca se , lo w er  
co u r ts , in c lu d in g  th e  N in th  C ir cu it , h a v e  fa i le d  to  ta k e  
in to  a c co u n t th e  d if fe r e n t  p u rp o ses  o f  th e  tw o  th eo r ie s . 
T h e C o u rt in  th is  c a se  h a s  a n  o p p o r tu n ity  to  m a k e  c le a r  
th a t  th e  sa m e  se t  o f  s ta t is t ic s  m a y  n o t n e c e ssa r ily  e s ta b ­
lish  d isp a r a te  tr e a tm e n t  and  im p a c t in a  p a r t ic u la r  ca se .

397



3

That is, although statistical disparities may sometimes 
be probative of disparate treatment, they are not proba­
tive of disparate impact unless a causal connection is 
shown between the disparities and some specific, facially 
neutral employment practice.

Furthermore, in Watson a majority of the Court recog­
nized the difficulties of validating subjective selection cri­
teria and a plurality of the Court stated that the em­
ployer’s rebuttal burden is to state a legitimate business 
reason. For the benefit of Title VII litigants and the 
lower courts, this Court should clearly adopt the Watson 
plurality’s discussion of the defendant’s rebuttal burden 
in a disparate impact case, and should make clear that 
it is the same for both objective and subjective criteria.

EEAC has filed amicus curiae briefs in numerous cases 
concerning the appropriate use of statistics and the ap­
plicability of the disparate impact theory to subjective 
criteria. See Watson v. Fort Worth Bank & Trust, 108 
S.Ct. 2777 (1988); Atonio v. Wards Cove Packing Com­
pany, Inc., 810 F.2d 1477 (9th Cir. 1987) (en bane); 
Green v. USX Corporation, 843 F.2d 1511, (3d Cir. 
1988), pet. for cert, pending (No. 88-141); and Pouncy 
v. Prudential Insurance Co., 668 F.2d 795 (5th Cir. 
1982). EEAC also has filed several briefs amicus curiae 
in this Court in cases involving burden of proof and sta­
tistical issues. See, e.g., Bazemore v. Friday, 106 S. Ct. 
3000 (1986) ; United States Postal Service Board of Gov­
ernors v. Aikens, 460 U.S. 711 (1983); Connecticut v. 
Teal, 457 U.S. 440 (1982); Texas Department of Com­
munity Affairs v. Burdine, 450 U.S. 248 (1981) ; Furnco 
Construction Corp. v. Waters, 438 U.S. 567 (1978); 
United Air Lines, Inc. v. Evans, 431 U.S. 553 (1977); 
and International Brotherhood of Teamsters v. United 
States, 431 U.S. 324 (1977).

398



4

STATEM ENT OF TIIE CASE

The defendant companies operate five canneries located 
in remote and widely separated areas of Alaska. The 
canneries are open only for a short period of time during 
the salmon runs each summer and are vacant the rest of 
the year. Skilled workers are brought in prior to the 
fishing season to assemble canning equipment, repair 
winter damage to the facilities and otherwise prepare the 
entire cannery for the season. They are retained during 
and after the season to maintain and disassemble equip­
ment. The trial court found there was too little time 
during the preseason to train unskilled workers for these 
skilled jobs, because the work is intense and involves ex­
tensive overtime. Pet. for Cert. 1-18-19. The unskilled 
cannery workers, who comprise most of the summer work 
force, arrive shortly before the fishing begins and remain 
at the cannery as long as there are fish to be canned. Be­
cause salmon are very perishable, the canneries operate 
virtually around the clock during fishing season.

Hiring for all jobs except for some cannery workers 
living in Alaska occurs at the defendants’ home offices in 
Seattle, Washington and Astoria, Oregon during the first 
three months of the year. Many of the jobs at the de­
fendants’ facilities are covered by union contracts which 
have rehire preference clauses. The defendants receive 
many more applications than there are vacancies for the 
upcoming season. The district court found that the ma­
jority of the applicants for skilled positions were whites, 
and relatively few non-whites applied for those positions. 
Id. at 1-31-32. It also concluded that, because of the 
sparse population in Alaska, it would not be reasonable 
from a business standpoint to seek applicants there for 
skilled jobs. Id. at 1-32. Unskilled cannery jobs are filled 
through rehire preferences or through the dispatch pro­
cedure of a Filipino union local in Seattle, and workers 
performing those jobs are predominantly non-white. The

399



5

trial court also found that the skills acquired in most 
cannery worker jobs did not provide training for skilled 
jobs. Id. at 1-40.

The plaintiffs brought a class action against the Peti­
tioner companies alleging disparate treatment and impact 
under Title VII of the 1964 Civil Rights Act, as amended, 
42 U.S.C. §§ 2000e et seq., and the Civil Rights Act of 
1866, 42 U.S.C. § 1981. Specifically, they claimed that as 
unskilled cannery workers, they were discriminated 
against in hiring and promotion to skilled jobs, as well 
as with respect to the companies’ housing and messing 
practices. In a wide-ranging attack, the plaintiffs iden­
tified 16 practices which they asserted caused a concen­
tration of nonwhites in the cannery positions, including 
English language skill requirements and nepotism. In 
addition to anecdotal evidence, they attempted to support 
their claims of disparate treatment and impact with two 
kinds of general statistical evidence— (1) comparisons 
between the racial composition of the defendants’ skilled 
jobs and the racial composition of the available external 
labor supply, and (2) comparisons between the racial 
composition of defendants’ skilled jobs and the racial com­
position of the defendants’ unskilled jobs.

With respect to the allegations of disparate treatment, 
the trial court concluded that the plaintiffs had not proved 
the individual instances of discrimination and accorded 
the plaintiffs’ statistics little probative value because they 
did not reflect the pool of employees who had the requisite 
skills or who were available for preseason work. The dis­
trict court applied a disparate impact analysis to the 
English language requirement and nepotism claims, and 
found for the defendants. It declined to apply the dis­
parate impact theory to subjective criteria and practices.

On appeal, a Ninth Circuit panel in Atonio l affirmed 
the decision of the lower court. In particular, it held 
that:

400



6

[PJractices and policies such as a lack of well- 
defined criteria, subjective decision making, hiring 
from different sources or channels, word-of-mouth 
recruitment, and segregated housing and messing, 
which are not facially neutral, lend themselves far 
better to scrutiny for intentional discrimination. 
Consequently, we hold that disparate impact analysis 
was correctly withheld by the district court when 
considering these claims.

Atonio v. Wards Cove Packing Company, 768 F.2d 1120, 
1133 (9th Cir. 1985).

Thereafter, the case was presented for en banc review. 
Consistent with this Court’s subsequent decision in Wat­
son v. Fort Worth Bank & Trust, 108 S.Ct. 2777 (1988), 
the en banc panel in Atonio II held that:

[D]isparate impact analysis may be applied to chal­
lenge subjective employment practices or criteria 
provided the plaintiffs have proved a causal connec­
tion between those practices and the demonstrated 
impact on members of a protected class. The three 
elements of the plaintiffs’ prima facie case are that 
they must (1) show a significant disparate impact 
on a protected class, (2) identify specific employ­
ment practices or selection criteria and (3) show the 
causal relationship between the identified practices 
and the impact.

810 F.2d 1477, 1482 (9th Cir. 1987). Significantly, a 
concurring opinion stressed that the disparate impact 
theory is designed to be applied to certain types of cases 
only, and that the disparate treatment and impact theories 
may not be used interchangeably in any given fact situa­
tion. Id. at 1486-1494.

Contrary to the plurality of this Court that reached 
the issue in Watson, however, the en banc panel also held 
that:

401



7

The crucial difference between a treatment and im­
pact allegation is the intermediate burden on the 
employer. To rebut the prima facie showing of dis­
parate impact the employer may refute the statisti­
cal evidence as in the treatment claim and show that 
no disparity exists. But if the employer defends by 
explaining the reason for the disparity he must do 
more than articulate the reason. He must prove the 
job relatedness or business necessity of the practice.

Id. at 1485.

On remand from the Ninth Circuit en banc, the origi­
nal panel erroneously held in Atonio III that the “quan­
tity and quality of the statistical evidence which will give 
rise to an inference [of disparate impact] is the same as 
that which will give rise to an inference of discrimina­
tory intent.” 827 F.2d 439, 442 (9th Cir. 1987). The 
panel noted that the district court had found the plain­
tiffs’ comparative statistics showing a concentration of 
minorities in unskilled cannery worker jobs to be proba­
tive of intentional discrimination. Nevertheless, it found 
that they had established disparate impact, even though 
the plaintiffs did not show that the workforce imbalance 
was specifically due to particular criteria, as required by 
Atonio II.

Moreover, the panel simply disregarded the defendants’ 
statistics showing that, while external availability of non­
whites for skilled jobs was 2.5 percent to 20 percent, non­
whites actually were employed in about 21 percent of the 
non-cannery positions in the defendants’ Alaska opera­
tions. In addition, even though the plaintiffs had not 
shown that their statistical disparities were caused by 
the adverse impact of the identified selection criteria and 
practices, as required by the en banc decision, the Atonio 
III panel proceeded to consider defendants’ business ex­
planations for those criteria and practices and found 
them insufficient to prove business necessity.

402



8

SUMMARY OF ARGUMENT

Numerous federal appellate courts, and at least four 
justices of this Court, have correctly recognized that the 
disparate impact theory is appropriate, not for wide- 
ranging attacks on a company’s employment practices, 
but only for challenges aimed at clearly delineated, fa­
cially neutral employment policies that can be shown to 
have significantly disparate effects on different race or 
sex groups. See, e.g., Watson v. Fort Worth Bank & 
Trust, 108 S.Ct. 2777 (1988) (plurality opinion); Con­
necticut v. Teal, 457 U.S. 440 (1982) ; Bouncy v. Pruden­
tial Insurance Co., 668 F.2d 795 (5th Cir. 1982), and 
Maddox v. Claytor, 764 F.2d 1539, 1548 (11th Cir. 1985). 
The Ninth Circuit therefore erred in requiring the trial 
court to evaluate the business necessity of almost all of 
the defendants’ hiring and promotion practices and cri­
teria, when none had been shown to cause the workforce 
imbalance. Unless such a causal connection is shown, it 
cannot be said with any degree of certainty that the sta­
tistical disparity to which the plaintiffs point was an 
“effect”, or resulted from the “impact”, of the defendant’s 
employment practices. Hence, there is no basis under a 
disparate impact or “effects” test for requiring the de­
fendants to justify those practices.

If plaintiffs can establish prima facie discrimination 
under the impact theory simply by introducing general 
comparative statistics showing a workforce imbalance, 
without having to show that the imbalance is causally 
linked to one or more specific practices of the employer, 
then employers will effectively be forced to justify their 
entire selection processes in virtually every Title VII 
lawsuit, unless they can rebut the plaintiff’s statistics. 
And if, to meet this rebuttal burden, employers are then 
required to prove legitimate business reasons for their 
selection procedures by validation or some other strict 
standard, a prima facie case will almost inevitably lead 
to a finding of discrimination. As a consequence, “quotas

403



9

and preferential treatment [may] become the only cost- 
effective means of avoiding expensive litigation and po­
tentially catastrophic liability” for employers. Watson, 
108 S.Ct. at 2788 (O’Connor, J., for a plurality of the 
Court). That result would be directly contrary to the 
expressed will of the Congress that enacted Title VII. 
Accordingly, this Court should eschew the reasoning of 
the Ninth Circuit panel in Atonio III, and instead apply 
the principles articulated by the Watson plurality.

ARGUMENT

I. THE DISPARATE IMPACT THEORY OF TITLE YII 
MAY BE U SE D  ONLY W HERE A SPECIFIC, FA ­
CIALLY NEU TRAL EMPLOYMENT PRACTICE OR 
CRITERION APPLYING  TO A GROUP OF EM­
PLOYEES IS SHOWN TO CAUSE AN ADVERSE  
IM PACT ON A PROTECTED CLASS OF EMPLOY­
EES.

In International Brotherhood of Teamsters v. United 
States, 431 U.S. 324, 335 n.15 (1977), the Supreme Court 
stated that, in appropriate circumstances, either a dis­
parate treatment or disparate impact theory may be ap­
plied to a particular set of facts. For both theories to be 
applicable in a particular case, however, the plaintiff 
must allege facts that give rise to the application of both 
theories. Although the plaintiffs here have identified 16 
practices or criteria that they contend resulted in ad­
verse impact, they have failed to show a causal link be­
tween any of those criteria or practices and the statistical 
disparities to which they point. Proof of such a causal 
connection must be recognized as a sine qua non of the 
disparate impact theory. Without such a connection, 
there is no basis for finding that the perceived disparity 
was in fact an “effect,” or resulted from the “impact”, 
of the challenged criteria or practices.

If a prima facie case can be based on statistics like 
those relied on by the plaintiffs here—showing only that

404



10

a workforce imbalance exists but telling nothing whatever 
about its cause—then an employer charged under that 
theory will effectively be required to justify all of its 
employment practices, when in fact, none of them may 
have caused the imbalance. Such a result would be 
contrary to well established Title VII principles and pre­
cedents, as discussed below. Glaring workforce imbal­
ances revealed by general comparative statistics may be 
sufficient in some cases to raise an inference of discrim­
inatory intent, but in this case the trial court specifically 
found that the plaintiffs’ statistics were inadequate to 
show disparate treatment. For the court of appeals to 
hold, nevertheless, that these same statistics could estab­
lish a prima facie case under the impact theory was, we 
submit, clear error.

A. This Court has Applied the Disparate Impact 
Theory Only in Cases Challenging Specific, Facially  
N eutral Em ploym ent Practices or Criteria When 
the P laintiff Has Shown a Direct Link Between 
Those Practices or Criteria and Adverse Impact on 
a Group of Employees.

The Court has carefully guarded the distinction be­
tween the disparate impact and disparate treatment 
theories of Title VII. Consistently, the Court has applied 
disparate impact only to claims challenging specific, 
facially neutral employment practices or criteria that 
adversely affect a protected class of employees.

The Court adopted the impact theory in Griggs v. Duke 
Power Co., 401 U.S. 424, (1971), as a judicial gloss on 
Section 703(a) (2) of Title VII.1 Griggs held that “prac­

1 Section 7 0 3 (a )(2 ) provides:
It shall be an unlawful employment practice for an employer—
* * * * * to limit, segregate or classify his employees or applicants 
for employment in any way which would deprive or tend to 
deprive any individual of employment opportunities or other­
wise adversely affect his status as an employee, because of 
such individual’s race, color, religion, sex, or national origin.

42 U.S.C. § 2000e-2(a)(2) (1981).
405



11

tices, procedures or tests neutral on their face, and even 
neutral in terms of intent, cannot be maintained if they 
operate to ‘freeze’ the status quo of prior discriminatory 
employment practices.” 401 U.S. at 430 (emphasis 
added). The “practices, procedures or tests” at issue in 
Griggs were requirements for a high school education and 
a passing score on a standardized general intelligence 
test. The Court found those requirements to be in viola­
tion of Title VII because they were insufficiently related 
to the jobs for which they were used. Id. at 431-34. Thus, 
GHggs interpreted Section 703(a) (2) to prohibit specific, 
facially neutral employment practices or criteria that 
operate to discriminate against protected classes of 
employees.2

In Furnco Construction Corp. v. Waters, 438 U.S. 567
(1978), the Court expressly contrasted devices like those 
at issue in Griggs with selection procedures with multiple 
practices or criteria. The plaintiffs in Fumco challenged

2 A number of other decisions of this Court applying disparate 
impact analysis also have been limited to the narrow context of 
specific, facially neutral practices or criteria. For example, Albe­
marle Paper Co. v. Moody, 422 U.S. 405, 425 (1975) applied the 
impact theory to employment tests that “solectfedl applicants for 
hire or promotion in a racial pattern significantly different from 
that of the pool of applicants." In General Electric Co. v. Gilbert, 
429 U.S. 125 (1976), the Court relied on the impact theory in 
reviewing a rule that excluded women from a disability plan based 
on pregnancy. Dothard v. Rawlinson, 433 U.S. 321 (1977), pre­
sented the Court with height and weight criteria that adversely 
affected women and could not be shown to have a “business neces­
sity." The Court recognized that in dealing with such “facially 
neutral qualification standards" impact analysis should be allowed, 
and emphasized that a prima facie case is shown under the impact 
theory by demonstrating that the particular criteria in question 
actually caused the selection of applicants in a discriminatory man­
ner. Id. at 328-29. See also Nashville Gas Co. v. Satty,  434 U.S. 
136 (1977) (impact analysis applied to rule denying accumulated 
seniority to employees return to work after pregnancy); New York 
City Transit Authority v. Beazer, 440 U.S. 568 (1979) (impact 
theory used to review anti-narcotics ru le).

406



12

practices by which the employer hired only persons he 
considered to be experienced and competent or who were 
referred to him as similarly skilled. The Court refused 
to apply the impact theory, noting that the case “did not 
involve employment tests . . .  or particularized require­
ments such as . . . height and weight specifications . . . .” 
Id. at 575 n. 7 (emphasis added). Moreover, this Court 
in Connecticut v. Teal, 457 U.S. 440, 455 (1982), ack­
nowledged that the practices in Furnco, like some of the 
selection practices identified in Atonio I, 768 F.2d at 1133 
(quoted at p. 6, supra) involved facially discriminatory 
rather than facially neutral policies.

In Teal, the Court for the first time applied the dis­
parate impact theory to a multicriteria selection process 
involving a test, past work performance, supervisors’ 
recommendations and seniority. Unlike the plaintiffs 
herein and in Furnco, the plaintiffs in Teal were able to 
demonstrate that one of the criteria—the test—had an 
adverse impact against minorities even though the over­
all selection process did not have adverse impact against 
minorities. The Court emphasized that its decisions ap­
plying the impact theory had “consistently focused on 
employment and promotion requirements that create a 
discriminatory bar to opportunities” and had “never read 
§ 703(a) (2) as requiring the focus to be placed instead 
on the overall number of minority or female applicants 
actually hired or promoted.” 457 U.S. at 450 (emphasis 
in original). In the context of this case, Teal means that, 
just as a racially balanced “bottom line” does not insulate 
an employer from liabilty from disparate impact under 
Title VII, a workforce imbalance that has not been shown 
to be caused by a specific, facially neutral selection cri­
terion or practice cannot serve as the basis for establish­
ing a prima facie case of disparate impact.

Finally, in Watson, Justice O’Connor, in an opinion 
joined by at least a plurality of the Court, effectively 
summarized the rules for establishing a prima facie case

407



13

of disparate impact that have been consistently applied 
by the Court since Griggs:

First, we note that the plaintiff’s burden goes beyond 
the need to show that there are statistical disparities 
in the employer’s work force. The plaintiff must 
begin by identifying the specific employment prac­
tice that is challenged. . . . [T]he plaintiff is in our 
view responsible for isolating and identifying the 
specific employment practices that are allegedly re­
sponsible for any observed statistical disparities. Cf. 
Connecticut v. Teal, 457 U.S. 440 . . . (1982).
Once the employment practice at issue has been iden­
tified, causation must be proved; that is, the plaintiff 
must offer statistical evidence of a kind and degree 
sufficient to show that the practice in question caused 
the exclusion of applicants for jobs or promotions 
because of their membership in a protected group. 
Our formulations . . . have consistently stressed that 
statistical disparities must be sufficiently subtantial 
that they raise such an inference of causation.

108 S.Ct. at 2788-2789 (O’Connor, J . ) ; id. at 2792 n.2 
(Blackmun, J .), See also AFSCME v. State of Washing­
ton, 770 F.2d 1401, 1405 (9th Cir. 1985) (Kennedy, J.) 
(“Disparate impact analysis is confined to cases that chal­
lenge a specific, clearly delineated employment practice 
applied at a single point in the job selection process” ). 
Amicus urges the full Court to endorse these principles 
here.

In this case, while the plaintiffs have identified and 
challenged en masse 16 specific employment practices and 
criteria, they have not isolated any particular criterion 
or practice nor shown it to be causally linked to the sta­
tistical disparities to which they point. Thus, they have 
failed to establish the sine qua non of disparate impact— 
a showing that the disparity was, in fact, the effect of 
the employer’s practice.

408



14

To the extent that the court below ignored the require­
ment of proof of causation articulated in Watson, it also 
erred in asserting as a general proposition that “statis­
tical evidence, which will give rise to an inference [of 
disparate impact] is the same as that which will give 
rise to an inference of discriminatory intent.” 827 F.2d 
at 442. The proposition, as stated, is overly broad. Glar­
ing statistical disparities standing alone may sometimes 
be sufficient to raise an inference of discriminatory in­
tent, but without proof of a causal link to some specific 
employment criterion or practice, the same statistics will 
not suffice to make out a case of disparate impact.

In any event, amicus further contends that qualified 
applicant flow statistics showing selection rates based on 
specified criteria, not representation statistics, are the 
appropriate statistical basis for determining whether 
there was disparate impact in this case. In Teamsters, 
431 U.S. at 342 n.23, and Hazelwood School District v. 
United States, 433 U.S. at 308 n.13, the Court recognized 
the superiority of actual applicant flow data over repre­
sentation statistics for purposes of pass-fail comparisons 
of adverse impact. Moreover, in Beazer, the Court re­
quired the use of actual applicant flow data to determine 
adverse impact, holding that general population data 
“tells us nothing about the class of otherwise qualified 
applicants and employees” and, therefore are “virtually 
irrelevant.” 440 U.S. at 585-86. Indeed, as the trial 
court found in this case, nonwhites did not apply nor were 
they deterred from applying to skilled jobs. The absence 
of interest in these positions by minorities serves as an 
absolute defense to allegations of job segregation.3 8

8 See, e.g., Ste. Marie v. Eastern Railroad Association, 6B0 F.2d 
395, 403 (2d Cir. 1981) (women not interested in being railroad 
inspectors); Mazus v. Department of Transportation, 629 F.2d 870 
(3d Cir. 1980), cert, denied, 449 U.S. 1126 (1981) (few  women 
sought road maintenance positions) ; Hilton v. Wyman-Gordon Co., 
624 F.2d 379 (1st Cir. 1980) (distribution of workers due to nu-

409



15

B. Numerous Well Reasoned Decisions of Courts of 
Appeals Have Held That The Disparate Impact 
Theory Is Limited To Claims Involving The Ap­
plication Of Specific, Facially Neutral Employment 
Practices Or Criteria, Because Only In Such Cases 
Is It Possible To Demonstrate A Causal Connection 
Between A Particular Practice Or Criterion And 
Adverse Impact On Protected Employees.

Under a proper interpretation of Title VII, disparate 
impact analysis is not the “appropriate vehicle from 
which to launch a wide ranging attack on the cumulative 
effect of a company’s employment practices.” Spaulding 
v. University of Washington, 740 F.2d 686, 707 (9th 
Cir.), cert, denied, 469 U.S. 1036 (1984), quoting Pouncy 
v. Prudential Insurance Co., 668 F.2d 795, 800 (5th Cir. 
1982). When removed from the context of challenges to 
“clearly delineated neutral policies of employers,” the 
disparate impact theory becomes too vague to be appli­
cable. Id. at 708 (emphasis added). Thus, because im­
pact analsyis was developed “to handle specific employ-

merous factors including employee preferences, job qualifications, 
and economic conditions affecting job availab ility ); and Lee v. 
Washington County Bd. of Educ., 625 F.2d 1235 (5th Cir. 1980) 
(school board rebutted prima facie case by demonstrating that lack 
of blacks hired for positions in question was due solely to lack of 
black applicants).

Even with respect to the plaintiff’s disparate treatment claims, 
it  is well settled that disparities in female and minority representa­
tion in higher level jobs compared to their representation in lower 
levels is not probative of discrimination absent proof that the women 
and m inorities in lower positions were qualified for the higher 
level positions. See, e.g., Hill v. Western Electric Co., Inc., 596 F.2d 
99, 105 (4th Cir. 1979) (“The assumption that minimally qualified 
hourly rated employees were qualified for promotion to a salaried 
position is simply unfounded”), Pack v. Energy Research and De­
velopment Administration,  566 F.2d 1111, 1113 (9th Cir. 1977) 
("No evidence whatsoever was introduced to demonstrate that the 
lower-grade professional women were qualified to occupy the higher 
positions or that there elsewhere existed a pool of qualified women 
applicants”).

410



16

ment practices not obviously job-related,” id, at 707, it 
clearly should not be applied to attacks on employment 
practices or criteria without establishing a clear link to 
the practices causing the adverse impact.

The First Circuit in Latinos Unidos de Chelsea En 
Accion v. Secretary of Housing and Urban Development, 
799 F.2d 774, 786-87 (1st Cir. 1986), has pointed out 
that:

Without the threshold of a specific, facially-neutral 
procedure (or possibly, a combination of procedures, 
see Griffin v. Carlin, 755 F.2d at 1525), the dispar­
ate impact test is simply a stripped-down version of 
the discriminatory treatment test. We do not be­
lieve the Supreme Court in Griggs intended to set up 
an alternative test for finding discrimination that 
simply dropped the requirement of intent. Rather, 
the disparate impact model was created “to challenge 
those specific, facially-neutral practices that result 
in a discriminatory impact and that by their nature 
make intentional discrimination difficult or impos­
sible to prove”. If plaintiffs’ claims do not focus on 
a specific practice, it is impossible to apply the 
Griggs analysis, which envisions the employer re­
butting a prima facie case of discrimination by 
showing that the practice leading to a disparate im­
pact was justified as necessary to the employer’s 
business, Griggs, 401 U.S. at 432, 91 S.Ct. at 854. 
(Footnote omitted.)

Numerous other federal appellate courts have adopted 
this interpretation of the appropriate context for apply­
ing the disparate impact theory. See Maddnx v. Claytor, 
764 F.2d 1539, 1548 (11th Cir. 1985) (even where im­
pact analysis is applied to subjective practices, plaintiffs 
must identify particular steps in the selection process). 
See also Atonio 11, 810 F.2d at 1485; Robinson v. Polar­
oid Carp., 732 F.2d 1010, 1014 (1st Cir. 1984); Talley 
v. United States Postal Service, 720 F.2d 505, 507 (8th 
Cir. 1983), cert, denied, 466 U.S. 952 (1984); Cai'roll

411



17

V. Sears, Roebuck & Co., 708 F.2d 183, 189-90 (5th Cir. 
1983); Pope v. City of Hickory, North Carolina, 679 F.2d 
20, 22 (4th Cir. 1982).

The Fifth Circuit’s decision in Pouncy, the first case 
to address whether ihipact analysis should be applied to 
a wide-ranging attack on multiple employment practices 
or criteria, explains why the disparate impact theory is 
inappropriate for such an attack. After holding that 
the theory “applies only when an employer has instituted 
a specific procedure . . . that can be shown to have a 
causal connection to a class-based imbalance in the work 
force,” Pouncy, 668 F.2d at 800 (emphasis added), the 
appellate court explained *.

Identification by the aggrieved party of the specific 
employment practice responsible for the disparate 
impact is necessary so that the employer can respond 
by offering proof of its legitimacy . . . .  We do not 
permit a plaintiff to challenge an entire range of 
employment practices merely because the employer’s 
work force reflects a racial imbalance that might be 
causally related to any one or more of several 
practices. . . .

Id. at 801.
Only specific, facially neutral practices or criteria are 

amenable to the required showing of a causal connection. 
As recognized by Atonio /:

Were the facial neutrality threshold to disappear or 
be ignored, the distinction between disparate impact 
and disparate treatment would diminish and intent 
would become a largely discarded elemeht. Rather 
than being an irrelevant factor as envisioned, race 
(or sex, etc.) could then become an overriding factor 
in employment decisions. Employers with work forces 
disproportionate to the minority representation in 
the labor force could then face the choice of either 
hiring by quota or defending their selection proce­
dures against Title VII attack. We do not find such

412



18

a result has been mandated by Congress or through 
Supreme Court interpretation of Title VII. There­
fore, practices and policies such as a lack of well- 
defined criteria, subjective decision making, hiring 
from different sources or channels, word-of-mouth 
recruitment, and segregated housing and messing, 
which are not facially neutral, lend themselves far 
better to scrutiny for intentional discrimination.

768 F.2d at 1133. See also Griffin v. Bd. of Regents of 
Regency Universities, 795 F.2d 1281, 1288 n.14 (7th Cir.
1986) endorsing this view. As the Fifth Circuit explained 
further in Pauncy:

The disparate impact model requires proof of a 
causal connection between a challenged employment 
practice and the composition of the work force. Apti­
tude tests, height and iveight requirements, and sim­
ilar selection criteria all may be shown to affect one 
class of employees more harshly than another by con­
trolling for the impact of the employment practice 
on one class in the employer's work force so that it 
can be measured.

668 F.2d at 801 (emphasis added).
Other federal circuit courts also have recognized that 

proof of a causal connection between adverse impact and 
a particular practice or criterion is central to the dis­
parate impact theory. The First Circuit has ruled that 
“plaintiffs must show a causal connection between the 
application of the criterion in question and an alleged 
discriminatory impact on the protected class,” and that 
the causal link must be shown “independent of other 
factors.” Robinson, 732 F.2d at 1016 (emphasis added). 
In Carroll, 708 F.2d at 189, the Fifth Circuit held that 
the plaintiffs had not made a prima facie showing of 
disparate impact because they had failed “to establish the 
required causal connection between the challenged em­
ployment practice (testing) and discrimination in the 
work force.” The court further observed:

413



21

Court’s prior decisions applying the theory and with 
Congressional intent. This case affords the full Court an 
opportunity to endorse the Watson plurality’s approach, 
and thus add both clarity and reason to an area of Title 
VII law that has too long lacked both.

To rule otherwise—that is, to adopt an evidentiary rule 
that requires an employer to prove by a preponderance 
of the evidence the business necessity of both its objective 
and subjective business practices and criteria that have 
been shown to cause adverse impact—would mean that 
virtually no selection practices, except those which are 
recognized by a consensus as “best” practices, will sur­
vive under such a burden. As a result, even employers 
with the best of will and intention will be forced to hire 
by the numbers in order to avoid litigation; that is, es­
tablish quotas based on workforce and population statis­
tics.

Like the defendants in this case, the vast majority of 
large and small employers rely on selection procedures 
involving a mix of objective and subjective criteria and 
practices. Relatively few employers rely exclusively on 
validated tests to screen candidates for hire or promo­
tion. See Bureau of National Affairs, Recruiting and 
Selection Procedures, PF Survey No. 146 (May 1988) 
at 17-25. Ninety (90) percent of all employers use in­
terviews to screen applicants for employment, relying on 
a wide variety of objective and subjective criteria re­
lated to job performance. Id. at 20-21.

The reason why employers do not currently rely ex­
tensively on objective, validated selection practices is be­
cause of the complexity and cost of validation4 and the

4 The expenditure to employers to validate selection criteria is 
substantial. For example, the measurement of one simple char­
acteristic reportedly costs up to $100,000, according to a 1979 
study. Gwartney, Asher, Haworth & Haworth, Statistics, the Law 
and Title VII: An Economist's View, 54 Notre Dame L. Rev. 633,

416



22

difficulty of compliance with the Uniform Guidelines on 
Employee Selection Procedures, 29 C.F.R. §§ 1607 et scq., 
which vary from generally accepted professional prac­
tices in test development. See generally, Potter, Employee 
Selection: Legal and Practical Alternatives to Compliance 
and Litigation (1986)/’ This is true even though com­
petitive pressures to select the most productive employees 
available dictate that companies should use validated, ob­
jective procedures. In 1982, after a 3-year study, the 
National Research Council found that “most fcourtl de­
cisions have ruled against the challenged tests; no selec­
tion program seems to have survived when the Guidelines 
were applied in any detail.” National Research Council, 
Ability 7'ests: Uses, Consequences and Controversies
(1982) at 105. Moreover, a plurality of the Court in 
Watson, stated that: “It is self-evident that many jobs 
. . . require personal qualities that have never been con­
sidered amenable to standardized testing.” 108 S.Ct. at 
2791. Moreover, Justice Blackman in Wat,son recognized 
that the formal validation techniques of the Uniform

643 (1979). Moreover, in 1978, the then EEOC Vice Chairman 
Daniel E. Leach stated that the cost of a criterion related validity 
study ranged from $100,000 to $400,000. Daily Lab. Rep. at D-14 
(Dec. 5, 1978).

n In April 1976, David L. Rose, Chief of the Employment Section, 
Civil Rights Division, Department of Justice, stated in a memo­
randum to the Deputy Attorney General, that the thrust of the 
Guidelines was to: "place almost all test users in a posture of non- 
compliance; to give great discretion to enforcement personnel to 
determine who would be prosecuted; and to set aside objective selec­
tion procedures in favor of numerical hiring.” 122 Cong. Rec. 22950 
(daily ed. July, 1976) (emphasis supplied). See also Ballcw, Courts, 
Psychologists, and the EEOC's Uniform Guidelines: An Analysis 
of Recent Trends Affecting Testing as a Means of Employee. Selec­
tion, 36 Emory L.J. 203, 212-217 (1987), which documents the 
declining deference that this and other courts give to the technical 
validation requirements of the Uniform Guidelines and the differ­
ences between accepted professional practices and the Guidelines’ 
requirements.

417



2 3

Guidelines “may sometimes not be effective in measuring 
the job-relatedness of subjective-selection processes.” 108 
S.Ct. at 2795.

In Watson, a plurality of the Court stated that the 
“business necessity” or “job relatedness” defense under 
Griggs did not shift the burden of persuasion to the de­
fendant. 108 S.Ct. at 2790. It said that “the ultimate 
burden of proving that discrimination against a protected 
group has been caused by a specific employment practice 
remains with the plaintiff at all times. Id. Even with re­
spect to defending standardized or objective tests, the 
plurality held that employers are not required to intro­
duce formal validation studies showing that particular 
criteria predict actual on-the-job performance. Id. Once 
“the defendant has met its burden of producing evidence* 
that its employment practices are based on legitimate 
business reasons, the plaintiff must ‘show that other tests 
or selection devices, without a similarly undesirable facial 
effect, would also serve the employer’s interest in efficient 
and trustworthy workmanship.’ ” Id. (citation omitted)." 
This burden of proof scheme is consistent with the prin­
ciples articulated by this Court in Furnco and Burdine 
and the legislative history of Title VII.

Most if not all employment-related decisions are based 
on the employer’s desire to choose the best person for the 
job. See Furnco, 438 U.S. at 577 (“ [W]e know from 
our experience that more often than not people do not 
act in a totally arbitrary manner, without any underly-

fl In meeting the pretext burden, the plurality pointed out that: 
Factors such as the cost or other burdens of proposed alterna­
tive selection devices are relevant in determining whether they 
would be equally as effective as the challenged practic in serv­
ing the employer’s legitim ate business goals. The same factors 
would also be relevant in determining whether the challenged 
practice has operated as the functional equivalent of a pretext 
for discriminatory treatment.

108 S.Ct. at 2690 (citation om itted).

418



24

ing reasons, especially in a business setting.”) The pur­
pose of Title VII is to ensure that those decisions are 
made without consideration of illegal and inappropriate 
factors such as race, sex, and national origin. Griggs, 
401 U.S. at 431. However, as the Court in United Steel- 
woi'kers v. Weber 443 U.S. 193, 206 (1979) recognized, 
the legislative history to Title VII shows that the statute 
would not have been enacted without recognition of and 
preservation of, managerial discretion in employment 
decisions:

Title VII could not have been enacted into law with­
out substantial support from legislators in both 
Houses who traditionally resisted federal regulation 
of private business. Those legislators demanded as a 
price for their support that “management preroga­
tives . . .  be left undisturbed to the greatest extent 
possible. H.R. Rep. No. 914, 88th Cong., 1st Sess., 
pt. 2, p 29 (1963), U.S. Code Cong. & Admin. News 
1964, p. 2391.

In Furnco, the Court held that Title VII does not re­
quire businesses to adopt “best” hiring procedures or re­
quire an employer to “pursue [1 the course which would 
both enable [it] to achieve fits] own busines goal and 
allow [it] to consider the most employment applications.” 
438 U.S. at 577 (emphasis in original). Importantly, with 
respect to the lower court’s conclusion in Furnco that 
“different practices would have enabled the employer at 
least consider, and perhaps hire, more minority employ­
ees,” this Court concluded that “courts may not impose 
such a remedy on an employer at least until a violation 
of Title VII has been proved.” Id. at 578. It also em­
phasized that: “Courts are generally less competent than 
employers to restructure business practices, and unless 
mandated to do so by Congress they should not attempt 
it.” Id.

In sum, under Furnco, an employer is not required to 
adopt “best” hiring procedures that would permit it “to

419



2 5

at least consider . . . the most minority employees.” Id. 
Indeed, a unanimous Court expressly held in Burdine 
that Title VII “was not intended to ‘diminish traditional 
management prerogatives’ ” and that an “employer has 
discretion to choose among equally qualified candidates.” 
450 U.S. at 259 (citation omitted).

However, a disparate impact burden of proof scheme 
that imposes on employers a burden of persuasion to 
prove the business necessity of its practices in effect 
compels employers to use “best” hiring practices or, al­
ternatively to establish quotas. As recognized by the 
Court’s plurality decision in Watson:

If quotas and preferential treatment become the only 
cost-effective means of avoiding expensive litigation 
and potentially catastrophic liability, such measures 
will be widely adopted. The prudent employer will 
be careful to ensure that its programs are discussed 
in euphemistic terms, but will be equally careful to 
ensure that the quotas are met. Allowing the evolu­
tion of disparate impact analysis to lead to this re­
sult would be contrary to Congress’ expressed intent, 
and it should not be the effect of our decision today.

108 S.Ct. at 2788.7
As discussed above, even validated objective selection 

procedures have been difficult to justify under Title VII

T One commentor has stated that the purpose of the disparate 
impact theory “is better served by placing a moderate burden upon 
defendants rather than a heavy burden that would in effect, prevent 
hidden discrimination only by requiring reverse discrimination”. 
Eutherglen, “Disparate Impact Under Title V II : An Objective
Theory of Discrimination,” 73 Va. L. Rev. 1297, 1315 (1987). In 
addition, the commentator points out that:

A heavy burden serves the more controversial purpose of pro­
moting equal opportunity directly by discouraging employment 
practices with adverse impact but only at the risk of con­
tradicting the prohibition against required preferences in sec­
tion 703( j ).

Id. at 1315-16.

420



26

because lower courts for the most part have required 
employers to show that the practice is absolutely essential 
or necessary to the operation of the business” This has 
occurred even though no alternative to tests is currently 
available that is equally informative, and as technically 
and economically viable, with respect to assessing the 
capabilities of individuals. Ability Tests at 143-44. To 
impose a burden of persuasion on employers to justify 
validated as well as less precise measures literally con­
demns all employment selection practices and criteria 
when a company has an internal workforce imbalance or 
a workforce that is different from the general population. 
Indeed, three Justices of this Court in Teal recognized 
that “there are few if any tests . . . that accurately re­
flect the skills of every individual candidate. Teal, 457 
U.S. at 463. (Powell, J., dissenting).

A requirement that the employer prove the business 
necessity of its practices also misapprehends the orderly 
burden of proof scheme established in Griggs and Albe­
marle. It merges the rebuttal and pretext stages, because 
the employer in proving that its job related criteria are 
essential to its business must also explain why it did not 
rely on other criteria and practices as well. For example, 
Section 3B of the Uniform Guidelines provides that:

[Wjhenever a validity study is called for by these 
guidelines, the user should include, as part of the 
validity study, an investigation of suitable alterna- *

* See, e.g., Kirby  v. Colony Furntnrc Co., Inc., 613 F.2d 696 (8th 
Cir. 1980); United States v. Bethlehem Steel Corp., 446 F.2<1 652. 
662 (2d Cir. 1971). Compare EEOC v. Kimbrough Investment Co., 
703 F.2d 98, 100 (5th Cir. 1983) ( “If the plaintiff succeeds in this 
showing [of a prima facie case of discrimination], the focus of 
attention sh ifts to the employer to persuade the court of the ex­
istence of a ‘legitim ate business reason’ by a preponderance of the 
evidence.” ) ; and Chrisner v. Complete Auto Transit, Inc., 645 F.2d 
1251 (6th Cir. 1981) (The practice must substantially promote the 
proficient operation of the business). See also Schlei and Grossman, 
Employment Discrimination Law at 1329-1330 n.148 (1983).

421



2 7

tive selection procedures and suitable alternative 
methods of using the selection procedure which have 
as little adverse impact as possible, to determine the 
appropriateness of using or validating them in ac­
cord with these guidelines.

An employer may conduct this search voluntarily, but it 
is inconsistent under Albemarle to require the employer 
prove this as part of its rebuttal burden.0 Indeed, the 
Court in Albemarle clearly stated that once the defend­
ant meets its burden of proof, the complaining party can 
show the discriminatory pretext of the criteria or prac­
tices used by the employer by demonstrating “that other 
tests or selection devices, without a similarly undesirable 
racial effect, would also serve the employer’s legitimate 
interest in ‘efficient and truthworthy workmanship.’ ’’ 
422 U.S. at 425.

In sum, shifting the burden of proof to the defendant 
to prove the job relatedness of its selection procedures 
unnecessarily interferes with the rights of employers to 
determine its selection practices. Alternatively, permit- 
ing the defendant to state or produce its legitimate busi­
ness reasons for relying on particular criteria or prac­
tices will not force employers to abandon nondiscrimina- 
tory employment practices or to engage in illegal quotas, 
but instead allows plaintiffs to show that the criteria or 
practice is a pretext for discrimination in accordance 
with Albemarle.

0 In addition, several lower courts have required the defendant to 
assume this burden as a part of showing the necessity of its selec­
tion criteria. See, e.ej., Kirby,  613 F.2d at 703-04, lUalte v. Ci ty  of 
Lns Anpeles, 595 F.2d 1367, 1376 (9th Cir. 1979) cert, denied, 446 
U.S. 928 (1980); and Pettway v. American Cast Iron Pipe Co., 
494 F.2d 211, 244 n.87 (5th Cir. 1974), cert, denied, 467 U.S. 1243 
(1984).

422



28

C O N C L U S IO N

F o r  th e  rea so n s  s ta te d  h ere in , E E  A C  r e sp e c tfu lly  su b ­
m its  th a t  th e  d ec is io n  o f  th e  N in th  C ir c u it  sh ould  be 
rev ersed .

R e sp e c tfu lly  su b m itted ,

Robert E. W illiams 
Douglas S. M cDowell 
Edward E. Potter, P.C.* 
M cGuiness & W illiams 
1015 Fifteenth Street, N.W. 
Suite 1200
Washington, D.C. 20005 
(202) 789-8600

Attorneys for the Amiens Curiae 
Equal Employment Advisory 
Council

* Counsel of Record
September 10,1988

423





No. 87-1387

In the

Supreme Cffourt of life litniteii States
October T erm , 1988

Wards  Cove Packing Com pany , Inc ., 
Castle & Cooke, Inc .,

v.

Frank A ntonio , et al..

P etitioners,

R esponden ts.

On Writ of Certiorari to the United States 
Court of Appeals for the Ninth Circuit

BRIEF FOR THE LAWYERS’ COMMITTEE FOR 
CIVIL RIGHTS UNDER LAW AS 

AMICUS CURIAE IN SUPPORT OF THE RESPONDENTS

Conrad Harper 
Stuart J. Land 

Co-Chairmen 
Norman Redlich 

Trustee
Richard T. Seymour 
James C. G ray, J r.

Lawyers’ Committee for 
Civil R ights Under Law 

1400 “Eye" Street, N.W.
Suite 400
Washington, D.C. 20005 
(202) 371-1212

N icholas deB. Katzf.nbacii*
Alan E. Kraus 
David Arciszf.wski

Rikf.r, Danzig, Scherer. Hyland 
& Perretti 

Headquarters Plaza 
One Speedwell Avenue 
Morristown, New Jersey 07960-1981 
(201)538-0800

A tto rn e y s  f o r  A m icu s  C uriae  
L a w yers ' C o m m itte e  f o r  C iv il R igh ts  

U nder Law

November 4, 1988 'Counsel of Record

425



T A B L E  O F  C O N T E N T S

IN T E R E S T  O F  A M IC U S  C U R I A E ...............................  1
Q U E S T IO N S  P R E S E N T E D ...........................................  2
S U M M A R Y  O F  A R G U M E N T .......................................  3
A R G U M E N T ......................................................................  6

I. IN  A  D IS P A R A T E  IM P A C T  C A S E , A S  T H E
N IN T H  C IR C U IT  P R O P E R L Y  H E L D , O N C E  
T H E  P L A IN T IF F  H A S  M A D E  A P R IM A  
F A C IE  S H O W IN G , T H E  B U R D E N  O F  P E R ­
S U A S IO N  S H IF T S  T O  T H E  E M P L O Y E R  T O  
R E B U T  T H A T  P R IM A  F A C IE  S H O W IN G  . 6

A. The Differences between Individual Disparate 
Treatment Cases and Classwide Disparate 
Impact Cases Warrant a Different Allocation
of the Burdens of P ro o f.....................................  7

B. This Court Has Uniformly Held that the Bur­
den of Proving Business Necessity Shifts to 
the Employer Following a Prima Facie Show­
ing in a Disparate Impact Case......................... 13

C. Sound Reasons of Public Policy and Practical­
ity Warrant the Shifting of the Burden of Per­
suasion to the Employer in Disparate Impact 
C ase s....................................................................  15

II. S P E C IF IC  C A U S A T IO N  IS  N O T  T H E
A P P R O P R IA T E  S T A N D A R D  IN  A D IS P A ­
R A T E  IM P A C T  C A S E .........................................  17

II I .  T H E N IN T H  C IR C U IT  C O U R T O F  
A P P E A L S  C O R R E C T L Y  H E L D  T H A T  
R E S P O N D E N T S ’ S T A T IS T IC S  M A D E  O U T  
A P R IM A  F A C IE  C A S E  O F  D IS P A R A T E  
IM P A C T ....................................................................  25

C O N C L U S IO N ..................................................................  29

Page

426



11

A lb e rm a r le  P a p e r  Co. v. M o o d y , 422 U S. 405 
(1975)................................................................................ p a ss im

A n to n io  v. W a rd s  C o ve  P a ck in g  C o., 827 F.2d 439 
(9th C ir. 1987) .........................................................  6

B a rn ett v. W. 7. G ran t C o., 518 F.2d 543 
(4th C ir. 1975) .........................................................  19

B a ze m o re  v. F rid a y , 478 U.S. 385 (1 9 8 6 ) .................  20

B elch er v. B a sse tt F u rn itu re  In d u s tr ie s  Inc., 588 
F.2d 904 (4th Cir. 1978) ........................................  15

B o n illa  v. O a k la n d  S c a v en g er  C o., 697
F.2d 1297 (9th Cir. 1982), cert, d en ied , 467
U.S. 1251 (1984) ..................................................... 18

B row n  v. G a sto n  C o u n ty  D yein g  M a ch in e  Co., 457 
F.2d 1377 (4th C ir.), cert, den ied , 409 U.S. 982
(1972).........................................................................  19

B urns v. T h io k o l C h e m ic a l C orp ., 483 F.2d 300 
(5th Cir. 1973) .........................................................  26

C o n n ec ticu t v. T ea l, 457 U.S. 440 (1982) .................  p a ss im

D a v is  v. C ity  o f  D a lla s , 483 F. Supp. 54 
(N .D . Tex 1979)....................................................... 27

D o m in g o  v. N e w  E n g la n d  F ish C o., 445 F. Supp.
421 (W .D. Wash. 1977), a f f d ,  727 F.2d 1429
m o d if ie d , 742 F.2d 520 (9th Cir. 1 9 84 )................  19

D o th a rd  v. R a w lin so n , 433 U.S. 321 (1 9 7 7 )............  11,12,14

E E O C  v. In la n d  M a rin e  In d u str ies , 729 F.2d 1229 
(9th C ir.), cert, den ied , 469 U.S. 855 ( 19 8 4 )........  18

T A B L E  O F  A U T H O R I T I E S

Page

427



Ill

F alcon  v. G en era l T e lep h o n e  Co. o f  th e  S o u th w e s t,
626 F.2d 369 (5th C ir. 1980), v a ca te d , 450 U S.
1036, a f f d  on  reh ea rin g , 647 F.2d 633
(5th C ir. 1 9 8 1 ).......................................................... 27

F ra n k s  v. B o w m a n  T ra n sp o r ta tio n  C o.,
424 U.S. 747 (1 9 7 6 )................................................  6,28

F urnco  C o n s tru c tio n  Co. v. W aters , 438 U.S. 567 
(1978)..........................................................................  9,11,15

G ibson  v. L oca l 40, 543 F.2d 1259 (9th Cir. 1976) . . .  18

G reen  v. U S X  C o rp ., 843 F.2d 1511 (3d Cir. 1988), 
p e t i t io n  f o r  cert, f i l e d ,  57 U .S.L.W . 3123 (U .S.
July 23, 1988) (No. 88-141) ...................................  22,24

G rig g s  v. D u k e  P o w er  C o., 401 U.S. 424 (1971) . . .  p a s s im

H a z e lw o o d  S c h o o l D is tr ic t  v. U n ite d  S ta te s ,  433 
U.S. 299 (1977).......................................................... 2,27

In te rn a tio n a l B hd. o f  T e a m s te rs  v. U n ited  S ta te s ,
431 U.S. 324 (1 9 7 7 ) ..........................................   p a s s im

L ew is  v. B lo o m sb u rg  M ills , Inc., 773 F .2d 561 
(4th C ir. 1 9 8 5 ).......................................................... I

L illy  v. H a rris -L 'ea ter  S u p e r m a r k e t, 842 F.2d 1496 
(4th C ir. 1 9 8 8 ).......................................................... 19

N e w  Y ork  C ity  T ra n s it A u th o r ity  v. B eazer. 440 
U.S. 568 (1979).......................*.................................  25

M cD o n n e ll D o u g la s  C orp . i\ G reen, 41 I U.S. 792
(1973)..........................................................................  9,11

P a yn e  v. T ra v en o l L a b o ra to r ie s , Inc., 673 F.2d 798 
(5th C ir.), cert, d en ied , 459 U.S. 1038 ( 1982) . . .  . 2,8

T A B L E  O F  A U T H O R I T I E S — C on tin u ed

Page

428



IV

P ou n cy  v. P ru d e n tia l In su ran ce C o., 668 F.2d 795
(5th C ir. 1982) .........................................................  17

R o w e v. G en era l M o to rs  C orp ., 457 F.2d 348 
(5th C ir. 1972) .........................................................  18

S e g a r  v. S m ith ,  738 F.2d 1249 (D .C. Cir. 1984), 
cert, d en ied , 471 U.S. 111 5 ( 19 8 5 ) ........................  24

S le d g e  v. J. P. S te v e n s  c? C o., 585 F.2d 625 (4th Cir.
1978), cert, d en ied , 440 U.S. 981 (1 9 7 9 )..............  2

T a rver  v. C ity  o f  H o u sto n , 22 FP D  1130, 689 (S.D.
Tex. 1980) . ................................................................ 21

T e x a s  D e p a r tm e n t o f  C o m m u n ity  A ffa ir s  v.
B ur d in e , 450 U.S. 248 (1 9 8 1 )................................  p a ss im

U n ite d  S ta te s  v. G eo rg ia  P ow er C o., 474 F.2d 906 
(5th C ir. 1973) .........................................................  19

U n ite d  S ta te s  v. C o u n ty  o f  F a irfa x , 629 F.2d 932 
(4th Cir. 1980) . . .  .’ ................................................. 8,16

V uyan ich  v. R e p u b lic  N a tio n a l B ank, 521 F.Supp.
656 (N .D . Tex. 1981), rev  ’d o n  o th e r  g rou n ds, 723
F.2d I 195 (5th C ir.), cert, den ied . 469 U.S. 1073
(1984).........................................................................  8

W atson  v. F ort W orth  B ank a n d  T ru st, 487 U.S.
108 S.Ct. 2777 (1988) ............................................  p a ss im

W ilk in s  v. U n iv e rs ity  o f  H o u sto n , 654 F.2d 388 (5th 
Cir. 1981), v a ca te d , 459 U.S. 809 (1982), a f f d  on 
reh ea rin g , 695 F.2d 134 (5th Cir. 1983) 20

St a t u t e s , R egulations a n d  Rules

Civil Rights Act of 1964, Title V II, as amended, 42 
U .S.C. §2000e-2....................................................... 22

T A B L E  O F  A U T H O R I T I E S — C ontinued

Page

429



V

Uniform Guidelines on Eimployee Selection
Procedures (1978), 29 C .F .R . §1607 ........................  16

29 C .F .R . §1607.1 6 Q ...............................................  23
29 C .F .R . §1 6 1 5 ........................................................ 16

Fed. R. Civ. P. 8 ............................................................ 12

C ongressional  H istory:

S.Rep. No. 415, 92nd Congress, 1st Sess. 5 (1971) 22

M iscellaneous:

110 Daily Labor Rep. (B N A ) A-3 (June 7, 1985) 17

Friedman and Williams, C u rren t U se o f  T es ts  fo r  
E m p lo y m e n t , in 2 Ability Testing; Uses, C onse­
quences, and Controversies ( 1982) ......................... 16

Report of the Director, Administrative Office of the 
U.S. Courts (1 9 7 7 )...................................................  29

Rose, S u b je c tiv e  E m p lo y m e n t P ra c tices: D oes the  
D isc r im in a to ry  Im p a c t A n a ly s is  A p p ly ? . 25 San 
Diego L. Rev. 63 (1 9 8 8 )...........................................  17

B. Schlei and P. Grossman, E m p lo y m e n t D isc r im i­
n a tio n  L aw  (2d cd. 1 983).........................................  I 1,26

T A B L E  O F  A U T H O R I T I E S — ('o n lin u c d

Page

430



In the

§itpm tie Cffourt of fife Jlnttei* §tatcs
O ctober T erm , 1988

No. 87-1387

W ards  Cove Packing Co m pa n y . Inc ., 
C astle & Cooke, Inc .,

v.
P e tit io n e r s ,

Frank  A n tonio , e t a!..

R e sp o n d e n ts .

On Writ of Certiorari to the United States 
Court of Appeals for the Ninth Circuit

BRIEF FOR THE LAWYERS’ COMMITTEE FOR 
CIVIL RIGHTS UNDER LAW AS 

AMICUS CURIAE IN SUPPORT OF THE RESPONDENTS

INTEREST OF AMICUS CURIAE

The Lawyers’ Committee for Civil Rights Under Law submits 
this brief as amicus curiae urging affirmance.1

The Lawyers’ Committee is a nonprofit organization estab­
lished in 1963 at the request of the President of the United 
States to involve leading members of the bar throughout the 
country in the national effort to insure civil rights to all Ameri­
cans. It has represented and assisted other lawyers in repre­
senting numerous persons in administrative proceedings and 
lawsuits under Title V II. E.g., Lewis v. Bloomsburg Mills, Inc.,

' Pursuant to Rule 36.2, the Lawyers' Committee is Tiling herewith written 
consents of the parties to the submission of this brief as a m i c u s  c u r ia e .

431



2

773 F.2d 561 (4lh Cir. 1985); Payne v. Travenol Laboratories, 
Inc., 673 F.2d 798 (5th C ir.), cert, denied, 459 U.S. 1038 
(1982); Sledge v. J. P. Stevens Co., 585 F.2d 625 (4th Cir. 
1978), cert, denied, 440 U.S. 981 (1979). The Lawyers' 
Committee has also represented parties and participated as an 
amicus in Title V II  cases before this Court. E  g., Watson v. Fort 
Worth Bank & Trust, 487 U.S. , 108 S.Ct. 2777 (1988); 
Connecticut v. Teal, 457 U.S. 440 (1982); Hazelwood School 
District v. United States, 433 U.S. 299 (1977).

The questions presented by this case raise important and 
recurring issues in Title V II law. In particular, the allocation of 
the burdens of proof in disparate impact cases and the nature of 
plaintiffs' burden in presenting a prima facie  case of disparate 
impact are issues that affect virtually every class action brought 
under Title V II. This Court's decision will undoubtedly have 
significant implications on present and future Title V II  suits in 
which the Lawyers’ Committee participates. In addition, the 
Lawyers’ Committee has a longstanding interest in persuading 
the Court to adopt principles that will result in the sound admin­
istration of the discrimination laws, so that findings of liability 
will be obtainable by persons with legitimate claims and limited 
resources. In this case, the Lawyers' Committee also brings to 
the Court the benefit of its actual experience in marshalling the 
facts in complex employment discrimination class actions, and 
discusses in practical terms the flaws in the approach taken by 
petitioners and the United States as amicus curiae to disparate 
impact cases.

Q U E S T IO N S  P R E S E N T E D
1. Whether, in a disparate impact case under Title V II, 

plaintiffs’ prima facie  showing shifts a burden of persuasion to 
the employer to prove the business necessity of the personnel 
practices at issue or merely a burden of production?

2. Whether, in order to make a prima facie  showing in a 
disparate impact case, plaintiffs are required to identify specific 
employment practices at issue and prove a specific causal link 
between each practice and an identifiable disparate impact?

3. Whether the statistical showing made by plaintiffs in this 
case, which demonstrated a significant racial disparity between

432



3

petitioners’ cannery workers and non-canncrv workers, was 
sufficient to make out a prime facie case?

SU M M A R Y  O F A R G U M EN T

1. Petitioners contend that the Ninth Circuit Court of 
Appeals erred by placing the burden on them of disproving 
plaintiffs’ statistical showing of disparate impact before, in peti­
tioners’ view, plaintiffs had even made out a prime facie case. 
(Pet. Brief at 37-47.) Petitioners have misconstrued the Ninth 
Circuit’s holding: The Ninth Circuit held only that, if employers 
attempt to dispute plaintiffs’ statistical showing with statistics of 
their own, those statistics must be probative and relevant. Here, 
petitioners’ labor force statistics relied on job criteria and 
qualifications that were not shown to be actually used by peti­
tioners. The Ninth Circuit properly held that those labor force 
statistics were not sufficiently probative to dispute plaintiffs’ 
prime facie  showing of disparate impact.

The United States as amicus curiae goes further than peti­
tioners. The United States argues that, on this appeal, this Court 
should discard the order of proof in disparate impact cases that 
has been controlling since Griggs Duke Power Co., 401 IJ.S. 
424 (1971) and, for the first time, expressly hold that the 
employer's burden in rebutting a prima facie showing of dispa­
rate impact is one of mere production, rather than persuasion. In 
short, the United States urges this Court to apply the individual 
disparate treatment order of proof, as set forth in / exes Depart­
ment o f Community Affairs v. Burdine, 450 U.S. 248 (1981), to 
disparate impact cases.

The argument of the United States is, we submit, fatally 
flawed. It ignores the critical differences between individual dis­
parate treatment cases and disparate impact cases that make 
Burdine inappropriate and inapplicable in the disparate impact 
context. In individual disparate treatment cases, where the 
employer’s motivation is the ultimate issue, the plaintiff s prima 
facie case eliminates only the “ most common nondis- 
criminatory reasons for differences in treatment. It is not until 
the employer responds by articulating the “ real reason for its 
different treatment of plaintiff that the plaintiff ultimately



4

proves discrimination by eliminating the so-called real reason as 
well.

In disparate impact cases, in contrast, the plaintiffs face a 
much heavier prima facie  burden. There, plaintiffs, in order to 
make a prima facie showing, must demonstrate that the employ­
er’s facially neutral practices have caused a significant disparate 
impact upon minorities. Because proof of discriminatory motive 
is not necessary in disparate impact cases, a prima facie show­
ing, unless rebutted by the employer by proof of the business 
necessity of the practices at issue, mandates a holding in favor of 
plaintiffs. Obviously, employers have far better access to evi­
dence concerning the business necessity of their employment 
practices. There is nothing unfair about asking them to carry the 
burden of proving that affirmative defense.

In sum, the far heavier prima Jacie  burden imposed on dispa­
rate impact plaintiffs amply justifies a heavier rebuttal burden 
on defendant employers. This Court has uniformly applied that 
very order of proof in past disparate impact decisions and there 
is, we submit, no reason to alter that rule now.

2. Petitioners further contend that plaintiffs should not be 
allowed to challenge the cumulative disparate impact of petition­
ers’ personnel system. Petitioners assert that disparate impact 
plaintiffs must be required to identify specific practices at issue 
and demonstrate in detail the specific disparate impact caused 
by each practice. Otherwise, petitioners contend, the employer 
faces the unfair burden of defending the business necessity of 
every aspect of his personnel system. That issue is not fairly 
presented on this appeal.

This is not a case where the plaintiffs launched a “ shotgun” 
attack on an employer’s entire personnel system. Here, as the 
Ninth Circuit found, plaintiffs proved the disparate impact of 
six specific employment practices. Those six practices— the use 
of subjective criteria, nepotism, separate hiring channels, word- 
of-mouth recruiting, race labelling and segregated facili­
ties— are well-established causes of disparate impact. Thus, 
plaintiffs amply carried their prima facie  burden of proving that 
the practices at issue caused the disparate impact shown.

434



5

Moreover, there is nothing unfair in shifting the burden of 
proof to an employer once disparate impact of his personnel sys­
tem has been demonstrated. It is well-established that disparate 
impact creates a presumption of discrimination because a fair, 
nondiscriminatory employment system should ordinarily pro­
duce a racially balanced workforce. Thus, to hold that an 
employer with a demonstrably unbalanced workforce is immune 
from Title V II  challenge simply because the plaintiffs are unable 
to prove which specific practices caused what part of the dispar­
ity would, in all likelihood, allow significant violations of Title 
V II  to go unremedied.

Furthermore, in many instances, the plaintiffs are unable to 
prove specific causation because either the evidence, such as 
detailed employment records, is unavailable or because plaintiffs 
lack sufficient expertise, resources or access to the place of 
employment. There is nothing unfair in asking the 
employer— who controls the maintenance of employment 
records, who is, of course, most expert in his own employment 
system and job criteria, and who enjoys complete access to the 
place of employment— to carry the burden of explaining away a 
demonstrable, significant racial disparity in his workforce.

3. Finally, petitioners contend that plaintiffs’ statistics show­
ing a significant disparity between the racial composition of peti­
tioners' lower paid cannery workers and that of their higher paid 
non-cannery workers were not sufficient to carry plaintiffs 
pritna facie  burden. Petitioners contend that only comparisons 
to the qualified labor pool are relevant.

Petitioners miss the point. This case is about barriers to 
opportunity. The plaintiffs’ complaint here is that petitioners 
employment practices unfairly denied them even the opportunity 
to compete for the higher paying non-cannery jobs. As this 
Court long ago recognized in Griggs, a core objective of Title 
V II  “ was to achieve equality of employment opportunities and 
remove barriers that have operated in the past to favor an 
identifiable group of white employees over other employees. 
401 U.S. at 429-30. In short, plaintiffs’ statistics amply demon­
strate discrimination in job opportunities; petitioners labor force 
statistics are not even relevant to that claim.

435



6

ARGUMENT
I. IN A DISPARATE IMPACT CASE, AS THE NINTH CIR­

CUIT PROPERLY HELD, ONCE THE PLAINTIFF HAS 
MADE A PRIMA FACIE SHOWING, THE BURDEN OF 
PERSUASION SHIFTS TO THE EMPLOYER TO 
REBUT THAT PRIMA FACIE SHOWING.

Petitioners, supported by various amici including the United 
States, contend that in this case the Ninth Circuit Court of 
Appeals “ fashionfed] a new allocation of the burdens of proof in 
a [disparate] impact case, drastically lowering respondents’ and 
raising petitioners’.” (Pet. Brief at 37.) In summary, petitioners 
contend that the Ninth Circuit erred by holding that petitioners 
had the burden of disproving the plaintiffs’ prima facie  showing 
of disparate impact before plaintiffs had ever made out a prima 
facie  case. (Pet. Brief at 37-47.) Petitioners argue that the Ninth 
Circuit improperly placed on them the burden of proving, rather 
than merely articulating, flaws in plaintiffs’ prima facie  showing 
of disparate impact. (Id .)

Petitioners miss the point of the Ninth C ircuit’s holding. The 
Ninth Circuit did not place the burden of disproving plaintiffs’ 
statistics on petitioners; the Ninth Circuit held only that peti­
tioners’ objections to plaintiffs’ prima facie  showing must be rel­
evant and probative. Because petitioners’ objections were neither 
relevant nor probative, they failed to preclude a finding that 
plaintiffs had made out a prima facie  case. There is nothing new 
or novel about that holding.2

The United States goes even further than petitioners. The 
United States contends that the Ninth Circuit erred by holding 1

1 For example, the Ninth Circuit held that petitioners’ labor force statistics 
were not relevant to undermine plaintiffs’ workforce imbalance statistics 
because petitioners had failed to demonstrate that the qualifications criteria 
that underlay their labor force statistics were in fact used by petitioners. 
Indeed, plaintiffs presented evidence that those criteria were not used by peti­
tioners. A n t o n i o  v. W a r d s  C o v e  P a c k i n g  C o . ,  827 F.2d 439, 446 (9th Cir. 
1987). That holding is nothing more than a straightforward application of this 
Court’s holding in F r a n k s  v. B o w m a n  T r a n s p o r t a t i o n  C o . ,  424 U.S. 747, 773 
n. 32 (1976) that only non-discriminatory standards “actually applied" are 
relevant.

436



7

that, once plaintiffs in a Title V II case have made a prima facie 
showing of disparate impact, the burden of persuasion of 
rebutting that prima facie  case shifts to the defendant employer. 
Rather, the United States contends, the employer s burden 
should be only to articulate a rebuttal to plaintiffs prima facie 
case. In short, the United States argues that the allocation of 
burdens of proof in a disparate impact class action should be 
exactly the same as in an individual disparate treatment case. 
See Texas Department o f Community Affairs v. Burdine, 450 
U.S. 248, 252-53 (1981); (Brief for the United States as Amicus 
Curiae at 25-28.) Because of the far reaching implications of the 
United States’ argument, we address it in detail below.

Simply put, the United Stales is wrong. First, individual dis­
parate treatment cases and class action disparate impact cases 
are fundamentally different in theory and in practice and those 
differences amply justify a different allocation of burdens of 
proof. Second, the decisions of this Court applying disparate 
impact theory have, without exception, held that, once a prima 
facie  case of disparate impact is established, the burden of per­
suasion shifts to the employer to rebut that prima facie case. 
Thus, the Ninth C ircuit’s ruling is not only not a “ new alloca­
tion’’ of the burdens of proof in impact cases, it is the only ruling 
the Ninth Circuit could have made consistent with this Court's 
prior decisions. For this Court to adopt the United States' 
approach, it would have to overrule a long and unbroken line of 
authority set forth in this Court's own decisions. And third, 
there are sound public policy and practical reasons why in dispa­
rate impact cases the burden of persuasion should shift to the 
employer once a prima facie showing is made.

A. The D iffe ren ce s  betw een  In d iv id u a l D isp a ra te  T rea tm en t 
C a se s  a n d  C la s s -w id e  D isp a ra te  Im p a ct C a ses  M arran t a 
D iffe re n t A llo c a tio n  o f  th e  B urdens o f  Proof.

Title V II  cases generally fall into two categories: disparate 
treatment and disparate impact/ Disparate treatment cases •

•1 A particular case can utilize either or both methods of proof. As this C ourt 
has recognized, “(c|ilhcr theory may, of course, be applied to a particular set

437



8

seek to remedy the most obvious evil Title V II  was designed to 
eradicate, namely situations where an employer intentionally 
treats some people less favorably because of their race, color, 
religion, sex or national origin. In a disparate treatment case, 
“ [p]roof of discriminatory motive is critical, although it can in 
some situations be inferred from the mere fact of differences in 
treatment.” Teamsters, 431 U.S. at 335 n. 15.

Disparate impact claims, in contrast, focus on employment 
practices and procedures that are facially neutral in their treat-

of facts.” I n t e r n a t i o n a l  B h d .  o f  T e a m s t e r s  v. U n i t e d  S t a t e s ,  431 U.S. 324, 336 
n. 15 (1977).

Pattern and practice disparate treatment cases present yet a third category, 
combining elements of both disparate impact and individual disparate treat­
ment cases. Pattern and practice disparate treatment cases are, at bottom, class 
actions predicated on allegations of disparate treatment. Like an individual dis­
parate treatment case, the crux of a pattern and practice disparate treatment 
case is an employer’s intentional and less favorable treatment of minority 
employees. Like a disparate impact case, however, the proof of discrimination 
in a pattern and practice disparate treatment case is generally statistical. In 
such a case, the mere articulation of a defense—such as the employer's asser­
tion that it hires and promotes the “best-qualified" candidates—is insufficient. 
T e a m s t e r s ,  431 U.S. at 342 n. 24; P a y n e  v. T r a v e n o l  L a b o r a t o r i e s .  673 F.2d at 
818. Similarly, the mere articulation of a potential flaw in plaintiffs’ statistics 
is insufficent; the defendant has the burden of persuasion that the problems it 
cites are real, and that they explain so much of the disparities proven by 
plaintiffs that their probative value is destroyed. P a y n e ,  673 F.2d at 822; 
U n i t e d  S t a t e s  v. C o u n t y  o f  F a i r f a x ,  629 F.2d 932, 940 (4th Cir. 1980), cer t ,  

d e n . ,  449 U.S. 1078 (1981).
There is nothing unusual or unjust in these rulings. No matter how massive a 

plaintiffs statistical showing may be, it can never cover every possible factor. 
Universal analyses come only at infinite expense. To allow a probative statis­
tical showing to be defeated by mere articulation or speculation that other 
factors or analyses might lead to a different result, without imposing any bur­
den of persuasion on the defendant, would result in the defeat of every statis­
tical showing of disparate treatment. It is for those reasons that Judge 
Higginbotham stated in V u y a n ic h  v. R e p u b l i c  N a t i o n a l  B a n k  o f  D a l l a s ,  521 F. 
Supp. 656, 661 (N.D. Tex., 1981), r e v ' d  o n  o t h e r  g r o u n d s ,  723 F.2d I 195 (5th 
Cir ), c e r t ,  d e n . ,  469 U.S. 1073 (1984), that “|i|n a complex class action, 
utilizing statistical proof and counterproof, the value of the B u r d in e  
sequence—to highlight the issues in context is about as relevant as a minuet 
is to a thermonuclear battle.”

438



9

ment of different groups but nevertheless in fact fall more 
harshly on one or more groups.' Proof of discriminatory motive 
is not necessary in disparate impact cases. Indeed, “good intent 
or absence of discriminatory intent docs not redeem employment 
procedures or testing mechanisms that operate as ‘built-in 
headwinds’ for minority groups.’’ Griggs v. Duke Power Co., 401 
U.S. 424, 432 (1971).

As a unanimous Court recognized in Griggs, disparate impact 
analysis promotes Congress' intent in Title V II to outlaw not 
only overt, intentional discrimination but also more subtle, 
unintended discrimination:

What is required by Congress is the removal of artificial, arbi­
trary and unnecessary barriers to employment when the barriers 
operate invidiously to discriminate on the basis of racial or other 
impermissible classification.

. . . The Act proscribes not only overt discrimination but also 
practices that are fair in form, but discriminatory in operation.

Griggs, 401 U.S. at 431.

In an individual disparate treatment case, the plaintiffs bur­
den in establishing a prima facie case is “ not onerous”. Btirdine, 
450 U.S. at 253. Essentially, the plaintiff need only prove that 
he is a member of a racial minority and eliminate the most com­
mon reasons for his failure to be hired or promoted or otherwise 
treated equally. See, e.g., id. at 253-54; McDonnell Douglas 
Corp. v. Green, 41 I U.S. 792, 802 (1973). Such a minimal show­
ing “ raises an inference of discrimination only because we pre­
sume these acts, if otherwise unexplained, are more likely than 
not based on the consideration of impermissible factors.” Furnco 
Construction Co. v. Waters, 438 U.S. 567, 577 (1978).

The employer in an individual disparate treatment case then 
need only articulate a legitimate, nondiscriminatory reason for *

* Most disparate impact cases are class actions. Individual disparate impact 
cases do exist, however, and the order of proof in such cases is. and should be. 
the same as in class action disparate impact cases.

439



10

its treatment of plaintiff and produce sufficient evidence to raise 
“ a genuine issue of fact as to whether it discriminated against 
the plaintiff.” Burdine, 450 U.S. at 254. The minimal burden on 
the employer in responding to an individual disparate treatment 
plaintiffs prima facie  case is commensurate with the low prima 
facie  threshold for the plaintiff. The entire purpose of the order 
of proof in individual disparate treatment cases is to narrow the 
issue of the employer’s intent gradually. Thus, the ultimate issue 
of discriminatory motive is most often decided, assuming 
plaintiff made a prima facie  showing, at the final stage, when 
the plaintiff must prove that the employer’s articulated non- 
discriminatory motive is a mere pretext for discrimination.

Said another way, it is not until the third and final stage of the 
order of proof in a disparate treatment case that the plaintiff 
actually proves discrimination by eliminating not only the most 
common nondiscriminatory motivations for the employer’s 
apparently discriminatory treatment but also the particular 
nondiscriminatory reasons proferred by the employer. For that 
reason, this Court in Burdine refused in an individual disparate 
treatment case to place any burden of persuasion on the 
employer at the second stage in the order of proof.

In a disparate impact case, in contrast, the order of proof, and 
particularly the plaintiffs’ prima facie  burden, are significantly 
different. As this Court itself noted in Burdine, “ the factual 
issues, and therefore the character of the evidence presented, 
differ when the plaintiff claims that a facially neutral employ­
ment policy has a discriminatory impact on protected classes.” 
450 U.S. at 252 n. 5. See also Teamsters, 431 U.S. at 336 n. 15 
(“ [cjlaims of disparate treatment may be distinguished from 
claims that stress ‘disparate impact’ ” ).

Thus, in a disparate impact case, the plaintiffs face a much 
higher prima facie  burden of proof. There, the plaintiffs must 
prove, in order to make a prima facie  show ing, that the employ­
er’s facially neutral employment practices and procedures cause 
a disparate impact upon a protected class. Eg. ,  Albermarle 
Paper Co. v. Moody, 422 U.S. 405, 425 (1975). That standard 
requires a substantial showing. The plaintiffs must prove, gener­

440



ally through the use of probative statistics, that the practices and 
procedures about which plaintiffs complain have a substantially 
disproportionate exclusionary impact on minorities. Dot hard v. 
Rawlinson, 433 U.S. 321,328-30 ( 1977).'

Once that standard is met, however, disparate impact 
plaintiffs have done far more than simply dispel the “ most com­
mon” nondiscriniinatory explanations for differences in employ­
ment results (as is the case with a disparate treatment prima 
facie  case). See Furnco, 438 U.S. at 577. Rather, a prima facie 
showing in a disparate impact case convincingly demonstrates 
the very evil that that type of analysis is designed to uncover, in 
a fashion that unless rebutted by the employer will compel a rul­
ing for the plaintiffs.6 The heavier burden carried by plaintiffs in 
prima facie  showings in disparate impact cases thus amply 
justifies the shifting of a burden of persuasion to the defendant. 
See B. Schlei and P. Grossman, Employment Discrimination 
Law  at 1328 n. 139 (2d ed. 1983) (“ [tjhe heavier burden placed 
upon the defendant in responding to a prima facie case under the 
adverse impact model corresponds with the plaintiffs’ heavier 
burden of establishing a prima facie case” ).

Indeed, the employer’s burden in responding to a disparate 
impact prima facie  case— to justify the business necessity of the 
challenged practices and procedures— is in the nature of an * *

5 As this Court is well aware, amt as this case convincingly demonstrates, 
the probative value of the statistics relied upon by a plaintiff class in a dispa­
rate impact case is often hotly disputed. The plaintiffs' burden to establish dis­
parate impact by statistics is indeed an onerous one. S e e  Part III in fra .  S e e  
a l s o  T e a m s t e r s ,  431 U.S. at 340 &. n. 20 (statistics come in an “infinite 
variety” and their usefulness “depends on all surrounding facts and 
circumstances”).

* As this Court has noted, “absent explanation, it is ordinarily to be 
expected that nondiscriniinatory hiring practices will in time result in a work 
force more or less representative of the racial and ethnic composition of the 
population in the community from which employees are hired " T e a m s te r s .  
431 U.S. at 340 n. 20. Thus, as Justice Blackmun stated in W a ts o n ,  “|u|n!ikc a 
claim of intentional discrimination, which the M c D o n n e l l  D o u g la s  factors 
establish only by inference, the disparate impact caused by an employment 
practice is directly established by the numerical disparity 487 U.S. at 
l08S.C t.at 2794.

11

441



12

affirmative defense. Once the plaintiffs have made a prima facie 
showing in a disparate impact case that the employment prac­
tices at issue are presumptively illegal, the employer can “save” 
those practices by demonstrating their business necessity. Unlike 
individual disparate treatment analysis— where, because the 
ultimate issue is the employer’s intent, the employer need only 
articulate a nondiscriminatory motivation not already disproven 
by the plaintiffs prima facie  showing under disparate impact 
analysis the employer’s burden is, after plaintiffs’ have proven 
disparate impact, to avoid the conclusion of unlawful discrimina­
tion by proving the business necessity of the practices and proce­
dures causing the disparate impact upon minorities. See, e.g., 
Albermarle Paper Co., 422 U.S. at 425.7

In sum, disparate impact analysis focuses solely on the effect 
of an employer's practices. A prima facie  showing of a statistical 
disparity in such a case is thus complete proof of unlawful dis­
crimination by the employer that, unless rebutted by proof of the 
business necessity of the challenged practices, mandates a 
finding in favor of plaintiffs. In contrast, a prima facie  showing 
in an individual disparate treatment case is nothing more than 
the first step in a process designed to ferret out the employer’s 
intent. That is the difference between an individual disparate 
treatment prima facie  showing and a disparate impact prima 
facie  showing. And that is why it is entirely appropriate to shift 
the burden of proof, rather than merely production, to the 
employer in disparate impact analysis.

T That is not to say, of course, that the employer may not challenge the 
accuracy or significance of plaintiffs’ statistics. D o l h a r d , 433 U.S. at 431. 
However, once the court has found disparate impact, the employer can only 
r e b u t  that finding by proving the business necessity of the offending practice. 
E g . ,  G r i g g s ,  401 U.S. at 431-32. That is the very nature of an affirmative 
defense. S e e  Fed. R. Civ. P. 8. Significantly, even petitioners agree that “busi­
ness necessity" is an affirmative defense in disparate impact cases. ( S e e  Pet. 
Brief at 42 (“if the employer remains silent on the issue of disparate impact, 
that issue is established and he must come forward with what amounts to an 
affirmative defense of business necessity")).

442



13

B. T his C o u rt H a s  U n ifo rm ly  H e ld  th a t the B urden o f  Proving  
B u sin ess  N e c e s s i ty  S h if ts  to  th e  E m p lo yer  F o llow in g  a 
P r im a  F acie  S h o w in g  in  a D isp a ra te  Im p a c t C ase.

Contrary to the suggestion of the United States (Brief for the 
United States at 25-28), this Court has consistently, indeed 
uniformly, held that in disparate impact cases, following a p r im a  

fa c ie  showing by plaintiffs, the burden of persuasion, not merely 
production, shifts to the employer. For this Court now to hold 
that the employer’s burden in disparate impact cases is one of 
production alone would require the overruling of a long and 
unbroken line of decisions dating back to G riggs v. D u ke Pow er  
C o ., 401 U.S. 424 (1971). There is no basis, we submit, for such 
a radical departure from well-established authority.

Thus, in G rig g s  itself, this Court flatly held that, once a statis­
tical showing of disparate impact is made, “Congress has placed 
on the employer th e  b u rd en  o f  sh o w in g  that any given require­
ment must have a manifest relationship to the employment in 
question” . 401 U.S. at 432 (emphasis added). Indeed, in G riggs, 
the employer articulated— but failed to prove— that its high 
school degree and standardized test requirements were related to 
successful job performance. This Court flatly rejected that 
proffer as insufficient to carry the employer's burden and 
reversed the Fourth Circuit's holding in favor of the employer. 
Id. at 431-36.

Sim ilarly, in A lb e rm a r le  P a p er  C o., this Court expressly held 
that the burden of persuasion of business necessity shifts to the 
employer, once plaintiffs make a p r im a  fa c ie  showing:

Title VII forbids the use of employment tests that arc discrimina­
tory in effect unless t h e  e m p l o y e r  m e e t s  t h e  b u r d e n  o f  s h o w i n g  

t h a t  a n y  g i v e n  r e q u i r e m e n t  / h a s /  . . .  <7 m a n i f e s t  r e l a t i o n s h i p  t o  

t h e  e m p l o y m e n t  in  q u e s t i o n ' .  This burden arises, of course, only 
after the complaining party or class has made out a prima facie 
case of discrimination, i . e . ,  has shown that the tests in question 
select applicants for hire or promotion in a racial pattern 
significantly different from that of the pool of applicants. / /  a n  

e m p l o y e r  d o e s  t h e n  m e e t  t h e  b u r d e n  o f  p r o v i n g  t h a t  i t s  t e s t s  a r e  

' j o b  r e l a t e d ’, it remains open to the complaining party to show 
that other tests or selection devices, without a similarly undcsira-

443



14

ble racial effect, would also serve the employer's legitimate inter­
est in ‘efficient and trustworthy workmanship'.

422 U.S. at 425 (citations omitted; emphasis added).
In Albermarle, as in Griggs, the defendant employer 

argUed— but did not prove— that the tests at issue were job 
related, offering a post-litigation validation study done using job 
criteria that were not in fact used by the defendant but rather 
were created by defendant’s expert. This Court had little trouble 
in rejecting that “ proof’ as insufficient to carry the defendant’s 
burden. Id. at 429-36.

Likewise, in Dothard, this Court held that, once a prima facie 
showing of disparate impact is made, the burden shifts to the 
employer to “ prove[] that the challenged requirements are job 
related” . 433 U.S. at 329. The Court further held that, in a dis­
parate impact case, “ a discriminatory employment practice must 
be shown to be necessary to safe and efficient job performance to 
survive a Title V II  challenge” . Id. at 332 n. 14. In Dothard, the 
employer articulated that its height and weight requirements for 
prison guards were related to strength, which the employer 
further hypothesized was related to effective performance as a 
prison guard. This Court quickly rejected that business necessity 
defense on the ground that the employer had failed to prove the 
relationships it articulated;

We turn, therefore, to the appellants’ argument that they have 
rebutted the prima facie case of discrimination by showing that 
the height and weight requirements arc job related. . . .  In the 
District Court, however, the appellants produced no evidence 
correlating the height and weight requirements with the requisite 
amount of strength thought essential to good job performance. 
Indeed, they failed to offer evidence of any kind in specific 
justification of the statutory standards.

433 U.S. at 331.
In short, this Court has consistently, and often, held that the 

employer's burden in a disparate impact case is one of persua­
sion, not merely articulation. That is, and always has been, the 
understanding of Title V II  practitioners representing both 
defendants and plaintiffs. There is no reason, we respectfully 
submit, for this Court now to overrule that long line of authority.

444



15

C. S o u n d  R eason s o f  P u b lic  P o lic y  an d  P r a c tic a li ty  W arran t 
th e  S h if t in g  o f  th e  B urden  o f  P ersu asion  to  the E m p lo yer  in  
D isp a ra te  Im p a c t C ases.

The only practical allocation of the burdens of proof in dispa­
rate impact cases is to shift the burden of persuasion of business 
necessity to the employer. Once the discriminatory effect of an 
employer’s practices is shown by the plaintiffs printa facie case, 
only the employer can fairly be expected to demonstrate that the 
practices in question are necessary.

For example, in a disparate impact case concerning an 
employment test or other objective measurement, an employer 
will most often defend job relatedness based upon validation of 
the measurement in question. See, eg., Eurnco, 438 U.S. at 
579-80. Validation is a complex, time consuming process and, as 
a practical matter, only the employer has sufficient access to, 
and familiarity with, the employment records and jobs at issue 
to conduct a validation study.8 Indeed, validation studies gener­
ally cannot be done properly simply by reviewing existing 
records. The party conducting such a study needs substantial 
access to current employees in order to administer a lest and to 
compare job success as predicted by the job requirement at issue 
to actual job success. C ivil discovery and access to Equal 
Employment Opportunity Commission files are simply not ade­
quate substitutes for the everyday access to the workplace 
enjoyed by employers. Indeed, plaintiffs are sometimes barred 
from any access to the workplace. See Belcher v. Bassett Furni­
ture Industries Inc., 588 F.2d 904 (4th Cir. 1978) (order 
granting plaintiffs’ counsel and expert five days access to 
defendant's plant reversed as an abuse of discretion). Moreover, *

* C'f. T e a m s t e r s ,  431 U.S at 360 n.45:

[TJhe employer |is| in the best position to show why any individual 
employee was denied an employment opportunity. Insofar as the reasons 
related to available vacancies or the employer s evaluation of the appli­
cant's qualifications, the company's records |arc| the most relevant 
items of proof. If the refusal to hire (was) based on other factors, the 
employer and its agents |know| best what those factors were and the 
extent to which (hey influenced the decision making process.

445



16

validation requires scores of hours of work and a ti.w..   ̂
familiarity with the requirements of the jobs at issue. Very few, 
if any, Title V II  plaintiffs have the resources and the particular 
expertise necessary for such a validation, even if they had the 
requisite access to the workplace.”

Finally, sound public policy mandates that the burden of prov­
ing business necessity rest on the employer. If  an employer’s per­
sonnel practices and procedures result in a statistically 
significant disparate impact on a protected class or classes, the 
employer should immediately, as a matter of public policy, vali­
date the business necessity of those practices and procedures. 
S e e  U n ifo rm  G u id e lin e s  on  S e le c tio n  P ro c ed u re s , 29 C.F.R. §§ 
1607, 1615 (1978).'° The employer should not wait until he is

’ For much the same reason, plaintiffs in disparate impact cases and pattern 
and practice disparate treatment cases cannot be expected to foresee each and 
every objection that employers might articulate at trial to their statistics. 
Plaintiffs cannot, in discovery, prepare for every such objection. Thus, as a 
practical matter, only the employer has adequate access to the facts to prove 
that its objections to plaintiffs’ statistics are soundly based in fact, and not 
merely hypothetical, and the employer should bear the burden of proving the 
factual basis of its objections. See, e g.. U n i t e d  S t a t e s  v. C o u n t y  o f  F a i r f a x ,  
629 F.2d at 940.

,0 That is not to say that an employer has a legal obligation to conduct vali­
dation studies as soon as a racial disparity is observed. We suggest only that 
this Court should encourage, rather than discourage, such employers from 
attempting to discover why such disparities exist and determining if the job 
requirements causing the disparity are truly necessary. Moreover, not all prac­
tices require formal validation studies. Many are valid on their face.

The Equal Employment Advisory Council (“EEAC”) suggests in its a m i c u s  
brief that validation studies cost between $100,000.00 and $400,000.00. (Brief 
for EEAC at 21 n. 4.) That estimate appears substantially high. Indeed, a 
survey of 1339 employers found that most validation studies cost as little as 
$5,000.00. See Fri -d man and Williams, C u r r e n t  U s e  o f  J e s t s  f o r  E m p l o y m e n t ,  

in 2 Ability "Tes'iog: Uses, Consequences, and Controversies 104, 110-11 
(1982) (“In all size categories, most companies that validated their lest or 
nontest selection procedures spent less than $5000 per job studied"). The 
EEAC further suggests that the Uniform Cuiidclincs on Selection Procedures 
are inconsistent with “generally accepted professional practices in test develop­
ment". ( I d .  at 21-22.) The American Psychological Association (“APA"). 
however, has gone on record with exactly the opposite position. Thus, in 1985, 
the APA wrote to EEOC Chairman Clarence Thomas that there was “no

446



17

sued under Title V II  to verify the business necessity of such 
practices and procedures. If, however, the employer’s burden in 
a Title V II  disparate impact action is merely one of articulating 
business necessity, employers will be discouraged from con­
ducting job validation studies in advance of litigation because to 
do so will expose them to a greater risk of liability (if, for exam­
ple, the study fails to show validity) than they face in litigation 
where the plaintiffs are unlikely to be able to conduct a 
definitive validation study."

II. SPECIFIC CAUSATION IS NOT TIIE APPROPRIATE 
STANDARD IN A DISPARATE IMPACT CASE

Petitioners further contend that, in a disparate impact case, 
the plaintiffs should not be allowed to challenge the cumulative 
effect of an employer's personnel practices but rather should be 
required to identify specific practices and demonstrate a specific 
disparate impact causally associated with each practice at issue. 
(Pet. Brief at 30-36.) Relying on Pouncy v. Prudential Insur­
ance Co., 668 F.2d 795, 800 (5th Cir. 1982) , petitioners argue 
that the disparate impact model is not “ the appropriate vehicle 
from which to launch a wide-ranging attack on the cumulative 
effect of a company’s employment practices. (Pet. Brief at 30 
(quoting Pouncy, 668 F.2d at 800).)

Petitioners’ argument addresses an issue not fairly presented 
by this case. Indeed, the short answer to petitioner s contention

compelling reason for revising” the uniform Guidelines on technical 
grounds”. 110 Daily Labor Rep. (BNA) A-3 (June 7, 1985).

" That analysis does not change when an employer’s subjective personnel 
policies are at issue. F ir s t ,  subjective personnel policies, like objective tests and 
other measurements, can be validated. S e e ,  e g . . Rose, S u b j e c t i v e  E m p l o y m e n t  
P r a c t i c e s :  D o e s  th e  D i s c r i m i n a t o r y  I m p a c t  A n a l y s i s  A p p l y ? ,  25 San Diego L. 
Rev. 63, 87-89 (1988). And s e c o n d ,  in any event, the employer, who is by 
definition the party most familiar with the requirements or the jobs at issue, is 
still in the best position to defend the necessity of the practices at issue. That 
fact docs not change simply because the practices are subjective in nature. And 
if the employer cannot defend the business necessity of his subjective personnel 
system in a Title VII case where disparate impact has been proven, he should 
not prevail.

447



18

is that this is not a case in which the plaintiffs made a shotgun, 
undifferentiated attack on the cumulative effect of an employ­
er’s personnel practices and procedures. To the contrary, 
plaintiffs challenged sixteen specific personnel practices used by 
petitioners. With respect to six of those practices, the Ninth C ir ­
cuit Court of Appeals held that plaintiffs’ challenges were well- 
founded. Thus, upon a review of the trial record, the Ninth C ir ­
cuit found that petitioners’ use of subjective criteria in making 
hiring decisions, petitioners’ nepotism policy, petitioners’ use of 
separate hiring channels and word-of-mouth recruitment for 
cannery and non-cannery jobs, and petitioners’ race labelling 
and segregated facilities caused a discriminatory impact upon 
minorities. For each of those practices, we submit, simple logic 
and well-established legal authority in this Court and the courts 
of appeals amply demonstrate a causal connection to disparate 
impact.

1. S u b je c tiv e  C rite r ia :  As this Court recognized only last 
term, the use of subjective criteria by a predominantly 
white, male supervisory force inevitably raises problems 
of “subconscious stereotypes and prejudices’’. W atson  v. 
F ort W orth  B an k  a n d  T r u s t, 487 U S. , 108 S.Ct. at 
2780 (1988). Courts of appeals have likewise recognized 
that the use of subjective criteria in employment decision­
making presents a “ ready mechanism’’ for discrimination, 
intentional or unintentional. E .g ., E E O C  v. in la n d  
M a rin e  In d u s tr ie s , 729 F.2d 1229, 1236 (9th C ir.), cert, 
d e n ie d , 469 U.S. 855 (1984); R o w e  v. G en era l M o to rs  
C o rp ., 457 F.2d 348, 359 (5th Cir. 1972).

2. N e p o tism :  Nepotism is, by definition, a practice of giving 
preference to relatives of current employees. Where the 
current employees are predominantly white, nepotism 
necessarily has an adverse impact on minorities. S e e , e.g ., 
B o n illa  v. O a k la n d  S c a v en g er  C o ., 697 F.2d 1297 (9th 
Cir. 1982), cert, d e n ie d , 467 U.S. 1251 (1984) ; G ibson  v. 
L o c a l 40 , 543 F.2d 1259, 1268 (9th Cir. 1976) .

3. S e p a r a te  h ir in g  ch an n els  a n d  w o rd -o f-m o u th  
rec ru itm e n t:  Where two work forces within a company

448



19

have significantly different racial compositions and the 
company employs both separate hiring channels and 
word-of-mouth recruitment the potential indeed, the 
likelihood— for disparate impact upon minorities is obvi­
ous. Thus, where the already predominantly white super­
visory force hires through word-of-mouth recruiting, it is 
only logical to expect that a predominantly white 
workforce will be perpetuated. E .g., B arn ett v. W. T. 
G ra n t C o ., 518 F.2d 543, 549 (4th Cir. 1975) ; Brown  v. 
G a sto n  C o u n ty  D yein g  M ach ine C o., 457 F.2d 1377, 1383 
(4th C ir.), cert, d en ied , 409 U.S. 982 (1972).

4. R a ce  la b e llin g  a n d  se g re g a te d  fa c ili t ie s :  Race labelling 
and segregated facilities— particularly in combination 
with the use of subjective criteria and word-of-mouth 
recruiting— similarly have an obvious, and adverse, 
impact upon the employment opportunities of minorities. 
Race labelling and segregated facilities reflect an obvious 
attitudinal “ headwind” in the way of employment oppor­
tunities for minorities. S e e , e .g ., G riggs, 401 U.S. at 432; 
L illy  v. H a r r is -T e a te r  S u p e rm a rk e t, 842 F.2d 1496, 1506 
(4th Cir. 1988) . More to the point, if minorities are rele­
gated to segregated facilities, they are isolated from the 
news of job opportunities spread by word-of-mouth 
among white employees. S e e , e .g .. U n ited  S ta te s  v. 
G eo rg ia  P o w er  C o ., 474 F.2d 906, 925 (5th Cir. 1973) ; 
D o m in g o  v. N e w  E n g la n d  Eish C o., 445 F. Supp. 421, 
435 (W .D. Wash. 1977), a f f d .  111  F.2d 1429 (9th Cir.), 
m o d if ie d , 742 F.2d 520 (9th Cir. 1984).

In sum, the causal connection between the practices about 
which plaintiffs complain here and disparate impact upon 
minorities is well-established. For petitioners to claim that

449



20

plaintiffs did not prove a causal connection flies in the face of 
both law and logic.12

Moreover, it would be virtually impossible for plaintiffs to 
prove with any more specificity the causal connection between a 
particular subjective practice and a particular disparate impact. 
Employers often do not maintain records that clearly show why 
certain applicants were hired or promoted and others were not. 
Absent such records, plaintiffs cannot hope to prove specific 
causation of disparate impact in hiring or promotion. And even 
statistical techniques often cannot fill that evidentiary gap. For 
example, multiple regression analysis can identify the 
significance of specified objective criteria to pay rates or hire 
rates. Bazemore v. Friday , 478 U.S. 385 (1986); Wilkins v. Uni­
versity o f Houston, 654 F.2d 388 (5th Cir. 1981) , vacated, 459 
U.S. 809 (1982), a jfd  on rehearing, 695 F.2d 134 (5th Cir. 
1983). However, multiple regression analysis is ill-suited to deal 
with unquantifiable variables such as subjective hiring criteria. 
Indeed, it is difficult to envision any method of isolating the 
significance of an individual subjective practice in such a situa­
tion, particularly by the plaintiffs who necessarily have far less

12 Moreover, the alternatives to these practices are obvious and cannot seri­
ously be contended to be onerous. First, word of mouth recruiting can be easily 
replaced and/or supplemented with a job posting system at the canneries dur­
ing the season, and at recruitment sites throughout the year. Second, the effect 
of separate hiring channels can be modified or eliminated by enabling company 
recruiters to recruit and provide information for all jobs ( i .e . ,  a recruiter going 
to an Alaskan Native village would be in a position to recruit individuals with 
skill as mechanics and not just for cannery workers). Third, with regard to 
subjective criteria, it is not a tremendous burden for the employers to establish 
and use objective job descriptions; such job descriptions would allow an appli­
cant to determine his or her qualifications for a position and would provide a 
standard by which applicants could reasonably be judged. Job descriptions are, 
in fact, a reasonable and fairly standard managerial practice. Fourth, with 
regard to race labelling and segregated facilities, the alternatives are simple 
and obvious. What justification can there be for assigning employee numbers 
by ethnic origin or referring to facilities bunkhouses, mess halls, etc.- by 
racial terms. And fifth, nepotism plainly has no significant relationship to job 
performance. Relatives of existing workers have no special qualifications neces­
sarily for the jobs at issue. There can be no hardship in simply eliminating 
nepotistic hiring

450



21

familiarity with the personnel system at issue than the defendant 
employer.13

In any event, it lies ill in the mouths of these petitioners to 
contend that allowing a disparate impact attack on the cumula­
tive effect of multiple employment practices is somehow unfair 
to employers. (.SVe Pet. Brief at 30-36.) Notwithstanding their 
protestations of the inability of employers’ to respond to cumula­
tive attacks, petitioners flatly claim to have proven the business 
necessity of each and every practice at issue. (.See Pet. Brief at 
36 (“even if petitioners in this case had such a burden, they have 
met it” ).)

Finally, even if the issue of the viability of a cumulative 
effects challenge were properly before this Court in this case, 
there is nothing unfair or inconsistent with Title V II theory in 
such a challenge. Indeed, even the United States concedes in its 
a m ic u s  brief that, in at least a multistage decision case, 
multifactor selection decisions may be challenged as a whole. 
(Brief for the United States at 22). S e e  a lso  T ea l, 457 U.S. at 
450 (Powell, J., dissenting) (“our disparate impact cases consist­
ently have considered whether the result of an employer s to ta l  
se le c tio n  p ro c e s s  had an adverse impact upon the protected 
group” ) (emphasis in original).

The Government’s concession is, we submit, compelled by this 
Court’s prior decisions and simple logic. F irst, it should not be 
forgotten that the s in e  qu a  non of a cumulative effects challenge 
is a statistical showing of a significant inequality in the employ­
er’s workforce statistics. If the employer's personnel system were 
working fairly and impartially, one would not expect to see such 
a statistical disparity. E g ., T e a m ste rs , 431 U.S. at 340 n.20.

 ̂ Even to attempt such proof of specific causation is a daunting task, for 
example, in a disparate impact case against the City of Houston, Texas, the 
Lawyers’ Committee sought to “disaggregate" over twenty different standards 
used by the employer and identify their specific disparate impact. That effort 
required the duplication of almost 150,000 pages of the defendant s records 
and the employment of approximately twenty temporary workers to review 
those records. That effort was, to say the least, extremely expensive. S e e  
T o n e r  v. C i t y  o f  H o u s t o n .  22  EPD 30,689 at p 14,627 (S.D Tex. 1980).

451



22

Thus, to hold that an employer is immune from Title V II  chal­
lenge simply because the plaintiffs are unable to identify the 
specific practice or practices causing specific portions of the dis­
parate impact would, in all likelihood, allow significant examples 
of employment discrimination to go unremedied. See Green v. 
U S X  Corp., 843 F.2d 1511, 1521-22 (3d Cir. 1988) , petition fo r  
cert, f ile d , 57 U .S.L.W . 3123 (U .S. July 23, 1988) (No. 
88-141).

Second, the broad remedial purpose, and express statutory 
language, of Title V II  support the proposition that cumulative 
effects challenges are proper. Indeed, in a Senate Report pre­
pared during the passage of the 1972 amendments to Title V II, 
Congress noted that employment “systems” can be, and often 
are, the cause of discrimination:

Employment discrimination is viewed today as a ... complex and 
pervasive phenomenon. Experts familiar with the subject now 
generally describe the problem in terms of 'sys tem s' and ‘effects' 
rather than simple intentional wrongs.

S. Rep. No. 415, 92nd Cong., 1st Sess. 5 (1971) (emphasis 
added).

Sim ilarly, the express language of 703(a)(2) of Title V II  pro­
vides broadly that it is an unlawful employment practice for an 
employer

to limit, segregate or classify his employees or applicants for 
employment in any w ay  which would deprive or tend to deprive 
any individual of employment opportunities or otherwise 
adversely affect his status as an employee, because of such indi­
vidual’s race, color, religion, sex or national origin.

42 U .S.C . § 2000e-2 (emphasis added)."

Congress’ broad proscription of employment practices that 
discriminate “ in any way” is certainly broad enough to encom­
pass challenges to the cumulative effects of personnel systems. 
As this Court noted in Griggs, Congress intended in Title V II to

" In 7e a t .  457 U.S. at 448, this Court noted that disparate impact analysis 
is based on § 705(a)(2) of Title VII

4 5 2



23

outlaw any and all employment practices that unnecessarily 
operate “as ‘built-in headwinds’ for minority groups.” 401 U.S. 
at 432. Thus, if the cumulative effects of an employer's entire 
personnel system deprive minorities of employment opportuni­
ties, it would be flatly inconsistent with the purpose of Title V II 
to exonerate that system absent a showing by the employer that 
its system is justified by business necessity (or, at least, that the 
practices and procedures the employer shows caused the dispa­
rate impact are justified).

Indeed, the Uniform Guidelines on Employee Selection Proce­
dures promulgated by the E E O C , Civil Service Commission and 
Departments of Labor and Justice support that conclusion. 
Those Procedures specifically define the employment practices 
that are subject to disparate impact review to include combina­
tions of practices. The Guidelines provide that disparate impact 
analysis applies to

(a]ny measure, c o m b i n a t i o n  o f  m e a s u r e s ,  or procedure used as a 
basis for any employment decision. Selection procedures include 
the full range of assessment techniques from traditional paper 
and pencil tests, performance tests, training programs, or proba­
tionary periods and physical, educational and work experience 
requirements through informal or casual interviews and unscored 
application forms.

2 9 C .F .R . § 1607.16(Q) (emphasis added).,s

Third , contrary to the assertions of petitioners and the United 
States as amicus curiae, there is nothing unfair about shifting 
the burden to the employer to choose whether disaggregation 
would serve its interests and, if so, to identify the practices 
causing disparate impact and to justify the business necessity of 
those practices once the plaintiffs have shown a disparity. It is 
employers who are most knowledgeable about their own person­
nel systems. And it is employers who have the best access to 
evidence concerning those systems. As the I bird C ircuit recently 
noted: 15

15 This Court has expressly held that the Uniform Ciuidclines arc entitled to 
great deference” as “the administrative interpretation of | title V!f| by the 
enforcing agency”. ,H h c r m a r l c , 422 U.S. at 411: G r ig g s ,  401 U.S. at 411-14.

453



24

Applying disparate impact analysis to this employer's hiring ‘sys­
tem' and measuring the disproportionate ‘effects' on minority 
hiring that result may impose a difficult burden on the employer, 
but not an unfair one. I n  t h e s e  c a s e s  t h e  e m p l o y e r  h a s  f a r  b e t t e r  

a c c e s s  a n d  o p p o r t u n i t y  t h a n  t h e  p l a i n t i f f s  t o  e v a l u a t e  c r i t i c a l l y  

t h e  i n t e r - r e l a t i o n s h i p  o f  t h e  c r i t e r i a  t h a t  i t  u s e s  i n  h i r i n g  p r a c ­

t i c e ,  a n d  t o  d e t e r m i n e  w h i c h  a s p e c t s  a c t u a l l y  r e s u l t  in  

d i s c r i m i n a t i o n .

Green v. U S X  Corp., 843 F.2d at 1524 (emphasis added). See 
alsoSegarv. Sm ith , 738 F.2d 1249, 1271 (D .C . C ir. 1984), cert, 
denied, 471, U.S. 1115 (1985).

And fourth, should this Court hold that it is always the 
plaintiffs’ burden to link specific employment practices with 
specific disparities, the result will be to encourage employers to 
have no personnel system at all, or to structure their employment 
systems in as complicated a fashion as possible (which may be 
the functional equivalent of no system at all), and to maintain as 
few personnel records as possible. In that way, employers may 
well be able to render themselves immune from Title V II  attack, 
no matter how skewed their employment statistics might be, 
because plaintiffs will be unable to identify the practice or prac-

454



25

tices that caused discrimination and/or prove the causal link. 
This Court, we submit, should not encourage such a result."1

III .  T H E  N IN T H  C IR C U IT  C O U R T  OF A P P E A L S  C O R ­
R E C T L Y  H E L D  T H A T  R E S P O N D E N T S ’ S T A T IS T IC S  
M A D E  O U T A P R IM A  F A C IE  C A S E  O F D IS P A R A T E  
IM P A C T

Finally, petitioners contend, again supported by the United 
States as amicus curiae, that plaintiffs’ statistics which show a

IB In its a m i c u s  brief, the United States suggests that this Court adopt a 
single governing formulation with respect to the inquiry into business 
justification once a plaintiff has made out a prima facie case under the dispa­
rate impact mode. (Brief for the United States at 23-25.) The United States 
proposes that this Court adopt a standard allegedly “encapsulated" in N e w  
Y o r k  C i t y  T r a n s i t  A u t h o r i t y  v. B e a z e r ,  440 U S. 568, 587 n.3l (1979). Thus, 
the United Stales would find a challenged practice justified as a business 
necessity where the employer's “legitimate employment goals of safety and 
efficiency . . . are significantly served by—even if they do not require - (the 
challenged selection practice]". That issue is not presented in this case, how­
ever, and we respectfully submit that this Court should not address an issue not 
briefed, argued or decided in the courts below and not the subject of this 
Court's writ of certiorari.

Furthermore, the standard proposed by the United States is too low and 
would thwart the central purpose of Title VII. As Justice Blackmun recognized 
in his concurrence in W a ts o n ,  “(precisely what constitutes a business necessity 
cannot be reduced, of course, to a scientific formula I V at so n ,  487 US. at

, 108 S.Ct. at 2794. Nevertheless, it is well-established that a mere 
“significant" relationship to “legitimate employment goals is not enough. 
“Congress has placed on the employer the burden of showing that any given 
requirement must have a m a n i f e s t  r e l a t i o n s h i p  to the employment in ques­
tion." G r i g g s ,  401 U.S. at 432 (emphasis supplied). As G r ig g s  made clear, 
“(t)he touchstone is business necessity." 401 U.S. at 301.

Moreover, in B e a z e r ,  this Court did not follow the standard that the United 
States proposes but merely recognized that the district court had made a 
finding that the defendant's employment goals of safety and efficiency actually 
d i d  require the exclusion of all users of illegal narcotics. The Court did not 
adopt a mere “relationship" standard as the employers burden in a disparate 
impact case. To the contrary, the Court expressly followed the standard articu­
lated in G r ig g s ,  noting that the record in B e a z e r  sufficiently reflected that the 
defendant's rule demonstrated a “manifest relationship to the employment in 
question". B c a z e r , 440 U.S. at 587 n 31 ( q u o t in g  (>riggs, 401 U.S. at 4 38)

455



26

striking disparity between a concentration of minority workers 
in lower level, lower paid cannery jobs and a paucity of minori­
ties in higher level, higher paid non-cannery jobs fail to make a 
prima facie  showing of disparate impact. In summary, petition­
ers contend that only statistics which compare the number of 
minorities in non-cannery jobs and the number of minorities in 
the qualified labor force are relevant here. (Pet. Brief at 15-24. 
See also Brief for the United States at 16-21.)

Petitioners have missed the point. This case is not about a sim­
ple comparison of the number of minorities in the non-cannery 
jobs and in the qualified labor pool. This case is about petition­
ers’ recruiting practices and the systematic exclusion of minori­
ties in low paid cannery positions from the opportunity to even 
apply, much less be hired, for the higher paid non-cannery jobs. 
For those issues, plaintiffs’ comparison between the number of 
minorities in cannery jobs and the number of minorities in non­
cannery jobs is entirely proper.

As the Fifth Circuit has stated:

‘In the problem of racial discrimination, statistics often tell 
much, and Courts listen.’ . .  . Our wide experience with cases 
involving racial discrimination in education, employment, and 
other segments of society has led us to rely heavily in Title VII 
cases on the empirical data which shows an employer's overall 
pattern of conduct in determining whether lie has discriminated 
against particular individuals or a class as a whole.

Burns v. Ih io k o l Chemical Corp., 483 F.2d 300, 305 (5th C'ir. 
1973) (citations omitted).17

There is no uniform rule that determines what types of statis­
tics are useful in what types of cases. The relevancy of particular

17 S e e  a l s o  7c a n i s t e r s ,  431 U.S. at 340 n. 20, q u o t i n g  13 Schlci and P. 
Grossman, E m p l o y m e n t  D i s c r i m i n a t i o n  L a w  at 1 161-93 (1976).

Since the passage of the Civil Rights Act of 1964, the courts have 
frequently relied upon statistical evidence to prove a violation. . . In 
many cases the only available avenue of proof is the use of racial statis­
tics to uncover clandestine and covert discrimination by the employer or 
union involved.

456



27

statistical showings can only be determined on a case-by-case 
basis. See, e.g., Hazehvood School District v. United States, 433 
U.S. at 311-12 ; Falcon v. General Telephone Co. o f the South­
west, 626 F.2d 369, 382 (5th Cir. 1980), vacated, 450 U.S. 
1036, a ffd  on rehearing, 647 F.2d 633 (5th Cir. 1981); Davis v. 
City o f Dallas, 483 F. Supp. 54, 60 (N.D. Tex 1979). As this 
Court pointed out in Teamsters, statistics “come in infinite vari­
ety” and “ their usefulness depends on all of the surrounding 
facts and circumstances” . 431 U.S. at 340. There is, in short, no 
hard and fast rule that statistical comparison in Title V II cases 
must be between the employer’s workforce and the “qualified 
labor pool” .

This is a unique case, involving seasonal work, often 
performed by migrant workers under exceptionally difficult con­
ditions. Routine statistical analyses do not apply. The facts and 
circumstances of this case mandate an approach to the relevant 
statistics tailored to the facts of this case.

Thus, here, a comparison to the so-called qualified labor force 
is beside the point. The crux of the issues raised by plaintiffs’ 
challenges to petitioners’ employment practices is the claim that 
those practices denied cannery workers the opportunity to com­
pete fairly for higher paying non-cannery jobs. By employing 
such hiring techniques as nepotism, word-of-mouth recruiting, 
separate hiring channels and use of subjective criteria, petition­
ers effectively precluded minority cannery workers from 
applying or being hired for non-cannery jobs. Similarly, such 
practices as race labelling and segregated facilities contributed 
substantially to a lack of knowledge of job opportunities on the 
part of minority cannery workers."' In total, those techniques 
assured that the current racial make-up of petitioners’ non­
cannery workforce would be perpetuated.

Petitioners contend that plaintiffs' workforce comparison statistics arc 
irrelevant because petitioners do not have a “promote from within policy. 
That again misses the point. Whether or not petitioners have such a policy, it is 
a violation of Title VI! to use employment practices that actively preclude 
lower level minority workers from the opportunity even to be considered for 
higher level positions where those practices cause disparate impact.

457



28

The fact that petitioners’ non-cannery workforce may reflect 
the racial breakdown of the qualified outside workforce is simply 
irrelevant to the issue of opportunity here. As this Court recog­
nized in Connecticut v. Teal, 457 U.S. 440 (1982) , the “ bottom 
line” of petitioners’ hiring practices is not a defense to a claim 
that those practices unlawfully curtail employment opportunities 
for minorities:

In considering claims of disparate impact under § 703(a)(2) this 
Court has consistently focused on employment and promotion 
requirements that create a discriminatory bar to opportun ities .  
This Court has never read § 703(a)(2) as requiring the focus to 
be placed instead on the overall number of minority or female 
applicants actually hired or promoted.

*  *  *

. . . The suggestion that disparate impact should be measured 
only at the bottom line ignores the fact that Title VII guarantees 
these individual respondents the o p p o r tu n ity  to compete equally 
with white workers on the basis of job related criteria.

Id. at 450-51 (emphasis in original).

Moreover, petitioners, as the Court of Appeals properly found, can­
not properly rely on their so-called qualified labor force statistics. As 
an initial matter, and as the district court found, Pet. App. I at 75-76, 
many of the non-cannery jobs at issue here were unskilled and hence 
required no particular qualification. Accordingly, the cannery workers 
plainly should have been eligible for those jobs. More generally, how­
ever, petitioners never proved that their purported job qualifications 
criteria were actually applied.19 Absent such proof, petitioners' 
qualified labor force statistics are irrelevant and worthless. Eg. ,  
Franks v. B ow m an Transporta tion  Co., 424 U.S. at 773 n. 32 (only 
non-discriminatory standards “actually applied"  by employers are 
relevant).

'* See, e g  . Pet App. at A-574-75 (trial testimony of l.arry L. DcPrance); 
Pet. App. at A-236 (Deposition of Warner Leonard).

458



29

In sum, the Ninth Circuit properly held that plaintiffs’ statistics 
made a p r im a  f a c ie  showing of disparate impact of discrimination in 
job opportunities.2"

CONCLUSION

In recent years, the number of employment discrimination 
class actions filed has declined precipitously, from a peak of 
1,174 in 1976 to only 46 in I988.2' A portion of that decrease 
may be attributable to a decline in employment discrimination 
in the United States, but there can be little doubt that private 
enforcement of Title V II  through class actions has suffered sub­
stantially in recent years as the cost in money and effort of 
prosecuting Title V II  class actions has risen substantially, if not 
exponentially. If the burdens of proof in disparate impact cases 
are revised as espoused by petitioners and the United Slates to 
further increase substantially, indeed drastically, the burden of 
proof on plaintiffs and correspondingly decrease the employer's 
rebuttal burden, we fear that no plaintiffs will have the resources 
or, indeed, the incentive to pursue Title V II  class actions. The 
most important method of enforcement of Title V II  the class 
action— may, for all practical purposes, cease to exist. That 
would, we submit, be a most unfortunate result for the cause of 
equal employment opportunity.

20 If this Court should conclude that the facts of record do not make out a 
p r i m a  f a c i e  case of disparate impact, then we respectfully suggest that the 
Court remand this matter for the presentation of further evidence and findings 
by the District Court. As the appellate process in this case demonstrates, given 
two panel opinions and one en  b a n c  decision, as well as this Court’s opinion, the 
legal standards governing plaintiffs’ case have shifted considerably during the 
litigation. Accordingly, plaintiffs should be given an appropriate opportunity to 
conform the evidence to the proper legal standard. S e c ,  e .g . ,  A l b c r m a r l e ,  422 
U S. at 436 (where the Court remanded the case to the District Court to allow 
both the plaintiffs and the defendant to revise their evidentiary showings to 
conform to the new legal standards set forth in the Court's opinion).

21 1977 Report of the Director, Administrative Office of the U S. Courts, 
Table 32, p. 239; Table X-5. Unpublished Computer Analysis prepared by the 
Administrative Office of the U. S. Courts.

459



30

Accordingly, for the reasons set forth above, the Lawyers’ 
Committee for C ivil Rights Under Law respectfully submits 
that the decision of the Ninth Circuit Court of Appeals should 
be affirmed.

Respectfully submitted,

Co n r a d  H arper  
St u a r t  J. La n d  
Co-Chairmen 
N o r m a n  R edlich 
Trustee
R ichard  T . S eymour  
James  C . G r a y , Jr .

La w yers’ C ommittee  
For C ivil R ights 
U n d e r  Law 
1400 “ Eye” Street, N.W . 
Suite 400
Washington, D. C. 20005 
(202) 371-1212

N icholas df.B. Ka t z e n b a c h* 
A lan E. Kraus 
D avid  A rciszewski

R iker , D a nz ig , Scherer , 
H y l a n d  & Perretti 

Headquarters Plaza 
One Speedwell Avenue 
Morristown, New Jersey 

07960-1981 
(201) 538-0800

Attorneys for Amicus Curiae 
Lawyers’ Committee for C ivil Rights 

Under Law

November 4, 1988

^Counsel of Record

460



No. 87-1387

I n  t h e

S'itjirtmf (Emtrt nl %  lu itfi Slatra
October T erm, 1988

W ards Cove P acking Company, I nc ., and  
Castle & Cooke, I nc.,

Petitioners,
v.

F rank A tonio, et al.,
Respondents.

on writ of certiorari to t h e  u n ited  states 
COURT OF APPEALS FOR THE NINTH CIRCUIT

BRIEF FOR THE NAACP LEGAL DEFENSE AND 
EDUCATIONAL FUND, INC., THE MEXICAN AMERICAN 

LEGAL DEFENSE AND EDUCATIONAL FUND, AND 
THE PUERTO RICAN LEGAL DEFENSE AND 

EDUCATION FUND AS AMICI CURIAE  
SUPPORTING RESPONDENTS

A ntonia H ernandez 
E. R ichard Larson 
J ose Roberto J uarez, J r. 

Mexican American Legal 
Defense and Educational 
Fund

634 South Spring Street 
11th Floor
Los Angeles, CA 90014 
(213) 629-2512

R uben F ranco 
Kenneth  K imerlino

Puerto Rican Legal Defense 
and Education Fund 

99 Hudson Street 
New York, N.Y. 10013 
(212) 219-3360

J ulius LeV onne Chambers 
Charles Stephen Ralston 
Ronald L. E llis 

NAACP Legal Defense and 
Educational Fund, Inc.

99 Hudson Street 
New York, N.Y. 10013
(212) 219-1900

B ill Lann Lee*
P atrick O. P atterson, J r. 
Theodore M. S haw 

NAACP Legal Defense and 
Educational Fund, Inc. 

634 South Spring Street 
Suite 800
Los Angeles, CA 90014
(213) 624-2405

C o u n s e l  f o r  A m i c i  C u r i a e

•Counsel of Record

461



QUESTIONS PRESENTED

1. Whether, on the facts of this 
case, the court of appeals correctly held 
that the evidence established a prima 
facie case of disparate impact.

2. Whether this Court should 
overrule the evidentiary standards for 
disparate impact cases articulated in 
Griaas v. Duke Power Co. and its progeny.

3. Whether, on the facts of this 
case, the court of appeals correctly 
considered the cumulative effect of a 
range of employment practices as 
demonstrating the consequences of 
discriminatory practices that had already 
been independently established.

i

462



TABLE OF CONTENTS
Page

Interest of Amici Curiae .......... 1
Summary of Argument ............... 2
ARGUMENT

I. TITLE VII, BY ITS TERMS,PROHIBITS DISPARATE 
IMPACT DISCRIMINATION AS WELL AS DISPARATETREATMENT DISCRIMINATION . . . .  9

II. THE LEGISLATIVE HISTORY OF 
TITLE VII, THE 1972 AMENDMENTS, AND THE UNIFORM ADMINISTRATIVE INTERPRETATION 
OF THE STATUTE DEMONSTRATE 
THAT THE EVIDENTIARY STANDARDS 
ARTICULATED IN GRIGGS AND ITS 
PROGENY ARE CONSISTENT WITH 
THE INTENT OF CONGRESS........13
A. In Enacting §703(a)(2)In 1964, Congress Specifically Intended To 

Prohibit "Institutionalized" 
Disparate Impact Discrim­
ination Not Motivated By Any 
Discriminatory Purpose . . .  13

B. In Amending Title VII In 1972, 
Congress Ratified
The §703(a)(2) Evidentiary 
Standards Articulated In Griggs.................. 20

ii

463



III.

C. The Evidentiary Standards Of 
Griggs And Its Progeny 
Have Been Uniformly Confirmed 
By Administrative Interpretations Of 
§703 (a) ( 2 ) ..............  26

THE SEPARATE EVIDENTIARY ANALYSES 
DEVELOPED BY THE COURT REFLECT THE 
DISTINCT NATURE OF THE 
DISCRIMINATORY PRACTICES CONGRESS 
INTENDED TO PROSCRIBE IN 
§§703(a)(1) AND 703(a)(2) . . .  29
A. The Court Has Articulated 

Evidentiary Standards 
For Analyzing Disparate Treatment Claims Under Section 703(a)(1) . . . .  30
1• Individual DisparateTreatment . . . .  31
2. Direct Evidence of Intentional

Discrimination........ 33
3. Pattern or Practice of Intentional

Discrimination . . . .  35

B. The Court Has Articulated
Separate Evidentiary 
Standards For Analyzing 
Disparate Impact Claims Under 
Section 703(a)(2) ........  36

C. The Griggs Disparate Impact 
Analysis Is Analogous To The

iii

464



Teamsters And Thurston
Disparate Treatment
Analyses............ 44

IV. OVERRULING THE EVIDENTIARY 
STANDARDS OF GRIGGS AND ITS 
PROGENY WOULD BE CONTRARY TO THE REMEDIAL
PURPOSE OF TITLE VII.........  47

V. THE FIRST AND THIRD QUESTIONS 
PRESENTED IN THE PETITIONFOR CERTIORARI ARE NOT PRESENTED 
BY THE FACTS OF THIS CASE........54

CONCLUSION........................ 62

i v
465



Table of Authorities

Cases:
Page

Albemarle Paper Co. v. Moody,422 U.S. 405 (1975) . . 21, 27, 38,................ 41, 60
Colby v. J.C. Penney Co.,811 F.2d 1119 (7th Cir. 1987) . 11
Connecticut v. Teal, 457 U.S. 440 . 6,.............. 20, 21, 26, 38, 59
Dothard v. Rawlinson, 433 U.S. 321(1977).................. 37, 41
Espinoza v. Farah Mfg. Co., 414 U.S.

86 (1973) ..................  27
Firefighters Inst, for RacialEquality v. City of St. Louis,

616 F.2d 350 (8th Cir. 1980), 
cert, denied. 452 U.S. 938 
(1981) ................... 53

Franks v. Bowman Transportation Co.,424 U.S. 747 (1976) . . . .  21, 35
Furnco Construction Corp. v. Waters,

438 U.S. 567 (1978) . . . 32, 43
General Electric Co. v. Gilbert,

429 U.S. 141 ( 1 9 7 6 ) .......... 27
Green v. USX Corp., 843 F.2d 1511(3rd Cir. 1988)............ 59

v

466



Griggs v. Duke Power Co., 401 U.S.
424 (1971) .............. passim

Guardians Association v. Civil Service Commission,
463 U.S. 582 (1983) .......... 42

Hazelwood School District v. United
States, 433 U.S. 299 (1977) . . 36

International Brotherhood ofTeamsters v. United States,
431 U.S. 324 (1977) . . . .  passim

Johnson v. Railway Express Agency,
421 U.S. 454 (1975)........... 21

Local 28, Sheet Metal Workers v.
EEOC, 478 U.S.421 (1986) 25, 26, 27

Local 93, Firefighters v. City of
Cleveland, 478 U.S. 501 (1986). .27

Los Angeles Department of Water
& Power v. Manhart, 435 U.S.
702 (1978) . . . . . .  .11, 34, 35

Lowe v. City of Monrovia, 775 F.2d 998
(9th Cir. 1985)................ 38

McDonnell Douglas Corp. v. Green,
411 U.S. 792 (1973)........ passim

Nashville Gas Co. v. Satty, 434 U.S.136 (1977)........ 11, 12, 13, 39
New York City Transit Authority v.Beazer, 440 U.S. 568 (1979) . . 41
Phillips v. Martin Marietta Corp.,400 U.S. 542 (1971) . . . .  34, 35

Vi
467



Segar v. Smith, 738 F.2d 1249 (D.C. Cir. 1984) . . . . 59

Texas Department of Community
Affairs v. Burdine 450 U.S.
248, n. 8 (1981) . • • 31, 32, 33

40, 43, 50
Tillery v. Pacific Tel. Co. , 34

FEP Cases 54 (N.D. Cal . 1982) . 53
Trans World Airlines v. Thurston,

469 U.S. Ill (1985) • • . 8, 31, 33
, 44, 45, 46

Wade v. Mississippi Coop. Extension Serv., 615 F. Supp. 1574 (N.D.Miss. 1985).................... 53
Wambheim v. J.C. Penney Co., 705

F.2d 1492 (9th Cir. 1983), cert, 
denied. 467 U.S. 1255 (1984). . 11

Watson v. Fort Worth Bank and Trust, 108 
S. Ct. 2777 (1988) . . . .  3, 5, 32, 38,

.......................... 41, 48
Wilson v. Michigan Bell Tel. Co., 

550 F. Supp. 1296 (E.D. Mich.1982) ....................... 53
Legislative Materials:
H.R. 405 ..........................  15
H.R. Rep. No. 88-570 ...............  16
H.R. Rep. No. 92-238 ........ 22, 23, 24
88 Cong., 1st Sess. 144-45 (1963) . . 19
110 Cong. Rec. 6307 (1964)...........  19

v i i

468



117 Cong. Rec. 32108 (1971)......... 51
117 Cong. Rec. 38402 (1971)......... 51
118 Cong. Rec. 697 (1972) . . . .  22, 25
118 Cong. Rec. 7166 (1972).......... 25
S. Rep. No. 88-867 (1964)............ 17
S. Rep. No. 92-415 .............. 22, 23
Statutes:
42 U.S.C. § 2000e-2 (a) (1)........ passim
42 U.S.C. § 2000e-2(a)(2) . . . .  passim
Administrative Materials:
29 C.F.R. § 1607 (1986).............. 28
29 C.F.R. § 1607.3 (1970) . . . .  28, 29
35 Fed. Reg. 12333 (1970)  28
35 Fed. Reg. 12336 (29 C.F.R.§ 1607.11)...................... 28
43 Fed. Reg. (1978)  28
Other Authorities:
B. Schlei & P. Grossman, Employment Discrimination Law, 202 (2d ed.

1983)........................  52
Rose, Subjective Employment Practices:_ 

Does the Discriminatory Impact 
Analysis Apply?# 25 San Diego 
L.R. 63 (1988).......... 14, 52

viii

469



No. 8 7 - 1 3 8 7

IN THESUPREME COURT OF THE UNITED STATES 
October Term, 1988

WARDS COVE PACKING COMPANY, INC., and 
CASTLE & COOKE, INC. ,

Petitioners.
v.

FRANK ATONIO, et al., 
Respondents.

On Writ of Certiorari to the United States Court of Appeals for the Ninth Circuit

BRIEF FOR THE NAACP LEGAL DEFENSE AND EDUCATIONAL FUND, INC., THE MEXICAN AMERICAN LEGAL DEFENSE AND EDUCATIONAL 
FUND, AND THE PUERTO RICAN LEGAL DEFENSE AND EDUCATION FUND AS AMICI CURIAE 

SUPPORTING RESPONDENTS

INTEREST OF AMICI CURIAE 
Amicus NAACP Legal Defense and 

Educational Fund, Inc., is a national 
civil rights legal organization that has 
litigated many cases on behalf of black

470



2

persons seeking vindication of their civil
rights, including Griggs v. Duke_Power
Co., 401 U.S. 424 (1971). Amicus Mexican
American Legal Defense and Educational 
Fund and amicus Puerto Rican Legal Defense 
and Education Fund are national civil 
rights organizations that have brought 
various lawsuits on behalf of Latino 
persons subject to discrimination in 
employment, education, voting rights and 
other areas of public life. Letters from 
the parties consenting to the filing of 
this brief have been filed with the Court.

SUMMARY OF ARGUMENT 
Amici, supporting respondents, 

principally address the important issue 
raised by the second question presented in 
the petition for certiorari —  viz., the 
continued vitality of Griggs v. Duke Power 
Co.

471



3

In Watson v. Fort Worth Bank and 
Trust, 108 S. Ct. 2777, 2785 (1988) (part
IIA) , Justice O'Connor, writing for the 
Court and citing Griggs, reiterated that 
Title VII proscribes not only intentional, 
disparate treatment discrimination but 
also disparate impact discrimination: 
"This Court has repeatedly reaffirmed the 
principle that some facially neutral 
employment practices may violate Title VII 
even in the absence of a demonstrated 
discriminatory intent." The Watson 
opinion also observed that "the necessary 
premise of the disparate impact approach 
is that some employment practices, adopted 
without a deliberately discriminatory 
motive, may in operation be functionally 
equivalent to intentional discrimination." 
Id. (emphasis added).

The petitioners in this case concede 
that, "[u]nder a strict reading of

472



4

Griggs . " once the plaintiff has 
established a prima facie case of 
disparate impact the employer "must come 
forward with what amounts to an 
affirmative defense of business 
necessity." Brief for Petitioners at 42 
(citation and footnote omitted). The 
Solicitor General, however, distorts the 
language of Watson to argue that Griggs' 
burden of proof standards are "[b]ased on 
the assumption that certain other 
exclusionary practices are 'functionally 
e q u i v a l e n t  to i n t e n t i o n a l  
discrimination.'" Brief for the United 
States as Amicus Curiae at 13. The 
Solicitor General then goes on to argue 
that, once the plaintiff has established a 
prima facie case of disparate impact 
discrimination, the employer's burden of 
demonstrating business necessity should be 
revised to conform to the employer's



5

minimal burden of production imposed under 
McDonnell Douglas Cora, v. Green. 411 U.S. 
792 (1973), in individual disparate
treatment cases. 14. at 27 ("Nothing 
about disparate impact cases justifies a 
departure from the model for litigating 
disparate treatment cases"). Compare 
Watson. 108 S. Ct. at 2787-2791 (parts II 
C&D) (O'Connor, J.).1

The Solicitor General's argument 
conflicts with the language of the 
statute, its legislative history and 
c o n t e m p o r a n e o u s  administrative 
interpretations, the prior decisions of

*In Watson. the Solicitor General argued that subjective employment practices could only be analyzed under an 
intentional discrimination standard. See 108 S. Ct* at 2786. The Court rejected the argument. In the present case, the 
Solicitor General seeks to accomplish indirectly —  through the subterfuge of modifying disparate impact standards of 
proof to conform to individual disparate treatment standards —  what the Court directly rejected in Watson.

474



6

this Court, and the remedial purpose of 
Title VII.

1. "A disparate impact claim
reflects the language of §703(a)(2)," 
Connecticut v. Teal. 457 U.S. 440, 448
(1982) , which proscribes practices that 
•'deprive or tend to deprive any individual 
of employment opportunities." 42 U.S.C. 
§2000e-2(a)(2). The individual disparate 
treatment analysis, on the other hand, is 
one of several evidentiary models for 
analyzing violations of §703(a)(l), 42
U.S.C. §2000e-2(a)(1).

2. The legislative history of Title
VII's enactment in 1964, and of its
amendment in 1972, both undermine the
Solicitor General's argument. In 1964, 
Congress made unmistakably clear that it 
intended to prohibit both intentional 
discrimination and disparate impact 
discrimination. Purposeful, overt

475



7

discrimination was not regarded as a 
paradigm? Congress expressly declared that 
Title VII reached beyond overt practices. 
In 1972, Congress specifically ratified 
Griggs and its evidentiary standards for 
disparate impact cases. Contemporaneous 
administrative interpretations of Title 
VII, including those of the Department of 
Justice and the EEOC, have uniformly 
applied the Griggs disparate impact 
analysis to all selection procedures with 
an adverse impact, and they have 
separately prohibited disparate treatment.

3. Based on the language and 
legislative history of §703(a), the Court 
has developed separate evidentiary 
analyses that recognize the basic 
differences between disparate treatment 
and disparate impact discrimination. The 
individual disparate treatment analysis of 
McDonnell Douglas serves different ends

476



8

than those served by the disparate impact 
analysis of Griggs,» the stages of the two 
evidentiary models are specific to each 
analysis and are in no way comparable. 
The more appropriate analogy for the 
employer's burden in a disparate impact 
case —  if an analogy is necessary—  
would be the employer's burden in class- 
based disparate treatment cases, such as 
International Brotherhood of Teamsters v. 
United States. 431 U.S. 324 (1977), and 
Trans World Airlines v. Thurston, 469 U.S. 
Ill (1985).

4. The Solicitor General's theory, 
if accepted, would frustrate the remedial 
purpose of Title VII by overruling Griggs 
and effectively repealing §703(a)(2)'s 
prohibition of arbitrary practices that 
have the effect of depriving minorities or 
women of employment opportunities.



9

Amici also submit that the first and 
third questions presented in the petition 
for certiorari are not actually presented 
by the facts of this case, and that the 
Court should not attempt to resolve those 
questions on this record.

ARGUMENT

I .  T IT L E  V I I ,  BY IT S  TERMS, PR O H IBITS  
DISPARATE IMPACT DISC RIM INA TIO N AS  
W E L L  A S  D I S P A R A T E  T R E A T M E N T  
D ISC R IM IN A T IO N .

The individual disparate treatment 
model of McDonnell Douglas. which the 
Solicitor General would extend to 
disparate impact cases, was developed to 
analyze claims of intentional 
discrimination against individual 
plaintiffs under 5703(a)(1) of Title VII. 
See McDonnell Douglas. 411 U.S. at 676-77. 
"A disparate impact claim," on the other 
hand, "reflects the language of 
5703(a)(2)." Teal. 457 U.S. at 448.

478



1 0

The two subparts of §703(a) state:
It shall be an unlawful 

employment practice for an 
employer:

1. to fail or refuse to hire 
or to discharge any individual, or 
otherwise to discriminate against 
any individual with respect to his compensation, terms, conditions, 
or privileges of employment, 
because of such individual's race, 
color, religion, sex, or national 
origin; or

2. to limit, segregate, or 
classify his employees or 
applicants for employment in any 
way which would deprive or tend to 
deprive any individual of employment opportunities or 
otherwise adversely affect his status as an employee, because of 
such individual's race, color, religion, sex, or national origin.

42 U.S.C. §2000e-2(a). This statutory
language establishes a comprehensive
framework embracing both 
employment discrimination: 
treatment and disparate impact.

The Court has applied §703 
variety of circumstances

forms of 
disparate

(a)(1) in a
involving

479



11
intentional discrimination. See e.q.. 
McDonnell Douglas (individual disparate 
treatment); Los Angeles Department of 
Water & Power v. Manhart. 435 U.S. 702 
(1978) (direct evidence of a policy of 
disparate treatment)? Teamsters (pattern 
or practice of disparate treatment). The 
Court, however, has "not decide[d] 
whether, when confronted by a facially 
neutral plan, it is necessary to prove 
intent to establish a prima facie 
violation of §703(a)(l)." Nashville Gas 
Co. v. Sattv. 434 U.S. 136, 144 (1977).2

The separate and distinct objective 
of Congress in enacting §703(a)(2) "is 
plain from the language of the statute."

2Several lower courts have held that 
disparate impact challenges may also be 
brought under §703(a)(l). See. e.q.. Colbv v. J.C. Penney Co.. 811 F.2d 1119, 1127 (7th Cir. 1987) ? Wambheim v. J.C. Penney Co.. 705 F. 2d 1492, 1494 (9th Cir. 
1983), cert, denied. 467 U.S. 1255 (1984).

480



1 2

Griggs, 401 U.S. at 429. Section
703 (a)(2) "speaks, not in terms of jobs 
and promotions, but in terms of 
limitations and classifications that would 
deprive any individual of employment
opportunities." Teal. 457 U.S. at 449
(original emphasis).

A disparate impact claim
reflects the language of §703(a) (2) and Congress' basic 
objectives in enacting that 
statute: "to achieve eguality of
employment opportunities and 
remove barriers that have operated 
in the past to favor an 
identifiable group of white
employees over other employees." r Griggs.] 401 U.S. at 429-430(emphasis added).

Id. (original emphasis). See Sattv. 434
U.S. at 141 (ruling that denial of 
pregnancy benefits is permissible under 
§703 (a)(1) "does not allow us to read 
§703(a)(2) to permit an employer to burden 
female employees in such a way as to 
d e p r i v e  t h e m  of e m p l o y m e n t  
opportunities").

481



13

"Proof of discriminatory motive . . . 
is not required," Teamsters. 431 U.S. at 
335 n.15, by the terms of §703(a)(2). As 
then-Justice Rehnquist put it, "Griggs 
held that a violation of §703(a)(2) can be 
established by proof of a discriminatory 
effect." Sattv. 434 U.S. at 144.

I I .  THE LEG ISLA TIV E HISTORY OF T IT L E  V I I ,  
THE 1 9 7 2  AMENDMENTS, AND THE
U N IFO R M  A D M IN IS T R A T IV E  IN T E R P R E ­
TATION OF THE STATUTE DEMONSTRATE 
T H A T  TH E E V I D E N T I A R Y  STANDARDS  
ARTICULATED IN  GRIGGS AND IT S  PROGENY 
ARE CONSISTENT WITH THE INTENT OF 
CONGRESS.

A. In Enacting §703(a)(2) In 1964,Congress Specifically Intended
To Prohibit "Institutionalized" 
Disparate Impact Discrimination 
Not Motivated By Any
Discriminatory Purpose.

The 1964 legislative history confirms
this Court's assessment of Title VII seven
years later in Griggs, 401 U.S. at 429-30,
that: "The objective of Congress in the
enactment of Title VII . . . was to

482



14

achieve equality of employment 
opportunities and remove barriers that 
have operated in the past to favor an 
identifiable group of white employees over 
other employees," whether those barriers 
were erected by intentional, racially 
motivated discrimination or by unjustified 
practices with a disparate impact.3 
Congress did not see disparate impact 
discrimination as another form of 
disparate treatment discrimination, but 
rather as a separate evil which Title VII 
separately addressed.

The forerunner of §703(a)(2) was
contained in House and Senate bills
introduced in the 88th Congress, from

3See Rose, Subjective Employment 
Practices: Does the Discriminatory Impact Analysis Apply?. 25 San Diego L.R. 63, 73- 
81 (1988) (author was chief of the section 
of the Department of Justice's Civil 
Rights Division responsible for enforcement of Title VII).

483



1 5

which Title VII of the omnibus Civil 
Rights Act of 1964 eventually emerged. 
Section 5(a)(2) of H.R. 405, which was 
favorably reported in H.R. Rep. No. 88- 
570 (1963), prohibited the limitation,
segregation, or classification of 
employees "in any way which would deprive 
or tend to deprive any person of 
employment opportunities or otherwise 
adversely affect his status as an 
employee" because of prohibited 
discrimination. Id. at 8.

The House Committee reported that 
discrimination in employment was "a 
pervasive practice" throughout the country 
and that it "permeate[d] the national 
social fabric —  North, South, East and 
West." Id. at 2.

. . . Job discrimination is
extant in almost every area of 
employment and in every area of 
the country. It ranges in degrees from patent absolute rejection to 
more subtle forms of invidious

4B4



1 6

distinctions. Most frequently, it 
manifests itself through relega­
tion to "traditional" positions 
and through discriminatory 
promotional practices.

Id. The House report attributed high
minority unemployment and underemployment 
in part to such discriminatory practices. 
Id. Opponents of the bill attacked the 
breadth of the prohibition.4 However, 
with the addition of sex as one of the 
prohibited bases for unlawful employment 
practices, H.R. 405 passed without any 
amendment of this substantive provision.

In the Senate, language similar to 
§703(a)(2) appeared in S. 1937, a bill 
introduced by Senator Humphrey, who was 
later the floor manager for the omnibus

4H.R. Rep. No. 88-570 at 110-11 
(minority view of Reps. Poll and Crames.)

485



17

Civil Rights Act of 1964.5 The bill was 
reported favorably out of the Senate Labor 
Committee on February 4, 1964. S. Rep.
No. 88-867 (1964). Section 4(a) of S.
1937 made unlawful the discriminatory 
denial of "equal employment opportunity," 
including any practice which "results or 
tends to result in material disadvantage 
or impediment to any individual in 
obtaining employment or the incidents of 
employment for which he is otherwise 
qualified." Id. at 24. The Senate 
report, written by Senator Clark, who was 
later the bipartisan floor leader for 
Title VII, explained that:

Overt or covert discrimi­
natory selection devices, intentional or unintentional, 
generally prevail throughout the 
major part of the white economic community. Deliberate procedures

5Senators Clark and Case, who were 
later the bipartisan Senate floor leaders 
for Title VII, were co-sponsors.

486



18

operate together with widespread 
built-in administrative processes 
through which nonwhite applicants 
are automatically excluded from 
job opportunities. Channels for 
job recruitment may be tradition­
ally directed to sources which by 
their nature do not include 
nonwhites? trainees may be 
selected from departments where 
Negroes have never worked; 
promotions may be based upon job experience which Negroes have 
never had.

As Secretary of Labor Wirtz 
stated in his testimony before the 
committee:

Discrimination has become, 
furthermore, institution­
alized so that it obtains today in some organizations 
and practices and areas as the product of inertia, 
preserved by forms and habits 
which can best be broken from 
the outside.

Id. at 5. According to the Committee, S. 
1937 defined "equal employment opportunity 
in broad terms to include a wide range of 
incidents and facilities, and 
e n c o m p a s s e  [ d ] all aspects of 
discrimination in employment because of 
race, color, religion, or national

487



19

origin." Id. at 10. The report declared 
that the substantive provision was 
"designed specifically to reach into all 
of the institutionalized areas and 
recesses of discrimination, including the 
so-called built-in practices preserved 
through form, habit or inertia." Id. at 
11. See also. Hearings on Equal 
Employment Opportunity Before the 
Subcommittee on Employment of the Senate 
Committee on Labor and Public Welfare, 
88th Cong., 1st Sess. 144-45 (1963)
(remarks of Sen. Humphrey).

Senator Humphrey, as principal floor 
manager, introduced the omnibus bill that 
contained Title VII, H. 7512, on the floor 
of the Senate on March 30, 1964. 110
Cong. Rec. 6307. While the omnibus bill 
opted for court enforcement as opposed to 
the administrative cease-and-desist 
authority proposed in the Labor Committee

488



2 0

bill, the substantive focus of §703(a)(2) 
-- the broad prohibition of practices 
resulting in the denial of employment 
opportunities —  remained the same. In 
explaining the bill, Senator Humphrey 
stated that, "at the present time Negroes 
and members of other minority groups do 
not have an equal chance to be hired, to 
be promoted, and to be given the most 
desirable assignments. . . . The crux of
the problem is to open employment 
opportunities for Negroes in occupations 
which have been traditionally closed to 
them." Id. at 6547, 6548.

The language of §703(a)(2) passed 
both houses intact.

B. In Amending Title VII In 1972, 
Congress Ratified The §703(a)(2) 
E v i d e n t i a r y  S t a n d a r d s  
Articulated In Griggs.

As the Court concluded in Teal. 
"[t]he legislative history of the 1972

489



21

amendments to Title VII . . . demonstrates 
that Congress recognized and endorsed the 
disparate impact analysis employed by the 
Court in Griggs." 457 U.S. at 447 n.8.*> 
The Court explained that "[b]oth the House 
and Senate reports cited Griggs with 
approval, the Senate report noting that:

'Employment discrimination as 
viewed today is a . . . complex and pervasive phenomenon. Experts * 16

6The legislative history of the 1972 
amendments is relevant here because those amendments extended the protection of 
§703(a)(2) to "applicants for employment" 
(such as the respondents in the present case) as well as employees, and because the amendments extended the coverage of Title VII to federal and state employees. §§701(a),(b), and (e) , 42 U.S.C. §§2000e-
(a), (b), and (e); §717, 42 U.S.C. §2000e-
16. See Teal, 457 U.S. at 447 n.8;Franks v. Bowman Transportation Co. 424 
U.S. 747, 764 n.21 (1976); see also, id.at 796 n.18 (Powell, J., concurring in
part and dissenting in part); Albemarle. 422 U.S. at 420-21; Johnson v. Railway 
Express Aaencv. 421 U.S. 454, 459 (1975).Compare Teamsters. 431 U.S. at 354 n.39(1972 legislative history entitled to 
little if any weight in construing §703(h), which was unaffected by 1972 amendments).

490



2 2

familiar with the subject now 
generally describe the problem in 
terms of "systems” and "effects" 
rather than simply intentional 
wrongs.'"

Id. (quoting S. Rep. No. 92-415 at 5 
(1971)). See also H.R. Rep. No. 92-238 at 
8 (1971).

Congress in 1972 reiterated in even 
stronger terms than in 1964 that Title VII 
prohibited disparate impact discrimination 
as well as disparate treatment discrimi­
nation. Indeed, congressional recognition 
that "institutional" discrimination was an 
evil different from discrimination 
motivated by ill will or animus was the 
impetus for several of the more 
significant amendments.7 "[WJhere

7Senator Dominick, who sponsored the 
Nixon Administration's court-enforcement 
approach as an alternative to the proposal to give EEOC cease-and-desist powers, 
stated that "'most discriminatory treatment is institutional; subtle 
practices that leave minorities at a disadvantage.'" 118 Cong. Rec. 697 (1972) 
(quoting Wall Street Journal article).

4 9 1



23

discrimination is institutional, rather 
than merely a matter of bad faith, . . . 
corrective measures appear to be urgently 
required." S. Rep. Ho. 92-415 at 14.* 8

See also 118 Cong. Rec. 944-45 (1972)(remarks of Sen. Spong) ("a significant part of the problem today is not the simple, willful act of some employer but rather the effect of long-established practices or systems in which there may be no intent to discriminate or even knowledge that such is the effect").

8Congress in 1972 extended Title VII to federal employees, who previously could 
invoke only Civil Service Commission 
administrative remedies. This change was 
necessary because the Commission had 
erroneously "assume[d] that employment 
discrimination in the Federal Government is solely a matter of malicious intent on the part of individuals," and "ha[d] not fully recognized that the general rules 
and procedures that it had promulgated may in themselves constitute systemic barriers 
to minorities and women." S. Rep. No. 92- 415 at 14? see also. H.R. Rep. No. 92-238 
at 24. Title VII was extended to state employees for similar reasons. See H.R. 
Rep. No. 92-238 at 17 ("widespread 
discrimination against minorities exists 
in state and local government employment 
and . . . the existence of this discrimination is perpetuated by the 
presence of both institutional and overt discriminatory practices").

492



24

In ratifying Griggs. Congress 
understood that such institutional 
practices could be justified only if the 
employer discharged a heavy burden of 
showing "overriding" business necessity. 
The House report summarized Griggs as 
holding that "employment tests, even if 
valid on their face and applied in a non- 
discriminatory manner, were invalid if 
they tended to discriminate against 
minorities and the company could not show 
an overriding reason whv tests were 
necessary." H.R. Rep. No. 92-238 at 21 
(emphasis added) ? see also id. at 22 ("If 
the use of the test acts to maintain 
existing or past discriminatory imbalances 
in the job, or tends to discriminate 
against applicants on the basis of race, 
color, religion, sex or national origin, 
the employer must show an overriding 
business necessity to justify use of the

493



25

test")? id. at 8 ("showing of an 
overriding business necessity for the use 
of such action").9

Finally, in language "that could 
hardly be more explicit," Franks. 424 
U.S. at 764 n.21, the section-by-section
analyses submitted to both houses 
"confirm[ed] Congress' resolve to accept 
prevailing judicial interpretation 
regarding the scope of Title VII." Local 
28. Sheet Metal Workers v. EEOC. 478 U.S. 
421, 470 (1986). See 118 Cong. Rec. 7166, 
7564 (1972) ("present case law as
developed by the courts would continue to

9Congress did not consider the 
employer's burden to be merely that of 
articulating a legitimate reason for 
engaging in practices that systematically 
excluded minorities or women. Id. 
Senator Dominick, for instance, explained 
that under Griggs, "'employment tests, 
even if fairly applied are invalid if they 
have a discriminatory effect and can't be 
justified on the basis of business 
necessity.'" 118 Cong. Rec. 697 (1972) 
(citation omitted) (emphasis added).

494



2 6

govern the applicability and construction 
of Title VII"). As the Court concluded in 
Teal, Congress made an explicit statement 
"that in any area not addressed by the 
amendments, present case law —  which as 
Congress had already recognized included 
our then recent decision in Griggs —  was 
intended to continue to govern." 457 U.S. 
at 447 n.8.

C. The Evidentiary Standards Of 
Griggs And Its Progeny Have Been 
Uniformly Confirmed By 
Administrative Interpretations 
Of §703(a) (2).

The Court's construction of 
§703(a) (2) in Griggs is "confirmed by the 
contemporaneous interpretations of . . .
both the Justice Department and the EEOC, 
the two federal agencies charged with 
enforcement responsibility]." Local 28. 
478 U.S. at 465-66. The enforcement 
agencies' administrative guidelines on 
this subject have been construed as

495



"express[ing] the will of Congress." 
Griggs. 401 U.S. at 434; see Albemarle.
422 U.S. at 431.10

In guidelines initially adopted in 
1966 and elaborated in 1970, see Griggs. 
401 U.S. at 434 n.9, the EEOC interpreted 
§703 (a) (2) as prohibiting the use of any 
test or other selection technique that was 
discriminatory in operation unless the 
e m p l o y e r  could establish j o b ­
relatedness . H  These guidelines, as

2 7

^ B e c a u s e  the guidelines are 
consistent with the statutory language and the legislative history, they are 
"entitled to great deference." Albemarle. 
422 U.S. at 431; Griggs. 401 U.S. at 433- 
34; see also Local 28. 478 U.S. at 465-
66• Local 93, Firefighters v. City of 
Cleveland. 478 U.S. 501,518 (1986). Cf.
General Electric Co. v. Gilbert. 429 U.S. 
125, 141-45 (1976)(EEOC guidelines on sex
discrimination not followed because they 
contradicted agency's earlier positions 
and were inconsistent with Congress' plain 
intent); Espinoza v. Farah Mfg. Co.. 414U.S. 86, 93-94 (1973).

1]-EEOC Guidelines on Employee 
Selection Procedures, 35 Fed. Reg. 12333, 
12334 (1970), codified at 29 C.F.R.

496



28

revised by the EEOC in 1970 prior to the 
Court's 1971 decision in Griggs. treated 
disparate impact discrimination as an evil 
separate from disparate treatment, and 
they interpreted Title VII as prohibiting 
both forms of discrimination.

The principle of disparate or 
unequal treatment must be 
distinguished from the concepts of 
validation. A test or other 
employee selection standard- 
even though validated against job 
performance in accordance with the 
guidelines in this part —  cannot 
be imposed upon any individual or 
class protected by Title VII where 
other employees, applicants or members have not been subject to 
that standard.

35 Fed. Reg. at 12336 (29 C.F.R. 
§1607.11).* 12

§§1607.3, 1607.13 (1970) (elaborating EEOC 
Guidelines on Employment Testing 
Procedures, reprinted in CCH Empl. Prac. 
Guide U6,904 (1967)) .

12The Uniform Guidelines on Employee 
Selection Procedures, 43 Fed. Reg. 38290 
(1978), codified at 29 C.F.R. §1607 (1986) —  which superseded the EEOC Guidelines 
and were adopted by the EEOC, the Department of Justice, and other agencies

497



29

III. THE SEPARATE EVIDENTIARY ANALYSES DEVELOPED BY THE COURT REFLECT THE 
DISTINCT NATURE OF THE DISCRIMINATORY 
PRACTICES CONGRESS INTENDED TO PRO­
SCRIBE IN ;§§703(a)(1) AND 703(a)(2).
Nothing on the face of the statute or

in its legislative history supports the
Solicitor General's argument that the
§703(a)(1) evidentiary standards of
McDonnell Douglas should supplant the
§703(a) (2) evidentiary standards of
Griggs. Indeed, this Court has developed
different standards precisely because it
is necessary to take into account the

in 1978 -- similarly require the
application of disparate impact analysis 
to "any selection procedure" and embrace 
the evidentiary standards of Griggs. See 
29 C.F.R. §1607.3 Like the EEOC
Guidelines, the Uniform Guidelines 
separately prohibit both unjustified 
disparate impact and disparate treatment 
in the use of selection procedures. See 
29 C.F.R. §1607.11 ("The principles of 
disparate or unequal treatment must be 
distinguished from the concepts of 
validation").

498



30

distinctions among various kinds of 
disparate treatment cases as well as the 
basic distinction between disparate 
treatment discrimination and disparate 
impact discrimination. Moreover, with 
respect to the separate disparate 
treatment and disparate impact analyses, 
the Court has ruled that ”[e]ither theory 
may, of course, be applied to a particular 
set of facts,” Teamsters. 431 U.S. at 335 
n.15, not that the two analyses are 
functionally indistinguishable.

A. *The Court Has Articulated Evidentiary Standards For Analyzing Disparate Treatment 
Claims Under Section 703(a)(1).

The Court has articulated several 
methods of analyzing disparate treatment 
claims under §703(a)(l). The proper 
analysis varies depending upon the nature 
of the claims and the evidence presented 
in each case.

499



3 1

1. individual Disparate Treatment.
The McDonnell Douglas model for 

individual disparate treatment cases is 
"intended progressively to sharpen the 
inquiry into the elusive factual question 
of intentional discrimination," Texas 
Department of Community Affairs v, 
Burdine. 450 U.S. 248, 254 n.8 (1981),
when direct evidence of discrimination is 
absent. Thurston. 469 U.S. at 121. Under 
the individual disparate treatment 
analysis, the plaintiff must establish a 
prima facie case through circumstantial 
evidence —  by showing, for example, that 
he or she belongs to a group protected by 
Title VII; that he or she applied and was 
qualified? that the application was 
rejected? and that the position remained 
open after the rejection. McDonnell 
Douglas. 411 U.S. at 802. "The prima 
facie case . . . eliminates the most

500



32

common non-discriminatory reasons for the 
plaintiff's rejection . . . [and] raises
an inference of discrimination only 
because we presume these acts, if 
otherwise unexplained, are more likely 
than not based on the consideration of 
impermissible factors.'" Burdine, 450 
U.S. at 253-55 (quoting Furnco
Construction Coro, v. Waters. 438 U.S. 
567, 577 (1978)).

A prima facie case of individual 
disparate treatment, however, is 
"insufficient to shift the burden of 
proving a lack of discriminatory intent to 
the defendant." Watson 108 S. Ct. at 2793 
(Blackmun, J., concurring in part and 
concurring in the judgment) (original 
emphasis). Such a prima facie showing
merely shifts to the employer the burden 
of producing admissible evidence that the 
plaintiff was rejected for a legitimate,

501



3 3

nondiscriminatory reason, thereby 
rebutting the presumption and raising a 
genuine issue of fact as to whether the 
employer discriminated against the 
plaintiff. Burdine. 450 U.S. at 254-55.
As a result, the employer "frames[s] the 
factual issue with sufficient clarity so 
that the plaintiff will have a full and 
fair opportunity to demonstrate pretext." 
Id-

2. Direct Evidence of Intentional
Discrimination.
" [T]he McDonnell Douglas test is 

inapplicable where the plaintiff presents 
direct evidence of discrimination." 
Thurston. 4 69 U.S. at 121? see Teamsters. 
431 U.S. at 358 n.44. Where plaintiff's 
direct evidence of discrimination is 
accepted, an employment practice is 
established as "discriminatory on its 
face" without further need to show a

502



34

discriminatory intent. Thurston. 469 U.S. 
at 121 (policy conditioning transfer 
rights on age of airline captains is 
discriminatory on its face under the Age 
Discrimination in Employment Act) ; 
Manha rt. 435 U.S. at 708 (policy
requiring female employees to make larger 
contributions to pension fund than male 
employees is discriminatory on its face 
under §703(a)(l))? Phillips v, Martin 
Marietta Coro.. 400 U.S. 542 (1971) (per
curiam) (policy of hiring men but not 
women with pre-school age children is 
discriminatory on its face under 
§703(a)(1)) .

Where plaintiffs' direct evidence 
establishes disparate treatment, the 
burden shifts to the employer to justify 
the practice by proving the applicability 
of any statutory immunities or affirmative 
defenses. See Thurston. 469 U.S. at 122-



35

25 (rejecting employer's statutory bona 
fide occupational qualification and bona 
fide seniority system defenses)? Manhart. 
435 U.S. at 716-17 (rejecting cost 
justification defense as unavailable in a 
disparate treatment case) ; Phillips. 400 
U.S. at 54 4 (remanding for evidence on 
bona fide occupational qualification 
defense).

3. Pattern or Practice of Inten­
tional Discrimination.
In class actions and other cases 

involving claims of widespread intentional 
discrimination against members of a race, 
sex, or ethnic group, statistical or other 
evidence of a "pattern or practice" of 
disparate treatment is sufficient to 
establish a prima facie violation in the 
absence of direct evidence of intentional 
discrimination. Teamsters. 431 U.S. at 
360; Franks. 424 U.S. at 751. "The burden

504



3 6

then shifts to the employer to defeat the 
prima facie showing of a pattern or 
practice by demonstrating that 
[plaintiffs'] proof is either inaccurate 
or insignificant.” Teamsters. 431 U.S. at 
360. See also Hazelwood School District v. 
United States. 433 U.S. 299, 310 (1977).
If the employer fails to rebut the prima 
facie case, the court concludes that a 
violation has occurred and enters 
appropriate classwide declaratory and 
injunctive relief without hearing further 
evidence. Teamsters. 431 U.S. at 361.

B. The Court Has Articulated 
Separate Evidentiary Standards For Analyzing Disparate Impact 
Claims Under Section 703(a)(2).

In enacting §703(a)(2), "Congress 
required 'the removal of artificial, 
arbitrary, and unnecessary barriers to 
employment when the barriers operate 
invidiously to discriminate on the basis



37

of racial or other impermissible 
classification.'" Dothard v. Rawlinson, 
433 U.S. 321, 328 (1977) (quoting Griggs, 
401 U.S. at 431).

The gist of [a 5703(a)(2)] 
claim . . . does not involve ana s s e r t i o n  of p u r p o s e f u l  
discriminatory motive. It is
asserted, rather, that these facially neutral qualifications 
work in fact disproportionately to 
exclude women from eligibility 
for employment. . . [T]oestablish a prima facie case of 
discrimination, a plaintiff need 
only show that the facially 
neutral standards in question 
select applicants for hire in a significantly discriminatory 
pattern.

Since it is shown that the 
employment standards are 
discriminatory in effect, the 
employer must meet "the burden of 
showing that any given requirement 
[has] . . .  a manifest relation 
to the employment in question." 
Griggs v. Duke Power Co., 401 U.S. 
at 432. If the employer proves that the challenged requirements 
are job related, the plaintiff may 
then show that other selection 
devices without a similar 
discriminatory effect would also 
'serve the employer's legitimate 
interest in 'efficient and 
trustworthy workmanship,'

506



38

Albemarle Paper Co. v. Moody, 422 
U.S. at 425 quoting McDonnell 
Douglas Corn, v. Green. 411 U.S.
792, 801.

Dothard. 433 U.S. at 329-30.13
When a plaintiff proves that a 

facially neutral practice has significant 
adverse impact, the plaintiff has 
established the very conduct that 
§703(a)(2) prohibits. Watson. 108 S. Ct. 
at 2794 (Blackmun, J., concurring in part 
and concurring in the judgment) ("unlike a 
claim of intentional discrimination, which 
the McDonnell Douglas factors establish 
only by inference, the disparate impact 
caused by an employment practice is

13This analysis is typically used in class actions under Rule 23, Fed. R. Civ. P., and government pattern or practice 
actions under §707 of Title VII, 42 U.S.C. §2000e-6, because disparate impact 
discrimination is by its nature broadly 
applicable to a group. However, the 
analysis has also been utilized in cases 
seeking relief only for individual 
plaintiffs. See, e.g., Teal, 457 U.S. at 442-44? Lowe v. City of Monrovia. 775 F.2d 
998, 1004 (9th Cir. 1985).

507



39

directly established by the numerical 
disparity") ? see Satty, 434 U.S. at 144
("Griggs held that a violation of 
§703 (a) (2) can be established by proof of 
a discriminatory effect"). Similarly, in 
both the direct evidence (Thurston) and 
pattern or practice intentional 
discrimination (Teamsters) models, the 
prima facie case directly establishes the 
discrimination prohibited by §703(a)(l). 
The direct evidence and pattern or 
practice models, like the disparate impact 
model, were developed for analyzing 
evidence concerning employment practices 
and policies that affect large numbers of 
people on a classwide basis.

The McDonnell Douglas individual 
disparate treatment model, on the other 
hand, was developed to analyze the very 
different kinds of evidence typically 
presented in a case involving a discrete

508



40

act of intentional discrimination against 
a single individual. A prima facie 
showing in a McDonnell Douglas case is not 
comparable in either its nature or its 
effect to a prima facie showing in a 
Griggs disparate impact case. A McDonnell 
Douglas prima facie case does not in 
itself establish the intentional 
discrimination prohibited by 5703(a)(1); 
it only "eliminates the most common 
nondiscriminatory reasons for the 
plaintiff's rejection." Burdine. 450 U.S. 
at 255; see Teamsters. 431 U.S. at 358 
n. 44.

This Court has uniformly held that, 
once the plaintiff establishes a prima 
facie disparate impact case under 
5703(a)(2), the burden shifts to the 
employer to prove that the challenged 
practice is justified. See, e.g., Teal, 
457 U.S. at 446 ("employer must . . .



4 1

demonstrate that any given requirement 
[has] a manifest relationship"); New York 
City Transit Authority v. Beazer. 440 U.S. 
568, 587 (1979) (prima facie case
"rebutted by [employer's] demonstration 
that its narcotics rule . . . 'is job
related'")? Dothard, 433 U.S. at 329
(employer must "prov[e] that the
challenged requirements are job related")? 
Albemarle. 422 U.S. at 425 (employer has 
"burden of proving that its tests are 'job 
related'"); Griggs. 401 U.S. at 431, 432
("The touchstone is business necessity"; 
"Congress has placed on the employer the 
burden of showing that any given 
requirement must have a manifest
relationship to the employment in 
question"); see also Watson. 108 S. Ct. at 
2794 (Blackmun, J . , concurring in part and 
concurring in the judgment).

510



42

While it is true that an evidentiary 
burden may be either one of persuasion or 
one of production, this Court in Title VII 
disparate impact cases has always imposed 
on the employer the burden to persuade the 
trier of fact of its justification for 
using practices that have a discriminatory 
impact. Indeed, as petitioners here 
concede, see Brief for Petitioners at 42, 
the employer has the burden of 
demonstrating business necessity as an 
"affirmative defense to claims of 
violation" of §703 (a)(2). Guardians 
Association v. Civil Service Commission. 
463 U.S. 582, 598 (1983) (White, J., 
announcing the Court's judgment and 
delivering an opinion joined by Rehnquist, 
J.) (Title VI case).

In trying to force the Griggs 
analysis into the McDonnell Douglas 
formula, the Solicitor General ignores the

511



43

Court's repeated admonitions that 
McDonnell Douglas does not provide the 
proper model for analyzing all Title VII 
claims.14 In an individual disparate 
treatment case, it is appropriate to 
impose a minimal burden of production on 
the employer because the plaintiff's prima 
facie showing is itself "not onerous," 
Burdine. 450 U.S. at 253, and does not in 
itself establish a violation of 
§703(a)(l). That same slight burden would 
be inappropriate in a disparate impact 
case, where the prima facie showing 
usually includes substantial statistical

14See. e.q.. McDonnell Douglas. 411 
U.S. at 802 n.13 ("The facts necessarily 
will vary in Title VII cases, and the 
specification . . .  of the prima facie 
proof required from the complainant in 
this case is not necessarily applicable in 
every respect to differing factual 
situations")? Teamsters, 431 U.S. at 358 
("Our decision in (McDonnell Douglas! . . 
. did not purport to create an inflexible 
formulation"); Furnco. 438 U.S. at 575 
(McDonnell Douglas formulation "was not intended to be an inflexible rule").

512



44

evidence of adverse impact and constitutes 
direct evidence of a violation of 
§703(a)(2) .

C. The Griggs Disparate Impact 
Analysis Is Analogous To The 
Teamsters And Thurston Disparate 
Treatment Analyses.

The Solicitor General's theory fails 
on its own terms. If there is a need 
analogize disparate impact analysis to 
some disparate treatment mode of proof, 
amici submit that the Teamsters ’’pattern 
or practice" model and the Thurston 
"direct evidence" model provide more 
appropriate analogies than the McDonnell 
Douglas "individual case" model. In the 
Teamsters and Thurston models, the 
allegedly discriminatory conduct is not a 
single, isolated decision affecting only 
one individual, but rather a broadly 
applicable practice of intentional 
discrimination affecting a class as a

513



4 5

whole. The purpose of these analyses is 
comparable to the purpose of the disparate 
impact model, with its parallel focus on 
"artificial, arbitrary, and unnecessary 
barriers to employment." Griggs. 401 U.S. 
at 431. In the Solicitor General's terms, 
c l a s s w i d e  d i s p a r a t e  treatment 
discrimination is the "functional 
equivalent" of disparate impact 
discrimination.

Because of the similarity in the 
practices analyzed, the evidentiary models 
are also similar. In the Teamsters and 
Thurston models, plaintiffs establish a 
prima facie case by introducing 
statistical or other evidence of a 
"standard operating procedure" of 
classwide disparate treatment, Teamsters. 
431 U.S. at 336, or by proving the 
classwide application of a facially 
discriminatory policy. Thurston. 469 U.S.

514



4 6

at 121. In the Griggs disparate impact 
model, plaintiffs establish a prima facie 
case by marshalling comparable evidence of 
a practice affecting an entire class of 
employees or applicants. Moreover, in the 
Teamsters and Thurston disparate treatment 
models, as in the Griggs disparate impact 
model, proof of a prima facie case shifts 
the burden of persuasion, not the burden 
of production, to the employer. See 
Teamsters. 431 U.S. at 360; Thurston. 469 
U.S. at 122-25. In all three models, 
plaintiff has borne his burden of proof to 
establish a violation of Title VII; 
defendant then has the burden of proving a 
justification, establishing what is, in 
essence, an affirmative defense.

In short, there is no need to change 
the Griggs disparate impact analysis to 
make it conform to the appropriate 
disparate treatment analysis. Existing

515



47

evidentiary standards for analyzing 
disparate impact discrimination are 
already closely analogous to the 
evidentiary standards for analyzing 
disparate treatment discrimination under 
Teamsters and Thurston.

IV. OVERRULING THE EVIDENTIARY STANDARDS 
OF GRIGGS AND ITS PROGENY WOULD BE 
CONTRARY TO THE REMEDIAL PURPOSE OF 
TITLE VII.
The Solicitor General argues, in 

essence, that Griggs and its progeny 
should be overruled in order to make the 
employer's burden in a Griggs disparate 
impact case conform to the employer's 
burden in a McDonnell Douglas individual 
disparate treatment case. Overruling the 
Court's prior decisions in this manner, 
however, would drastically alter the 
nature of disparate impact analysis under 
§703(a)(2). The employer's burden would 
be reduced to such an extent that all but

516



4 8

the most unimaginative employers —  unable 
even to articulate a legitimate reason for 
practices having a significant adverse 
impact —  would be able to rebut a showing 
of disparate impact discrimination, no 
matter how compelling. The result would 
be an effective repeal of §703(a)(2).

The Court in Griggs identified Title 
VII' s fundamental purpose as "the removal 
of artificial, arbitrary, and unnecessary 
barriers to employment when the barriers 
operate invidiously to discriminate on the 
basis of racial or other impermissible 
classification." 401 U.S. at 431. The 
statute "police[s]" not only the problem 
of intentional discrimination through the 
disparate treatment analyses available 
under §703(a)(l), but also "the problem of 
subconscious stereotypes and prejudices,"
Watson. 108 S. Ct. at 2786 (part IIB), and 
"built-in practices preserved through

517



49

form, habit or inertia." S. Rep. No. 88- 
867 at 11. The latter purpose derives 
from the terms of §703(a)(2) and, as 
Congress recognized, is enforced by 
application of the disparate impact 
analysis articulated in Griggs. The 
Solicitor General's proposal to overrule 
the evidentiary standards of Griggs and 
its progeny is contrary to Title VII' s 
fundamental purpose.

The Solicitor General would have the 
Court transmute the employer's burden of 
persuasion in a Griggs disparate impact 
case into the burden of production imposed 
on an employer in a McDonnell Douglas 
individual disparate treatment case —  a 
feat of judicial alchemy that would 
drastically change the nature of disparate 
impact analysis under §703(a)(2). The 
employer's burden in such cases of proving 
an "overriding business necessity," as

518



5 0

Congress termed it, is appropriately high 
because the challenged practice has been 
shown to violate §703(a)(2) as a prima 
facie matter. The Solicitor General's 
proposed standard, in contrast, would 
declare such practices lawful whenever the 
employer could simply articulate a 
"legitimate, nondiscriminatory reason" for 
its actions? the employer "need not [even] 
persuade the court that it was actually 
motivated by the proffered reason[ ]." 
Burdine. 450 U.S. at 254. The Solicitor 
General would then permit the plaintiff to 
introduce contrary evidence, but would put 
the risk of nonpersuasion of business 
necessity on the plaintiff. Failing this, 
all the plaintiff then could do to abate 
the exclusionary practice would be to 
present evidence of alternative selection 
devices. As a result, the plaintiff would 
have not only the burden of proving a



51

prima facie case of disparate impact, but 
also the burden of disproving business 
necessity.

The scheme proposed by the Solicitor 
General would thwart the specific remedial 
purpose of §703 (a) (2) by making it 
virtually impossible for a plaintiff to 
prevail on a claim of disparate impact 
discrimination. As a practical matter, 
§703(a)(2) would be repealed as an 
independent substantive provision, and the 
evils to which that provision is addressed 
—  "the problem of subconscious 
stereotypes and prejudices" and "built-in 
practices preserved through form, habit or 
inertia" —  would go unremedied.

Ignoring that the Griggs disparate 
impact standard directly reflects 
statutory language and congressional will, 
the Solicitor General attempts to justify 
its revision by raising the specter of

520



52

quotas and intrusion on managerial 
prerogatives. See Brief for the United 
States as Amicus Curiae at 25. Griggs 
itself rejected such claims, 401 U.S. at 
436, as did Congress when it ratified 
Griggs in 1972.15

Moreover, the suggestion that 
subjective selection procedures are 
impossible to validate16 is simply wrong. 
The courts have identified specific 
characteristics of valid subjective rating 
procedures, such as using specific 
guidelines for raters, rating only

1 5 C o n g r e s s i o n a l  o p p o n e n t s  
specifically objected to the 1972 
amendments on these grounds, but their 
views were not accepted. E.g. . 117 Cong. 
Rec. 32108 (1971) (comments of Rep. Rarick 
that bill would require preferential 
treatment and maintenance of racial 
balance); 117 Cong. Rec. 38402 (1971)
(comments of Sen. Allen that bill would 
infringe on discretion of state and local 
officials to select employees).

16See Brief for the United States as 
Amicus Curiae at 25 n.35; Brief for
Petitioners at 47.

521



53

observable behaviors or performance, 
requiring raters to have knowledge of job 
responsibilities, and using an evaluative 
device with fixed content that calls for 
discrete judgments.17 Subjective 
selection procedures can be and have been 
successfully validated.18 See Rose, 
Subjective Employment Practices. 25 San 
Diego L. Rev. at 87-92.

17See B. Schlei & P. Grossman, 
Employment Discrimination Law 202-05 (2ded. 1983) (collecting cases).

18See. e.g., Firefighters Inst, for 
Racial Equality v. City of St. Louis. 616 
F.2d 350, 362 (8th Cir. 1980), cert, 
denied, 452 U.S. 938 (1981) (interview and 
training simulations)? Wade v. Mississippi 
C o o p . Extension Serv.. 615 F. Supp. 1574 
(N.D. Miss. 1985) (promotional performance 
evaluation); Tillerv v. Pacific Tel. Co. , 
34 FEP Cases 54 (N.D. Cal. 1982); Wilson 
v. Michigan Bell Tel. Co.. 550 F. Supp. 
1296 (E.D. Mich. 1982) (formal assessment procedures).

522



54

V. THE FIRST AND THIRD QUESTIONS PRESENTED IN THE PETITION FOR 
CERTIORARI ARE NOT PRESENTED BY THE 
FACTS OF THIS CASE.
With respect to the first question

presented in the petition (concerning the
standards for establishing a prima facie
case of disparate impact) and the third
question presented (concerning the
application of disparate impact analysis
to multicomponent selection practices),
amici rely on respondents' brief.
However, as we briefly explain, it appears
that neither question is actually
presented by the record before the Court.

As to the first question, petitioners
argue that the Ninth Circuit's reliance
upon statistics comparing cannery with
noncannery positions is erroneous because
there was no showing of an internal
promotion system. Such statistics would
be marshalled as evidence of promotional
discrimination where an employer maintains

523



5 5

an internal promotion system in which 
lower level employees are the selection 
pool for upper level positions. See. 
e.g, . Paxton v. Union National Bank. 688 
F.2d 552, 564 (8th Cir. 1982), cert,
denied. 460 U.S. 1083 (1983). However,
petitioners err in arguing that 
comparative statistics can be used only 
where there are internal promotions.

In this case, plaintiffs challenged, 
on both disparate impact and disparate 
treatment grounds, several specific hiring 
practices —  nepotism, subjectively 
evaluated selection criteria, separate 
hiring channels and word of mouth 
recruitment, a rehire preference, and a 
series of related practices involving race 
labeling, housing and messing. Plaintiffs 
presented independent statistical or other 
evidence that each of these specific 
practices had a significant adverse impact

524



5 6

on minority class members. Except for the 
rehire preference, the district court 
erroneously failed to consider the 
challenge under, or erred in applying, the 
disparate impact standard. See App. Cert. 
VI-19-VI-39; see also. Brief for the 
United States as Amicus Curiae at 20 ("The 
district court did not apply disparate 
impact analysis to the selection of 
noncannery workers generally, and there is 
therefore no finding that respondents' 
statistics did not make out a prima facie 
case under the disparate impact model"). 
The Ninth Circuit, therefore, properly 
remanded these issues to the district 
court.

The comparative statistics to which 
petitioners object were not relied upon as 
the sole evidence of the disparate impact 
of the challenged practices. The Ninth 
Circuit upheld the use of these

525



57

comparative statistics on the limited 
ground that "such statistics can serve to 
demonstrate the consequences of 
discriminatory practices which have 
already been independently established." 
App. Cert. VI-16. The comparative
statistics, which do not appear strictly 
to be necessary to establish the disparate 
impact of each of the challenged 
practices, were presented as additional 
evidence that "some practice or 
combination of practices has caused the 
distribution of employees by race." App. 
Cert. VI-18.19

190n the facts of this case, the 
Ninth Circuit correctly considered these 
statistics given the difficulty of 
establishing the available labor pool for 
the migrant and seasonal noncannery jobs 
in question, the arbitrary nature of the 
qualifications actually imposed for the 
noncannery jobs, and the fact that 
minority cannery workers were apparently 
qualified and available. The Ninth 
Circuit's unwillingness to rely on 
petitioners' generalized census data, and its reliance instead on more probative

526



58

As to the third question presented, 
petitioners argue that only "cumulative" 
evidence of the impact of several 
employment practices was presented. For 
the reasons stated above, we believe 
petitioners have misstated the record: 
Specific, identified hiring practices were 
challenged, and both practice-specific 
evidence and cumulative statistical 
evidence were presented below.

However, if this were a case in which 
a plaintiff challenged a multicomponent 
employment practice, the adequacy of 
cumulative evidence of disparate impact 
would depend upon particular factual 
circumstances. If the practice consisted 
of a series of sequential steps, e.q. ,

practice-specific evidence of disparate 
impact coupled with respondents' 
comparative statistics, are understandable 
and proper in view of the record in this case.

527



59

Teal. 457 U.S. at 443-44 (a qualifying
written examination followed by 
consideration of other criteria) , the 
plaintiff might attack one or more steps, 
or the plaintiff might attack the process 
as a whole. While a plaintiff challenging 
one or more discrete steps in the process 
typically introduces evidence of the 
disparate impact of each challenged step, 
a plaintiff challenging the process as a 
whole is not required to introduce such 
evidence.20

Moreover, a plaintiff challenging a 
multicomponent practice in which the 
employer combines consideration of several 
factors, e.g.. Teal. 457 U.S. at 444
(employees promoted from a list of

20See Green v. USX Coro. . 84 3 F.2d
1511, 1524 (3rd Cir. 1988); Seqar v. 
Smith. 738 F.2d 1249, 1271 (D.C. Cir.
1984). See also. 29 C.F.R. §1607.16Q 
(Uniform Guidelines apply to any "measure [or] combination of measures").

528



6 0

successful test takers based on an amalgam 
of work performance, recommendations and 
seniority), should not be required to 
identify and present specific disparate 
impact evidence as to each factor. Title 
VII does not prohibit discrete 
discriminatory criteria in the abstract, 
but as "actually applied." Albemarle. 422 
U.S. at 433. If an employer uses an 
amalgam of factors as a practice, and that 
practice has a disparate impact, the 
plaintiff should not be required to go 
through the academic exercise of 
disentangling the factors in order to 
ascertain which particular factors caused 
the disparate impact of the practice as a 
whole. That burden should be borne by the 
employer.21

21It is the employer who presumably 
has an interest in distinguishing among several factors that produce a disparate 
impact in order to isolate the 
discriminatory factors and to save the

529



6 1

Amici respectfully submit that the 
first and third questions presented in the 
petition for certiorari are not actually 
presented by the facts of this case, and 
that those questions should not be decided 
on this record.

rest. It is the employer who may wish to conduct separate validation studies of the 
factors. Moreover, it is the employer who 
has the obligation under administrative guidelines to "maintain and have available 
records or other information showing which 
components [of a multicomponent selection 
procedure] have an adverse impact." Uniform Guidelines on Employee Selection 
Procedures, 29 C.F.R. §1607.15 (a) (2) (employers with 100 or more employees 
should maintain component data if overall practice has adverse impact or for two 
years after impact eliminated). See Brief 
for the United States as Amicus Curiae at 
22 ("certainly if [multiple] factors 
combine to produce a single ultimate 
selection decision and it is not possible 
to challenge each one, the decision may be challenged (and defended) as a whole").

530



62

CONCLUSION
The order of the Ninth Circuit 

remanding the case for further proceedings 
should be affirmed.

Respectfully Submitted,

JULIUS LEVONNE CHAMBERS CHARLES STEPHEN RALSTON 
RONALD L. ELLISNAACP Legal Defense and Educational Fund, Inc.
BILL LANN LEE*
PATRICK O. PATTERSON, JR. 
THEODORE M. SHAW
NAACP Legal Defense and 
Educational Fund, Inc.

ANTONIA HERNANDEZ E. RICHARD LARSON JOSE ROBERTO JUAREZ, JR.Mexican American Legal Defense 
and Educational Fund

RUBEN FRANCO KENNETH KIMERLINGPuerto Rican Legal Defense 
and Education Fund

Counsel for Amici Curiae
*Counsel of Record
November 1988.

531





No. 87-1387

In the

Supreme (tfnurt of tlje lituitEii g’tjitco
October Term , 1988

Wards Cove Packing Co m pa n y , In c .,
Castle &  Cooke, In c .,

P e tit io n e rs ,

Frank Atonio, e t a l . ,
R e sp o n d e n ts .

BRIEF OF AMICUS CURIAE 
NATIONAL ASSOCIATION FOR THE 

ADVANCEMENT OF COLORED PEOPLE * 15

Grover G. H ank ins*
G e n e ra l C o u n se l  

Sam uel  M. Walters 
A s s is ta n t G en era l C o u n se l  

N atio nal  A ssociation  for the  
A dvancem ent  of Colored  People 

4805 Mt. Hope Drive 
Baltimore, Maryland 21215 
(301)486-9191 

and
A lfred W. Blumrosen
15 Washington Street 
Newark, New Jersey 07102 
(201) 648-5332
C o u n se l f o r  A m ic u s  C u ria e

*Counsel of Record 

November 4, 1988

533



i

TABLE OF CONTENTS

PAGE

TABLE OF AUTHORITIES................................................  iii

INTEREST OF AMICUS CURIAE.................................... 1

SUMMARY OF ARGUMENT............................................  2

ARGUM ENT...........................................................................  4

I. Title VII unqualifiedly prohibits segregation of 
employees or applicants which deprive or tend to 
deprive any individual of employment opportu­
nities or otherwise adversely affect his status 
because of ra ce ........................................................  4

II. The facts found by the District Court establish
segregation of workers by race by the employer 5

III. The combination of segregated recruiting and
hiring channels, segregated job assignments, and 
refusal to consider minorities for promotion or 
transfer to white jobs establish a violation of 
Title V II .....................................................................  6

IV. The so-called "over representation” of minori­
ties in lower paying jobs, plus their exclusion 
from higher paying white jobs, does constitute 
illegal segregation under Title V II.......................  8 V.

V. There can be no "business necessity” justifica­
tion for maintaining job segregation................... 11

534



11

VI. This Court should affirm the holding of the 
Court of Appeals on the ground that illegal seg­
regation has been established rather than dismiss 
the writ as improvidently granted.................... 15

CONCLUSION.............................................................  16

PAGE

535



Ill

TABLE OF AUTHORITIES

Cases: PAGE

A lb e m a r le  P a p e r  C o . v. M o o d y , 422 U.S. 405
(1974)...................................................................... 4, 10, 13

C o n n e c tic u t  v. T e a l , 457 U.S. 440 (1982)...........................  11

C o rn in g  G la s s  W o rk s  v. B re n n a n , 417 U.S. 189 (1974). 12

D o th a r d  v. R a w lin s o n ,  433 U.S. 321 (1977)..................  13

F o r d  M o to r  C o . v. E E O C , 458 U.S. 219 (1982).............. 10

F u rn c o  C o n s tr u c t io n  C o . v. W a te r s , 438 U.S. 567
(1978).......................................................................... 11, 13

G rig g s  v. D u k e  P o w e r  C o . ,  420 F.2d 1225 (4th Cir.
1970), reversed in part, 401 U.S. 424 (1971)..............  7

G rig g s  v. D u k e  P o w e r  C o . ,  401 U.S. 424 (1971)......... 4, 7, 10

J o h n so n  v. S a n ta  C la ra  C o u n ty  T r a n s p o r ta tio n  A g e n c y ,
107 S. Ct. 1442 (1987)................................................. 10

L o c a l  189 , U n ite d  P a p e r m a k e r s  v. U n ite d  S ta te s ,  416 
F.2d 980 (5th Cir. 1969).............................................  9

L o s  A n g e le s  W a te r  a n d  P o w e r  C o . v. M a n h a r t, 435 U.S.
702 (1978)................................................................... 11

M c D o n n e l l  D o u g la s  v. G r e e n , 411 U.S. 732 (1973) . .7, 12, 14

M ille r  v. I n te r n a tio n a l  P a p e r  C o . ,  408 F.2d 283 (5th Cir.
1969)............................................................................ 9

P h il l ip s  v. M a r tin -M a r ie t ta  C o r p . ,  400 U.S. 542 (1971) 11

T e x a s  D e p a r tm e n t  o f  C o m m u n ity  A f f a i r s  v. B u r d in e ,
450 U.S. 248 (1981)....................................................  8

536



PAGE

T e a m ste rs  v. U n ite d  S ta te s ,  431 U.S. 333 (1977) . .4, 7, 11, 12

U n ite d  S ta te s  v. B e th le h e m  S te e l  C o r p . ,  446 F.2d 652 (2d 
Cir. 1971)....................................................................  9, 13

U n ite d  S ta te s  P o s ta l  S e rv ic e  v. A ik e n s , 460 U.S. 711
(1983).......................................................................... 8, 14

U n ite d  S te e lw o r k e r s  v. W e b e r , 443 U.S. 193 (1979) . . . .  10

W a tso n  v. F o r t  W o r th  B a n k  a n d  T ru s t, 487 U .S.___ ,
108 S. Ct. 2777 (1988)................................................  10

Statutes:
42 U.S.C. Sec. 2000e, e t s e q . ,  Civil Rights Act of 1964. 4

Regulations:
43 Fed. Reg. 19,260, 19,269 (May 4, 1978).................... 3

Uniform Guidelines on Employee Selection Procedures- 
1978, 29 C.F.R. Sec. 1607.4(C)(1)............................. 11

Other Authorities:
Blumrosen, T h e  L e g a c y  o f  G rig g s:  S o c ia l  P r o g r e s s  a n d  

S u b je c t iv e  J u d g m e n ts , 63 Chi. Kent L. Rev. 1 (1986) 16

Blumrosen, S e n io r i ty  a n d  E q u a l  E m p lo y m e n t  O p p o r tu ­
n ity :  A  G lim m e r  o f  H o p e ,  23 Rutgers L. Rev. 268 
(1969).......................................................................... 4, 9

Blumrosen, S tra n g e r s  in P a r a d is e :  G r ig g s  v. D u k e  P o w e r  
C o . a n d  th e  C o n c e p t  o f  E m p lo y m e n t  D is c r im in a tio n ,
71 Mich. L. Rev. 59 (1972).......................................  12

iv

537



IN THE

gatprcmc (Jdurt of tl]t UtiitEb IMate#
October Term, 1988 

No. 87-1387

Wards Cove Packing Company, Inc., 
Castle & Cooke, Inc.,

Petitioners,
—v.—

Frank Atonio, et al.t
Respondents.

BRIEF OF AMICUS CURIAE 
NATIONAL ASSOCIATION FOR THE 

ADVANCEMENT OF COLORED PEOPLE

INTEREST OF AMICUS CURIAE

Amicus curiae National Association for the Advancement of 
Colored People (NAACP) is an organization dedicated to the 
furtherance of racial equality and social and economic justice in 
this country. To promote these ends, the NAACP and its mem­
bers engage in activity protected by the United States Constitu­
tion, including petitioning the government for the redress of 
grievances. The NAACP and its members throughout the 
United States for more than twenty years have assisted workers 
in utilizing Title VII of the Civil Rights Act of 1964 to challenge 
employment discrimination against minorities and women. The 
NAACP has urged the Congress to strengthen Title VII and 
other provisions of the Civil Rights Act of 1964.

538



2

Open and notorious segregation of Black workers into infe­
rior jobs was one of the hallmarks of the system of segregation 
and discrimination in the South before the Civil Rights Act was 
adopted. Because of the litigation under Title VII, many of the 
overt forms of discrimination, such as hiring from dual segre­
gated labor markets, discrimination in job assignments, and 
discriminatory refusals to allow Blacks into better paying jobs, 
have been abandoned. However, there still remain circum­
stances in which minorities are restricted today, in precisely the 
same manner as in earlier years.

For the reasons explained below, the opinions of the District 
Court and Court of Appeals, for differing reasons, may permit 
the continued existence of blatant job segregation. The 
NAACP urges this Court to correct the errors of both the Dis­
trict Court and the Court of Appeals, and to reaffirm that the 
evil of job segregation remains unlawful under Title VII.

This amicus curiae brief is filed with the consent of the par­
ties, whose letters of consent have been filed with the Clerk of 
the Court.

SUMMARY OF ARGUMENT

Certiorari was granted to consider three questions relating to 
the concept of discrimination under Title VII of the Civil Rights 
Act of 1964 that deals with neutral practices which have a “ dis­
parate impact” on minorities or women.1 This case does not * 1 2 3

I The questions presented are:
1. Does statistical evidence that shows only a concentration of 

minorities in jobs not at issue fail as a matter of law to establish dis­
parate impact of hiring practices where the employer hires for at-issue 
jobs from outside his own work force, does not promote-from-within 
or provide training for such jobs, and where minorities are not under­
represented in the at-issue jobs?

2. In applying the disparate impact analysis, did the Ninth Circuit 
improperly shift the burden of proof to petitioners?

3. Did the Ninth Circuit commit error in allowing plaintiffs to chal­
lenge the cumulative effect of a wide range of non-racially motivated 
employment practices under the disparate impact model?

539



3

involve such practices. It involves racial segregation in hiring, 
job assignments, and promotions against Filipino and Alaskan 
Native workers in favor of whites.2 The employers hired minor­
ity workers through separate procedures and channels from 
those used to hire whites. They assigned minority workers to 
lower paying jobs and refused to consider them for promotion 
or transfer to white jobs. Whites were hired through separate 
procedures and channels from those used to hire minorities into 
higher paying jobs and were separately housed and fed from 
minorities. These facts, established by the District Court, con­
stitute racial segregation in violation of the statute.

The courts below did not recognize the job segregation of 
minorities as a violation of Title VII. The District Court dis­
counted evidence of segregation of minorities in low paying 
jobs as “over-representation” of minorities. It then analyzed 
several employment practices separately but never examined the 
interaction between segregated hiring, job assignment, and the 
refusal to consider minorities for promotion or transfer. The 
Court of Appeals analyzed employment procedures under the 
disparate impact principle and reversed the District Court. In 
applying the impact principle, it recognized a “business neces­
sity” defense to the maintenance of job segregation. This is not 
the law. Job segregation is illegal.

This court granted Certiorari to consider questions relating to 
the application of impact theory. However, the facts— 
segregation in hiring, job assignments, and refusal to transfer 
or promote minorities—make this case an inappropriate vehicle 
to resolve questions concerning disparate impact theory. The 
District Court analysis was clearly erroneous, and the Court of 
Appeals committed error in allowing a “ business necessity” 
defense to segregation.

Since the Court of Appeals found for the employees, albeit 
on an erroneous theory, its judgment should be affirmed. This

2 Both Alaskan Natives and persons of Filipino descent are considered 
as being in separate racial groups from whites for the purposes of the 
Federal reporting policies. See 43 Fed. Reg. 19,260, 19,269 (May 4, 
1978).

540



4

Court should remand, making clear that the segregation which 
has been established in this case is illegal and cannot be 
defended on grounds of business necessity.

THE ARGUMENT

I. TITLE VII UNQUALIFIEDLY PROHIBITS SEGREGA­
TION OF EMPLOYEES OR APPLICANTS WHICH 
DEPRIVE OR TEND TO DEPRIVE ANY INDIVIDUAL 
OF EMPLOYMENT OPPORTUNITIES OR OTHER­
WISE ADVERSELY AFFECT HIS STATUS BECAUSE 
OF RACE

The language of Sec. 703(a)(2), makes it an unlawful employ­
ment practice for an employer to:

**. . . limit, segregate or classify his employees or appli­
cants for employment in any way which would deprive or 
tend to deprive any individual of employment opportuni­
ties or otherwise adversely affect his status as an employee, 
because of such individual’s race . . . ”

Overt and current job segregation on the basis of race has 
never been defended before this Court. In the earliest cases 
under Title VII, employers admitted pre-act segregation against 
blacks, but stated that segregation had ended, and the post-act 
situation was justified by seniority or testing practices.3 None of 
the cases previously before this Court involved an employer 
who hired minorities through recruiting practices separate from 
those used to hire whites, assigned them to lower paying jobs 
and then, as a matter of general policy, refused to consider 
them for promotion or transfer to the better “white” jobs. The 
refusal to consider minorities for promotion out of segregated 
jobs is illegal per se as maintaining segregation.

3 G r ig g s  v. D u k e  P o w e r  C o . ,  401 U.S. 424 (1971); A lb e m a r le  P a p e r  
C o .  v. M o o d y ,  422 U.S. 405 (1974); T e a m s te r s  v. U n i t e d  S ta te s ,  431 
U.S. 333 (1977); Blumrosen, S e n io r i t y  a n d  E q u a l  E m p l o y m e n t  O p p o r ­

tu n i t y :  A  G l i m m e r  o f  H o p e ,  23 Rutgers L. Rev. 268 (1969).

541



5

542

II. THE FACTS FOUND BY THE DISTRICT COURT 
ESTABLISH SEGREGATION OF WORKERS BY RACE 
BY THE EMPLOYER

The facts found by the District Court establish that the defen­
dants did segregate employees and applicants in ways which 
deprived them of employment opportunities because of their 
race.4 The District Court found:

(1) Employees were segregated by race, with whites holding 
better, higher paying jobs and Alaskan Natives and Filipinos 
holding lower paying laborer and cannery jobs.5

(2) Filipinos and Alaskan Natives have been recruited from 
Alaskan Native communities and a local union in Seattle com­
posed mainly of workers of Filipino extraction.6

(3) They have been assigned to do low paid labor and can­
nery work.7 8

(4) The jobs done by these workers are characterized as “Fil­
ipino jobs,” or “ Eskimo jobs,” or “ native” jobs.*

4 References throughout are to the opinions appearing in appendices 
to the Joint Appendix. For convenience, the reference to Joint Appen­
dix is omitted when referring to the opinions of the Courts below. 
Appendix I contains the opinion of the District Court, which also 
appears in 34 E P D  1 34,437. Appendix III contains the Court of 
Appeals’ first opinion of Aug. 16, 1985, also appearing in 768 F.2d 
1120 (9th Cir. 1985). Appendix V contains the e n  b a n c opinion of the 
Court of Appeals of Feb. 23, 1987, appearing in 810 F.2d 1477 (9th 
Cir. 1987). Appendix VI contains the decision of the panel of the 
Court of Appeals on remand from the Court e n  b a n c , of Sept. 2, 1987, 
appearing in 827 F.2d 439 (9th Cir. 1987) as to which certiorari has 
been granted.

5 District Court findings #105 (1-36), #109 (1-38). See also A t o n i o  (VI- 
18) 827 F.2d 439 at 444, ("The statistics show only racial stratification 
by job category.”) S e e  a l s o ,  A t o n i o  (III-9) 768 F.2d 1120, 1124.

6 The District Court findings #90 (1-32), #105 (1-36), #109 (1-38); A t o ­

n io (V-6-7) 810 F.2d at 1479.

7 I d .

8 District Court finding #135-141 (1-76-80); A t o n i o  (VI-33) 827 F.2d at 
447.



6

(5) The employer does not consider members of the plaintiff 
class for employment, promotion, or transfer to the higher pay­
ing jobs held by whites, regardless of their possible qualifica­
tions.9

(6) Whites have been recruited primarily from the lower 48 
states for the higher paying jobs.10 11

(7) The whites had superior residential and eating facilities."

(8) The pay of white workers, both unskilled and skilled, was 
higher than that of Alaskan Native and Filipino workers.12

(9) When Filipino and Alaskan Native workers sought to 
apply for the white jobs, they were brushed off with a variety of 
excuses relating to the timeliness of their applications.13

III. THE COMBINATION OF SEGREGATED RECRUIT­
ING AND HIRING CHANNELS, SEGREGATED JOB 
ASSIGNMENTS, AND REFUSAL TO CONSIDER 
MINORITIES FOR PROMOTION OR TRANSFER TO 
WHITE JOBS ESTABLISH A VIOLATION OF TITLE 
VII

The employers devised segregated labor markets. For the 
higher paying jobs, they recruited whites from the lower 48. For 
the low paying jobs, they recruited minorities from the local vil­
lages or the Filipino union. They assigned minorities to the

9 District Court finding #86 (1-30), #89(1-31), #110(1-39). White jobs 
are filled from Seattle and Astoria. District Court findings #86 (1-30), 
#112 (1-39). The employers do not promote from within. District Court 
Finding #112 (1-39). “ Defendant’s cannery workers and laborers do 
not form a labor pool for other jobs at defendant’s facilities.” District 
Court finding #110 (1-39). S e e  a ls o Question Presented #1 on which 
certiorari was granted, note 1, s u p r a .

10 District Court finding #86 (1-30).

11 District Court findings #148, 149 (1-81-84).

12 A l o n i o  (111-91) 768 F.2d at 1124.

13 District Court findings #150-172 (1-84-94). 543



7

lower paying cannery and labor jobs, and whites to the higher 
paying jobs. They did not permit minority employees promo­
tion or transfer to better jobs.14 Thus the employers segregated 
the plaintiff class through its hiring practices, and maintained 
that segregation through job assignment practices and through 
refusal to consider minorities for promotion and transfer. 
These facts were all found by the District Court. This blanket 
refusal to consider minorities for better jobs locked them into 
the lower paying jobs for which they had been hired.15

This obvious violation of Title VII was obscured because of 
the efforts of the courts below to fit this case of brutal segrega­
tion into the framework of disparate impact or disparate treat­
ment.16 The concept of disparate impact was intended to 
address facially neutral practices.17 The concept of disparate 
treatment was intended to order the proofs in an individual case 
of discrimination.18 But these categories were never intended to 
be exclusive.19 They were not developed in, nor have they been 
applied to, cases of current work force segregation.

The emphasis on the proof process can obscure the ultimate 
issues of discrimination. In United States Postal Service v.

14 The concentration of minorities in the lower paying jobs and the 
denial of any consideration for promotion or transfer establishes a vio­
lation of Title VII, even though some whites were also in the lower 
paying jobs. T e a m s te r s  v. U n i t e d  S ta t e s ,  431 U.S. 333 at 337-338 
(1977). S e e  a l s o ,  G r ig g s  v. D u k e  P o w e r  C o . ,  420 F.2d 1225 at 1247 (4th 
Cir. 1970), r e v e r s e d  in  p a r t ,  401 U.S. 424 (1971), Sobeloff, J., d i s s e n t ­

in g .

15 T e a m s te r s  v. U n i t e d  S ta t e s ,  431 U.S. 345, 349-350 (1977) deals with 
a case of pre-act segregation perpetuated by post-act operation of a 
seniority system. In this case, post-act segregation is perpetuated by a 
refusal to consider those segregated for promotion or transfer to white 
jobs.

16 District Court (1-96-107). Court of Appeals: A t o n i o  (III-15, 43-47) 
768 F.2d 1125, 1131; (V-9-12) 810 F.2d 1480; (VI-4-9) 827 F.2d at 442.

17 G r ig g s  v. D u k e  P o w e r  C o . ,  401 U.S. 424 (1971).

18 M c D o n n e l l  D o u g la s  v. G r e e n , 411 U.S. 732 (1973).

19 T e a m s te r s  v . U n i t e d  S ta t e s ,  431 U.S. 338 at 358 (1977).

544



8

Aikens,20 this Court criticized the district court for addressing 
the existence of a prima facie case when all the evidence was in, 
rather than dealing with the question of discrimination vel non. 
In Texas Department o f Community Affairs v. Burdine,21 this 
Court noted how the lower court’s procedural rulings harbored 
a substantive error. The same errors were committed here.

IV. THE SO-CALLED “ OVER-REPRESENTATION’’ OF 
MINORITIES IN LOWER PAYING JOBS, PLUS 
THEIR EXCLUSION FROM HIGHER PAYING 
WHITE JOBS, DOES CONSTITUTE ILLEGAL SEGRE­
GATION UNDER TITLE VH

The District Court addressed skills requirements for the 
“white jobs’’ as matters of disparate treatment requiring proof 
of intent.22 On that issue, it examined the statistics showing the 
disparity between the large number of minorities in the lower 
paying jobs and their absence in the “ white” jobs. It dis­
counted this evidence, calling it “ over-representation” because 
minorities were only a tiny fraction of the total population of 
Alaska, Washington and Oregon.23 The District Court then 
viewed individual instances of rebuffed applicants, word of 
mouth recruiting among whites, racial labels, segregated hous­
ing and eating facilities as either justified or insignificant. It 
noted that “ this is not a promotion from within case,”24 but did 
not find that all minorities were unqualified for the “white 
jobs.”25

20 460 U.S. 711, 715-717 (1983).

21 450 U.S. 248, 258-259 (1981).

22 The District Court treated skills requirements as subjective and 
therefore not subject to the disparate impact rule. District Court (I- 
102) .

23 District Court findings #103 (1-35), #105 (1-36), #107 (1-37), #109 
(1-38-39), #121 (1-42).

24 District Court (I-114).

25 Any such finding would have been inconsistent with the District 
Court’s conclusion that some of the “at issue” jobs were unskilled. 
District Court finding #134 (1-75).

545



9

The Court of Appeals in reviewing the District Court stated:
Thus, when considering the skilled positions, the [dis­

trict] court found that statistics which merely highlight the 
segregation of whites and nonwhites between the at-issue 
and cannery worker jobs, without more, could not serve to 
raise an inference that the segregation is attributable to 
intentional discrimination against any particular race.26

To summarize, the District Court identified the “ over- 
representation” of minorities in the lower paying jobs. This 
“over-representation” was then relied upon to deemphasize the 
comparison of the number of minorities in lower paying jobs 
with whites in higher paying jobs. This “over-representation” 
is a euphemism for segregation. Treating segregation as “ over­
representation” obscured segregation as a violation.27

The argument that because plaintiffs are segregated they are 
entitled to no relief because they are over-represented is disin­
genuous. In early Title VII cases, employers did not argue they 
were entitled to keep Blacks in lower paying jobs because they 
had so many of them.28 Where the employer uses segregated 
recruiting processes to hire minorities or women into lower pay­
ing jobs and then refuses as a matter of policy to consider them 
for promotion or transfer, nothing more is needed to establish a 
violation of Title VII.29 This case is not analogous to Watson v.

26 A t o n i o  (VI-16) 827 F.2d at 444.

27 District Court finding #121 (1-42) treated “over-representation” as a 
reason not to credit statistics comparing proportions of minorities in 
lower paying jobs with whites in higher paying jobs.

28 L o c a l  1 8 9 , U n i t e d  P a p e r m a k e r s v. U n i t e d  S ta t e s , 416 F.2d 980 (5th 
Cir. 1969); U n i t e d  S t a t e s v. B e th le h e m  S te e l  C o r p . , 446 F.2d 652 (2d 
Cir. 1971). In the early years under Title VII, the E E O C  frequently 
obtained promises of “promotion from within” to end job segrega­
tion. Blumrosen, S e n io r i t y  a n d  E q u a l  E m p l o y m e n t  O p p o r t u n i t y :  A  
G l i m m e r  o f  H o p e , 23 Rutgers L. Rev. 268, 273-274, 303 (1969).

29 “And it is unthinkable that a citizen of this great country should be 
relegated to unremitting toil with never a glimmer of light in the mid­
night of it all.” Gwin, J. in M il l e r v. I n te r n a t io n a l  P a p e r  C o . , 408 F.2d 
283 (5th Cir. 1969). The sentiment is applicable to Alaska.

546



10

Fort Worth Bank and Trust,30 where four justices were con­
cerned with the risk of finding discrimination when it did not 
exist. Rather it is its opposite—a failure to see discrimination 
when it is blatant.

The Court of Appeals compounded the error of the District 
Court in the statement quoted above. It assumed that proof of 
segregation in hiring and assignment along with the refusal to 
allow promotion and transfer was not enough to show a viola­
tion of the statute, but that, in addition, intentional discrimina­
tion had to be shown.31 This double burden, a requirement of 
showing both segregation and discrimination, is not warranted. 
The statute makes segregation itself illegal.

The statute is intended to assist those who have been segre­
gated to break out of their situations, not to permit the fact of 
segregation to justify restrictions against them. The segregation 
into low paying jobs does not constitute favored treatment as 
the term “over-representation” suggests; rather, it constitutes 
the continued exploitation of minority workers trapped into 
low paying jobs. This Court has repeatedly said that the objec­
tive of the statute is to open opportunities to those who have 
traditionally been denied them.32 In this case, the group interest 
of minorities in freedom from job segregation is identical to the 
interest of each individual minority group member.

30 487 U.S______ _ 108 S. Ct. 2777 (1988).

31 “ . . . statistics which merely highlight the segregation of whites and 
nonwhites between the at-issue and cannery worker jobs, without 
more, could not serve to raise an inference that th e  s e g r e g a t io n  is  

a t t r i b u t a b l e  t o  i n t e n t i o n a l  d i s c r im in a t io n  a g a in s t  a n y  p a r t i c u la r  r a c e .”  

(emphasis added) A t o n i o  (VI-16) 827 F.2d at 444.

32 G r ig g s v. D u k e  P o w e r  C o . , 401 U.S. 424, 426, 429-432 (1971); A l b e ­

m a r le  P a p e r  C o . v. M o o d y , 422 U.S. 405, 417-418 (1975); U n i t e d  

S t e e l w o r k e r s  v . W e b e r , 443 U.S. 193, 202-203 (1979); F o r d  M o t o r  C o .  

v . E E O C , 458 U.S. 219, 228 (1982); J o h n s o n  v. S a n ta  C la r a  C o u n ty  

T r a n s p o r ta t io n  A g e n c y , 107 S. Ct. 1442 (1987).
The Uniform Guidelines on Employee Selection Procedure, while 

supporting the "bottom line” concept with respect to employers who 
employ at the availability level, expressly states that this concept is 
inapplicable to those employees who have been subject of prior restric­
tions on promotional opportunity. Uniform Guidelines on Employee

547



11

This Court has frequently noted that the statute proscribes 
discrimination against individuals.33 In Connecticut v. Teal,34 
the court stated that the employer could not “cancel out’’ dis­
crimination against some minorities by promoting others. Simi­
larly in Furnco35 and Teamsters,36 this Court held that the hiring 
or promotion of some minorities does not permit an employer 
to discriminate against others. In this case, each individual 
minority worker is a victim of the unlawful segregation of 
minorities. This deprivation of individual rights cannot be justi­
fied by a claim that the concentration of minorities in segre­
gated jobs constitutes “ over-representation.’’

V. THERE CAN BE NO “BUSINESS NECESSITY” JUSTI­
FICATION FOR MAINTAINING JOB SEGREGATION

The Court of Appeals reviewed the District Court’s analysis 
of the facts from the perspective of the disparate impact princi­
ple, with its corollary defense of business necessity. The Court 
of Appeals said:

. . . While the district court discounted the comparative 
statistics in evaluating the claims of intentional discrimina­
tion in skilled jobs we find them sufficiently probative of 
adverse impact. The statistics show only racial stratifica­
tion by job category. This is sufficient to raise an inference

Selection Procedures-1978, 29 C.F.R. Sec. 1607.4(C)(1) provides that 
the “bottom line” is not a justification “where the selection procedure 
is a significant factor in the continuation of patterns of assignment of 
incumbent employees caused by prior discriminatory employment 
practices.”

33 C o n n e c t i c u t v. T e a t , 457 U.S. 440 (1982); L o s  A n g e le s  W a te r  a n d  
P o w e r  C o . v. M a n h a r t , 435 U.S. 702 (1978). P h i l l i p s v. M a r t in -  

M a r i e t t a  C o r p . , 400 U.S. 542, 543-544 (1971) held that an employer 
could not justify exclusion of women with young children on the 
grounds that it hired many other women.

34 457 U.S. at 452-456 (1982).

35 F u r n c o  C o n s t r .  C o r p . v. W a te r s , 438 U.S. 567 at 579 (1978).

36 T e a m s te r s v. U n i t e d  S ta t e s , 431 U.S. 324, 341-342 (1977).

548



12

that some practice or combination of practices has caused 
the distribution of employees by race and to place the bur­
den on the employer to ju stify  the business necessity o f  the 
practices identified by the plaintiffs, [emphasis added]17

This analysis contained an error of law in assuming, without 
discussion, that the defense of business necessity was available 
in a case where the employer knowingly creates and maintains 
job segregation and does not consider minority workers for 
advancement into white jobs. The statute does not permit the 
defense of “ business necessity” in this type of case.

The “ business necessity” defense was developed by this 
Court as a component of the concept that practices with dispar­
ate impact on minorities are illegal.37 38 But it has no relevance to 
cases of overt discrimination. Any economic advantage which 
the employer may derive from such segregation is simply ille­
gal.39 The “ legitimate business reason” test was developed in 
M cDonnell Douglas as a method of ordering proof where the 
issue of the employer’s motive is clearly drawn between two 
possibilities, one legal and one illegal. Neither test is required to 
be applied to practices which segregate minorities in hiring, 
assignment, promotion, and transfer.40 Here, it is clear that the 
employer regularly and normally treated minorities less favor­
ably than whites.

The recruitment and hiring practices of the employer pro­
duced a segregated work force. The no-promotion policy main­
tained that segregation. There is no justification for this refusal 
to consider incumbent minority employees for promotion or

37 A t o n i o  (VI-18) 827 F.2d at 444.

38 S e e Blumrosen, S tr a n g e r s  in  P a r a d i s e :  G r ig g s  v. D u k e  P o w e r  C o . ,  
a n d  th e  C o n c e p t  o f  E m p l o y m e n t  D i s c r im in a t io n , 71 Mich L. Rev. 59, 
81-84 (1972).

39 C o m p a r e  C o r n in g  G la s s  W o r k s v. B r e n n a n , 417 U.S. 189, 205 
(1974).

40 The closest this Court has come to addressing a situation such as this 
is T e a m s te r s v. U n i t e d  S ta t e s , 431 U.S. 349 (1977).

549



13

transfer to “ white” jobs for which they may be qualified.41 The 
statute does not provide a Bona Fide Occupational Qualifica­
tion (BFOQ) defense for racial discrimination.42 The business 
necessity claim in a race case must be construed in a most lim­
ited way, so as not to defeat the purpose of the statute.43 Even 
where the statute does provide for a BFOQ defense, this Court 
has been careful to limit the scope of that defense to preserve 
the thrust of the prohibition on discrimination.44 While an 
employer may demonstrate that it could not recruit an inte­
grated labor force for a specific job because of availability, it 
cannot simultaneously refuse to consider the people it hires into 
a segregated job for other opportunities without violating Title
VII.45

These facts as a matter of law constitute the maintenance of a 
segregated work force which denied minorities opportunities 
for advancement. Intentional segregation is established by

41 These practices were not found to constitute a bona fide seniority 
system. C o m p a r e  District Court finding #101 (1-35).

42 S e e Sec. 703(e)(1). The District Court appeared to apply a lose form 
of a B F O Q  defense by its suggestions that many members of the class 
do not speak English, and prefer to fish rather than work in the sum­
mer time. District Court Finding #100 (1-34). The "business necessity” 
defense of G r ig g s has not been applied to cases of overt discrimination 
by this Court. In fact, A l b e m a r l e  P a p e r  C o . v. M o o d y , 422 U.S. 405 
(1975) suggests that claims of business necessity would not justify overt 
discrimination.

43 U n i t e d  S t a t e s v. B e th le h e m  S te e l  C o r p . , 446 F.2d 652, 662 (2d Cir. 
1971).

44 D o t h a r d  v . R a w l in s o n , 433 U.S. 321, 333 (1977).

45 In F u r n c o  C o n s t r .  C o . v. W a te r s , 438 U.S. 567 (1978), the employer 
justified its policy of not hiring at the gate on the grounds that it 
needed some information about the applicants’ capabilities before hir­
ing them. This argument cannot justify a blanket refusal to consider 
incumbent employees for promotion where these employees are well 
known to the employer. F u r n c o ’s caution against courts restructuring 
an employer’s recruitment and hiring practices has no application to a 
case where segregation exists. In F u r n c o , the employer’s statistics sug­
gested a lack of discriminatory intent. Here, the promotion and trans­
fer policies themselves constitute illegal maintenance of segregation.

550



14

proof of the fact of segregation by race in the hiring process and 
job assignments, along with its knowing maintenance by refus­
ing to permit promotions and transfers.46 All of the opinions 
below assumed that the facts in this case had to be fitted into 
the mold of either disparate treatment or disparate impact. As a 
consequence they treated separately these facts concerning seg­
regation in hiring, assignment, and refusal to permit promotion 
or transfer, which, taken together, establish segregation in vio­
lation of Title VII.47

As it did in Aikens, this Court should make clear that the 
lower courts must decide ultimate issues of segregation or dis­
crimination vel non when all the evidence is before them. When 
faced with blatant segregation, the lower courts need not fit the 
case into categories of disparate impact or disparate treatment. 
There can be no valid reason for the conscious maintenance of a 
racially segregated work force which flowed from the refusal to 
allow transfer and promotion to white jobs.

46 The decision to conduct all hiring in the lower 48 Tor white jobs, and 
not to consider applications from incumbent minority employees dur­
ing the time they are employed, obviously makes it easier for whites 
than minorities to make applications.

47 The District Court did not properly apply that aspect of M cD onnell 
D ou glas  v. Green which deals with statistics. M cD on n ell D ouglas  
states that “ [olther evidence which may be relevant to any showing of 
pretext includes facts as to . . . petitioner’s general policy and practice 
with respect to minority employment. On the latter point, statistics as 
to petitioner's employment policy may be helpful to a determination of 
whether petitioners’s refusal to rehire respondent in this case con­
formed to a general pattern of discrimination against blacks. (411 U.S. 
at 804-805). In the accompanying footnote, the Court stated that, 
“ [tjhe District court may, for example, determine after reasonable dis­
covery that, ‘the [racial) composition of defendant’s labor force is 
itself reflective of restrictive or exclusionary practices.’ ” Id. at 805 
(citation omitted). Contrary to these suggestions, the District Court 
found that such evidence established “over-representation,” not dis­
crimination. This was an error of law.

551



15

VI. THE COURT SHOULD AFFIRM THE HOLDING OF 
THE COURT OF APPEALS ON THE GROUND THAT 
ILLEGAL SEGREGATION HAS BEEN ESTABLISHED 
RATHER THAN DISMISS THE WRIT AS IMPROVI- 
DENTLY GRANTED

This court granted Certiorari to consider questions which 
relate to disparate impact theory. But the application of that 
theory to the facts of this case would permit a business necessity 
defense to a case of overt segregation.48 This is the substantive 
error embedded in the application of the disparate impact anal­
ysis to a case of job segregation. Thus, the Court of Appeals 
reasoning is in error.

While this Court could dismiss the writ as improvidently 
granted, to do so would leave uncorrected the error of the 
Court of Appeals in permitting a business necessity defense to 
job segregation. The error of the Court of Appeals should be 
corrected lest it generate other efforts to evade Title VII. At the 
same time, the Court of Appeals correctly overturned the Dis­
trict Court’s analysis that “ over-representation” of minorities 
detracted from the proof of discrimination. In this, the holding 
of the Court of Appeals should be affirmed. Therefore, the 
NAACP urges the Court to correct both the plain error of the 
District Court in its failure to appreciate the significance of the 
facts concerning job segregation, and the error of law of the 
Court of Appeals in recognizing a business necessity defense to 
maintenance of job segregation. This can be accomplished by 
affirming the Court of Appeals’ holding and remanding with 
instructions that the evidence of segregated hiring, job assign­
ments, and refusals to consider minorities for promotion and 
transfer constitute a violation of the statutory prohibition on 
segregation.

48 The statement of Question Presented #1, note 1, supra , assumes the 
legitimacy of the “ no promotion” rule which is illegal under the facts 
of this case.

552



16

CONCLUSION

A generation after Title VII of the Civil Rights Act was 
adopted, changed circumstances, some resulting from its imple­
mentation, have created new problems of interpretation. The 
improvement in minority and female employment under the 
statute, as interpreted by this Court, has moved the issues from 
those crude forms of discrimination of the 1960’s to more sub­
tle limitations on minority and female employment.49 But this 
case is not the proper vehicle to examine these subtle questions. 
It is a case of crude, currently maintained, segregation. To treat 
this case otherwise will permit an overt discriminator to rely on 
defenses tailored to more refined cases, and, thus, permit con­
tinued racial segregation. Pockets of continued segregation 
remain, as this case illustrates. Congress directly prohibited seg­
regated employment practices such as those found to exist in 
this case by the District Court. The plaintiffs are entitled to the 
full protection of the Civil Rights Act of 1964.

Respectfully submitted,

G ro v er  G. H a n k in s *
G e n e ra l C o u n s e l

Sa m u e l  M . W a l t e r s  
A s s is ta n t  G e n e ra l C o u n s e l  

N a t io n a l  A s s o c ia t io n  fo r  t h e  
A d v a n c e m e n t  o f  C o l o r e d  P e o p l e  
Special Contribution Fund, and

A l f r e d  W . Blu m r o se n

* C o u n s e l  o f  R e c o r d

49 Blumrosen, The le g a c y  o f  Griggs: Social P rogress an d  Subjective  
Ju dgm en ts, 63 Chi. Kent L. Rev. 1 (1986).

553





No. 87-1387

.................  .... ' ' ■ -  ' 1 " — t

3)n tlje Supreme Court of tlje duiteb states
October Term, 1988

Wards Cove Packing Company, Inc., et ai ., petitioners

v.

Frank Atonio, et ai .

O N  W R I T  O F  C E R  T I O R A  R l  
T O  T H E  U N I T E D  S T A  T E S  C O U R T  O F  A P P E A L S  

F O R  T H E  N I N T H  C I R C U I T

B R IEF FOR T H E  UN ITED STA TES AS AM ICUS C U R IA E 
SUPPORTING PETITIO N ERS

C h a r les  F ried  
S o lic ilo r  G en era l

Wm. B radeord  R eyn o ld s  
A s s is ta n t  A t to r n e y  G en era l

Roger C leg g
D e p u ty  A s s is ta n t A t to r n e y  G en era l

R ic h a r d  G. I a r a n io  
A s s is ta n t  to  th e  S o lic ito r  G en era l

Dav id  K. F lyn n  
L isa J. Stark  

A t to m e  vs

D e p a r tm e n t o f  J u stice  
W a sh in g to n , D .C . 2 0 5 3 0  
(2 02) 6 3 3 -2 2 1 7

555



Q U E S T IO N S  P R E S E N T E D

1. In this discrimination suit under Title VII of the Civil 
Rights Act of 1964, 42 U.S.C. 2000e et seq., whether the court 
of appeals correctly held that respondent-employees’ statistical 
evidence, which showed a marked disparity between the propor­
tion of minorities in the jobs at issue in the case and the propor­
tion of minorities in other jobs of the same employers, made out 
a prima facie case of disparate impact in selection for the jobs at 
issue.

2. Whether the court of appeals improperly allocated the 
burdens of proof and engaged in impermissible factfinding in 
applying disparate impact analysis to the challenged employ­
ment practices.

3. Whether disparate impact analysis allows employees to 
challenge the cumulative effect of a wide range of alleged em­
ployment practices.

(I)

556



TABLE OF CONTENTS

Page
Interest of the United States....................................................  I
Statement .................................................................................  I
Introduction and summary of argument.................................. 12
Argument:

I. The court of appeals incorrectly held that 
respondents’ statistics made out a prima facie case of
disparate impact........................................................  16

II. After a plaintiff makes out a prima facie case show­
ing that an identified selection mechanism causes a 
disparate impact, the employer has the burden of 
producing enough evidence to sustain a judgment in 
its favor that the challenged mechanism significantly 
serves legitimate business goals, and the plaintiff may 
then prevail by proving the contrary or by showing 
that an alternative practice with a less disparate im­
pact equally serves those goals.................................. 21

Conclusion ............................................................................... 29

T A B L E  OF A U TH O R ITIES

Cases:
A guilera  v. C o o k  C ou n ty  P olice & C orrections M erit 

B oard, 760 F.2d 844 (7th Cir.), cert, denied, 474 U.S.
907 (1985)...................................................................... 24

A lbem arle  P aper C o. v. M o o d y, 422 U.S. 405 (1975) . . . .  13, 14,
17, 23, 25, 28

B oard  o f  Trustees v. Sweeney, 439 U.S. 24 (1978)...........  26
B urw ell v. E astern A ir  Lines, Inc., 633 F.2d 361 (4th Cir.

1980), cert, denied, 450 U.S. 965 (1981)........................  24
Chrisner v. C om ple te  A u to  Transit, Inc., 645 F.2d 1251

(6th Cir. 1981)................................................................ 10. 24
C onnecticu t v. Teal, 457 U.S. 440(1982)........................  9, 13,

14, 22, 23
D o t hard  v. R aw linson, 433 U.S. 321 (1977)....................  13, 17,

23, 25
.E E O C  v. Rath Packing C o ., 787 F .2d 318 (8th Cir. 1986), 

cert, denied, No. 86-67 (Oct. 14, 1986) ......................

(HO

557



IV

C&ses — Continued: Page
Eubanks v. Pickens-Bond Constr. Co., 635 F.2d 1341 (8th

Cir. 1980)........................................................................  18
Furnco Constr. Corp. v. Waters, 438 U.S. 567 (1978) . . . .  25,

26, 29
Grano v. Dep’t o f Development, 637 F.2d 1073 (6th Cir.

1980) ..............................................................................  17
Griggs v. Duke Power Co., 401 U.S. 424(1971)...............  13, 17,

22, 23, 24, 25
Harnmon v. Barry, 813 F.2d 412 (D.C. Cir. 1987), cert.

denied, No. 87-1150 (May 31, 1988)...............................  17
Hazelwood School Dist. v. United States, 433 U.S. 299

(1977) ........................................................................ 16, 17, 18
Hester v. Southern Ry., 497 F.2d 1374 (5th Cir. 1974) . . . .  17
Johnson v. Transportation Agency, No. 85-1129 (Mar. 25,

1987)   25
Kinsey v. First Regional Securities, Inc., 557 F.2d 830

(D.C. Cir. 1977)..................... .’ ...................................... 24
Kirby v. Colony Furniture Co., 613 F.2d 696 (8th Cir.

1980) ..............................................................................  24
Lewis v. N LR B , 750 F.2d 1266 (5th Cir. 1985)................... 17
McDonnell Douglas Corp. v. Green, 411 U.S. 792

(1973)..............................................................................  26
Metrocare v. Washington Metro. Area Transit Authority,

679 F.2d 922 (D.C. Cir. 1982)..................................  17
Mister v. Illinois Cent. G ulf R .R., 832 F.2d 1427 (7th Cir.

1987) ..............................................................................  17
Moore v. Hughes Helicopters, Inc., 708 F.2d 475 (9th Cir.

1983) ..............................................................................  17
New York Transit Authority v. Beazer, 440 U.S. 568

(1979) .....................................................  13, 17, 23, 24, 25, 26
N LR B  v. Transportation Mgrnt. Corp., 462 U.S. 393

(1983)..............................................................................  26
Parson v. Kaiser Aluminum & Chemical Corp., 575 F.2d 

1374 (5th Cir. 1978), cert, denied, 441 U.S. 968
(1979)..............................................................................  24

Piva v. Xerox Corp., 654 F.2d 591 (9th Cir. 1981)...........  17
Reynolds v. Sheet Metal Workers, Local 102, 702 F.2d 221

(D.C. Cir. 1981).............................................................  18
Rowe v. Cleveland Pneumatic Co. Numerical Control,

690 F.2d 88 (6th Cir. 1982)............................................ 17

558



V

Cases-Continued: r>a8e
T ea m stersv . U n ited  Slates, 431 U.S. 324(1977)............... 13, 17,

18, 28
Texas D ep't o f  C om m u n ity  A ffa irs  v. Bur dine, 450 U.S.

248 (1981) .................................................................... 26, 27
U nited  S ta tes  v. C ounty  o f  Fairfax, 629 F.2d 932 (4th Cir.

1980), cert, denied, 449 U.S. 1078 (1981).....................  17
U n ited  S teelw orkers o f  A m erica  v. Weber, 443 U.S. 193

(1979) ....................................................................... 22, 25, 27
W am bheim  v. J .C . Penney C o ., 705 F.2d 1492 (9th Cir.

1983), cert, denied, 467 U.S. 1255 (1984)...................... 24
W ashington  v. D avis, 426 U.S. 229 (1976)......................  23, 25
W atson  v. F ort W orth Bank A Trust, No. 86-6139 (June

29, 1988) ..................................................................... ..p a s s im
W heeler v. C ity  o f  C olum bus, 686 F.2d 1144 (5th Cir.

1982) .............................................................................  18
W illiam s v. C olorado  Springs School D ist. N o. II, 641

F.2d 835 (10th Cir. 1981)............................................... 24

Statutes, regulations and rule:
Administrative Procedure Act, 5 U.S.C. 556(d) (§ 7(c)) .. 26
Civil Rights Act of 1964, Tit. V II, 42 U.S.C. 2000e

et ............................................................................................  !
42 U.S.C. 2000e-2(a)(l).............................................
42 U .S.C. 2000e-2(a)(2).............................................
42 U.S.C. 2000e-2(j)................................................... 18

29C.F.R . Pt. 1607 ............................................................  17
Fed. R. Evid. 301..............................................................  26

Miscellaneous:
E. Cleary, M cC orm ick  on E vidence  (2d ed. 1972)
44 Fed. Reg. 11998 (1979)..................................
Restatement (Second) of Torts (1965).................

559



3n tlje Supreme Court of tljc Uniteb ^tatefi
October Term, 1988

N o. 87-1387

Wards Cove Packing Company, Inc., et al„ petitioners

v.

Frank Atonio, et al.

O N  W R IT  O F  C E R T IO R A R I  
TO TH E U N IT E D  S T  A TES C O U R T  O F  A P P E A L S  

FOR TH E N IN T H  C IR C U IT

BRIEF FOR THE UNITED STATES AS AMICUS CURIAE 
SUPPORTING PETITIONERS

INTEREST OF THE UNITED STATES

This case presents important questions concerning the mean­
ing and application of Title V l l  of the C iv il Rights Act of 1964, 
42 U .S .C . 2000e et seq. The Attorney General has significant 
Title V l l  enforcement responsibilities. The United States, as the 
nation’s largest employer, is also subject to Title V II  re­
quirements.

STATEMENT

1. Petitioners Wards Cove Packing Company and Castle & 
Cooke, Inc., operate salmon canneries in Alaska (Pet. App. 14). 
Most of the canneries are located in remote, widely separated, 
and sparsely populated areas of Alaska (id. al 116-117, 132). 
They operate only during the salmon run for several months 
each summer: they lie vacant during the winter and are re­
opened and prepared for operation in May and June (the pre­
season) (id. at III3-II14). Accordingly, petitioners hire most of 
their employees from areas distant from the canneries, and the 
canneries furnish on-site housing and dining for the employees 
(id. at 117, 141, 1118).

560



2

Petitioners’ workforce is to a large extent racially stratified. 
The workforce as a whole has been approximately 43% minori­
ty (principally, Filipino and Alaska Native) since 1970 (Br. in 
Opp. 1), and that figure is representative of the entire Alaska 
salmon canning industry (ibid.). Minorities, however, are heavi­
ly concentrated in the lower paying cannery-line jobs and, at 
some canneries, in certain laborer positions (Pet. 4-5; Br. in 
Opp. 1-2). The higher paying noncannery jobs, including 
clerical, administrative, machinist, and other positions, are pre­
dominantly white (Pet. 4; Br. in Opp. 1-2).*

Respondents are a class of former and current nonwhite can­
nery employees of petitioners (Pet. App. 12). In 1974, they 
brought this suit under Title VII alleging that petitioners dis­
criminate on the basis of race in hiring, firing, paying, pro­
moting, housing, and dining at the canneries (ibid.).2 Pointing 
principally, though not only, to the disproportionate concentra­
tion of minorities in the cannery jobs, they sought to establish 
class-wide and individual liability both on disparate treatment 
and disparate impact theories.

2. After trial, the district court made detailed findings of 
fact on each of the many challenged practices and entered judg­
ment for petitioners (Pet. App. II-1130).

a. Describing petitioners’ employee-selection practices, the 
district court found that many jobs are filled pursuant to rehire- 
preference clauses of union contracts. Those clauses operate 
like seniority provisions, so that employees who have satis­
factorily worked in particular jobs in a prior season are rehired 
for the same jobs in the new season (Pet. App. 129, 135). The * 1

1 At issue in this case are the jobs other than those on rhe cannery line (non­
cannery jobs) (Pet. 4; Pet. App. 128). Respondents note some variation in the 
minority percentage in various noncannery jobs (Br. in Opp. 1-2), while peti­
tioners state that the overall percentage of minorities in noncannery jobs at the 
particular canneries at issue for the period at issue in the district court was 
21 °79 (Pet. 4). There is no dispute that minorities are heavily concentrated in 
the cannery jobs.

1 Suit was originally brought against Columbia Wards Fisheries as well as 
petitioners, but the claims against that defendant were dismissed (see Pet. 
App. Ill 13) and are not at issue in this Court.

561



3

court also found that, while some workers are hired from the 
areas surrounding the canneries, the remainder are hired at peti­
tioners’ home offices in Washington and Oregon and trans­
ported to the canneries when their jobs begin (id. at 130). Not­
withstanding those common elements, the channels for selection 
of cannery and noncannery workers are generally distinct. In 
particular, except for local Alaska residents and persons with a 
rehire preference, cannery jobs are filled through the dispatch 
procedure of Local 37 of the International Longshoremen’s 
Workers Union (Local 37) (id. at 132-133). By contrast, with the 
rehire-preference exception, noncannery jobs are filled by appli­
cations submitted during the fall and winter preceding the up­
coming season (id. at 130-131).3 Petitioners generally do not 
post notices at the canneries for any jobs (id. at 129).

Those selection mechanisms largely determine the workforce, 
because petitioners’ policy and practice have been to hire from 
outside its current workforce and not to promote employees 
from one position or department to another (Pet. App. 133-134, 
139). “Employees and non-employees are free to apply for any 
job for which they feel qualified,” however, and “[sjimilarly 
situated applicants are treated equally” (id. at 133). Neverthe­
less, most applicants for noncannery positions are white, and 
few nonwhites have applied for those positions (id. at 131-132).4 
By contrast, Local 37 “provides an oversupply of nonwhite can­
nery workers for all [but one of petitioners’ canneries]” (id. at 
135). The court found that most cannery workers are nonwhite 
and that that is so because Local 37 is the primary source of 
such workers and Local 37 is predominantly Filipino in its 
membership (id. at 136).5

’ Petitioners receive far more applications than there are vacancies, and 
they generally do not consider applications or oral inquiries made during or 
just after the preceding season (Pet. App. 131-132).

4 The court found (Pet. App. 140): “There has been a general lack of interest 
by cannery workers in applying for noncannery workers jobs."

5 Nonunion members may be hired “although they must join the union" 
(Pet. App. 133). At the one cannery where Local 37 has not asserted juris­
dictional rights and hence does not supply cannery workers, the minority 
percentage of the cannery workforce is “significantly less” than at the other

562



4

The court found that the job qualifications for the cannery 
and noncannery jobs are generally different. All but certain 
designated noncannery jobs require skills and experience, and 
some require off-season or preseason availability (Pet. App. 
130, 135-136,155-176).6 Cannery jobs generally require unskilled 
labor (e.g., id. at 137), and none requires preseason availability 
(id: at 140-141). In contrast, many of the noncannery jobs re­
quire skills that the unskilled cannery workers do not possess 
and cannot readily acquire on the job during the short season 
(id. at 135, 140, 147). Petitioners do not provide on-the-job 
training {id. at 145), and they try to hire experienced persons for 
all jobs {id. at 146). The court stated that cannery workers and 
laborers do not make up a labor pool for other jobs (id. at 139).

Analyzing the relevance of respondents’ statistics to determin­
ing the labor pool, the court found that the available labor sup­
ply for cannery, laborer, and other unskilled jobs is 90% white 
and that Filipinos make up only about 1% of the population 
and labor force of Alaska, California, and the Pacific North­
west (Pet. App. 136-137). That nonwhites fill so large a propor­
tion of cannery jobs thus means that they are greatly overrepre­
sented in those jobs (id. at 137). For that reason, although 48% 
of the employees in the Alaska salmon canning industry as a 
whole are nonwhite, the court declined to assign much weight to 
that fact, explaining that “[tjhe institutional factor of Local 37’s 
overrepresentation of non-whites accounts for this statistic” (id. 
at 142). Looking particularly at the noncannery jobs, the court

canneries ( i d .  at 137). Similarly, Alaska Natives make up a high proportion of 
the resident cannery workers in the canneries located in communities where 
there are substantial numbers of Alaska Natives and a significantly lower pro­
portion at the one cannery where there are not such numbers in the community 
( i d .  at 137-138).

« The court listed 16 supervising jobs that require management abilities and 
extensive experience to perform successfully (Pet. App. 155-156) and 27 jobs 
that require substantial skill and experience1 to perform successfully ( id .  at 
157-158). It also set forth, for numerous jobs, detailed lists of qualifications 
that are “reasonably required for successful performance" ( id .  at 158, 158-175). 
Finally, the court identified certain jobs that are the only noncannery jobs that 
are not skilled positions (i d . at 175-176; see also id .  at II2-I13 (correcting list)).

563



5

made no finding of a general underrepresentation of nonwhites 
in those jobs. To the contrary, the court listed (id. at 143-145) 
numerous noncannery jobs in which it found, by reference to 
the relevant labor supply, that either nonwhites were overrepre­
sented or whites were not overrepresented by a statistically sig­
nificant amount.7

b. Based on those factual findings, the district court re­
jected respondents’ challenges. After stating that the burden of 
proof shifts to the employer once employees have made out a 
prima facie case under the disparate impact theory of discrimi­
nation (Pet. App. 197-198), the court concluded that disparate 
impact analysis applies only to objective practices, not to sub­
jective employer decisionmaking (id. at 199-1102). In this case, 
the court stated, disparate impact analysis applies to petitioners’ 
English-language requirement for many jobs and to the 
"nepotism” that allegedly influenced the selection of employees 
for some jobs (id. at 1102-1105). The court, however, found no 
basis for liability in either area.* Although the court did not ex­
pressly find disparate impact analysis applicable to other prac­
tices, it examined the validity of the rehire preference without 
regard to the existence of discriminatory intent, finding (id. at 1

1 The court also found some use of racial and ethnic labeling at the can­
neries (Pet. App. 176-180); recounted individual instances of alleged 
discrimination, without making findings on whether there was discrimination 
( i d .  at 184-194); found that petitioners’ dining practices originated in peti­
tioners’ deference to the leadership of Local 37 ( i d .  at 180-181); and found that 
petitioners’ housing practices, which had segregative effects, were based on 
workers’ department and arrival time, not on race ( id .  at 181-184).

' As to the language requirement, the court found that petitioners had 
proved that that requirement was justified by business necessity (Pet. App. 
1102-1103). As to the “nepotism," the court found that, although “Irjelatives 
of whites and particular!) nonwhites appear in high incidence at the canneries” 
( i d .  at 1104-1105), those persons were highly qualified and “were chosen 
because of their qualifications.” In addition, the court found that respondents’ 
statistics failed to account for post-hiring marriages ( id .  at 1105). Accordingly, 
the court concluded, “the nepotism which is present in the at-issue jobs does 
not exist because of a ‘preference’ for relatives” ( i b i d . ) .  The court also noted 
that “numerous white persons who ‘knew’ someone were not hired due to inex­
perience” ( i d .  at 1122).

564



6

1121-1122) that the preference was justified by business neces­
sity, given the importance of experience in work involving a 
short season, perishable foods, susceptibility of the product to 
lethal diseases like botulism, and other dangers.’

In analyzing respondents’ other claims under the disparate 
treatment theory of Title VII liability, the court first reiterated 
that all noncannery jobs except certain designated ones were 
skilled and that even some of the exceptions required that the 
employees be available prior to the onset of the canning season 
(Pet. App. 1107-1109). The court then discussed the statistical 
evidence that respondents introduced in an effort to make out a 
prima facie case. It found, first, that respondents were incorrect 
in arguing that the historical percentage of Filipinos and Alaska 
Natives hired in the Alaska salmon canning industry as a whole 
represented the available labor pool, because institutional fac­
tors (notably, the use of Local 37) “greatly distort the racial 
composition of the workforce” and “Alaskan Natives and 
Filipinos, combined, represent only about one percent of the 
population of Alaska, Washington, and Oregon from which 
state[s] (petitioners] draw their workforce” (id. at I1I0-I111). 
Second, the court found that the high percentage of nonwhites 
in the cannery jobs was sufficient to make out a prima facie case 
of disparate treatment in certain unskilled noncannery jobs (id. •

• The court found (Pet. App. 1126-1127) that petitioners’ housing practices 
would survive disparate impact analysis as well as disparate treatment analysis. 
It explained that petitioners “established that workers arriving preseason and 
staying post-season required better insulated housing,” that "workers are 
hous(ed) departmentally because the various departments worked the same 
shifts” ( i d .  at 1125), and that “(ijt is not efficient or economically feasible to 
open all bunkhouses perseason to assign workers arriving preseason to dif­
ferent housing with a result of maintaining more housing than necessary for 
longer periods of time” ( id .  at 1126-1127). The court similarly found ( id .  at 
II27-1129) that petitioners’ dining hall practices would survive disparate im­
pact (as well as disparate treatment) scrutiny, because the creation of dining 
arrangements along essentially racial lines was the responsibility of Local 37, 
which asked for and received a separate mess and culinary crew for its 
members (white members included). As the court explained, the cooks "simply 
acceded to the wishes of the older workers who preferred the traditional food 
that was served” ( id .  at 1129).

665



7

at 1111-1112). But the court then found that respondents had ar­
ticulated legitimate nondiscriminatory reasons for the disparity 
between minority representation in the cannery jobs and in un­
skilled noncannery jobs —chiefly, the lack of timely and formal 
applications from nonwhites for those jobs. The court found 
that respondents had not shown that those reasons were pretexts 
(id. at 1112). Third, as to the skilled noncannery jobs, the court 
found that the statistics concerning the percentage of minorities 
in the cannery jobs “have little probative value” (id. at 1114). 
The court explained that cannery workers do not constitute the 
proper labor force because they do not possess the skills and 
preseason availability required for the skilled noncannery jobs 
(id. at 1113).'0

Although the district court found no prima facie case of a 
pattern or practice of discriminatory treatment in hiring, pro­
moting, paying, or firing with respect to the skilled noncannery 
jobs based on respondents’ statistics alone, the court neverthe­
less concluded that respondents had “raised a marginal in­
ference of discriminatory treatment” based on the collective ef­
fect of the statistical evidence, the nepotism evidence, and in­
dividual instances of claimed discrimination (none of which the 
court found separately to have much probative value) (Pet. 
App. 1118-1119).“ The court found, however, that petitioners 
had met their burden of producing evidence that their hiring, 
promoting, paying, and firing practices were motivated by 
reasons other than race. The court also found that respondents 10

10 The court observed that respondents “were generality aware of (the) im­
portant qualification (of preseason availability]’’ and that “this is not a 
promotion-from-within case" (Pet. App. 1114).

" The court pointed to the evidenpe of individual instances of discrimina­
tion, but it found that, with one exception, all of the applicants either had no 
preseason availability (as far as the evidence showed) or made only oral in­
quiries, which “are not treated as applications in the cannery industry!, as 
respondents] appeared to have understood” (Pet. App. 1115-1116). The court 
noted some evidence that some respondents were deterred from applying for 
better jobs. Although the court found that evidence insufficient to establish 
that petitioners' practices caused the deterrence, it observed that a prima facie 
case did not require such proof ( id .  at 1116 II18).

566



8

had not shown that petitioners’ asserted motivations were pre- 
textual or that petitioners had acted as a result of racial animus 
(id. at 1119-1124). Thus, the court found that petitioners’ 
statistics were significantly more probative than respondents’. 
In particular, it found that the census data, which showed a 
90% white unskilled labor force and only 1% Filipino and 
Alaska Native population in the states from which petitioners 
hired employees, provided the best evidence of the available 
labor pool (id. at 1119-1120).'2 In addition, few respondents 
made timely and proper applications for jobs.'* As for the hir­
ing officials’ decisions themselves, the court found that 
“regardless of the manner in which a prospective employee came 
to the attention of the hiring personnel, the person was 
evaluated according to job related criteria” (id. at 1122). *4

3. A three-judge panel of the court of appeals affirmed 
(Pet. App. Ill 1-III56). With respect to disparate treatment, it 
found that the relevant district court factual findings were not 
clearly erroneous and were sufficient to support the finding that 
none of the challenged practices resulted in disparate treatment 
because o f race (id. at III20-III43). With respect to disparate im­
pact, the court of appeals ruled that the district court had cor­
rectly declined to apply that theory of Title VII liability to the 
various subjective employment practices challenged by 
respondents (id. at 11143-11154).'5 11 * * 14

11 The court also criticized respondents’ statistics for not controlling for the
(substantial) group of rehires, because the rehire preference was justified by 
business necessity, and because past discriminatory employment practices had 
not been established in this case (Pet. App. 1120-1122).

11 While some made oral inquiries about jobs, those inquiries did not con­
stitute applications for jobs and "were generally made of persons without hir­
ing authority.” The court further found that applications were typically made 
too late in the season for preseason jobs or by applicants who were not 
available for those jobs. Pet. App. 1123.

14 The court found, too, that "whites hired were paid no more than non­
whites” (Pet. App. 1123), that it was "unable to find a practice of deterrence,” 
and that various instances of race labeling were “not persuasive evidence of 
discriminatory intent” ( i b i d . ) .

" The court rejected the disparate impact challenge to the alleged practice 
of nepotism, explaining that the district court had not erred in finding "that no

567



9

4. The court of appeals granted rehearing en banc and 
vacated the initial panel opinion (Pet. App. 1V1-IV2). The en 
banc court held that disparate impact analysis applies to subjec­
tive employment practices “provided the plaintiffs have proved 
a causal connection between those practices and the demon­
strated impact on members of a protected class” (id. at VI6; see 
id. at V5). The court also explained how such analysis should 
work and remanded the case to the panel for application of the 
standards.

To establish a prima facie’case of discrimination under the 
disparate impact theory, a plaintiff “must (1) show a significant 
disparate impact on a protected class, (2) identify specific em­
ployment practices or selection criteria and (3) show the causal 
relationship between the identified practices and the impact” 
(Pet. App. V19-V20).* 16 “Once the plaintiff class has shown dis­
parate impact caused by specific, identifiable employment prac­
tices or criteria, the burden shifts to the employer” (id. at V35). 
Although the employer in any Title VII case may refute the 
statistical evidence, the court of appeals held that the employer’s 
burden in a disparate impact case is different from its burden in 
a disparate treatment case. Whereas in the latter the employer 
must merely articulate a nondiscriminatory reason for the 
disparity, and the plaintiff retains the burden of persuasion, in 
the former the employer "must prove the job relatedness or 
business necessity of the practice” that is challenged, and the 
burden of persuasion on that issue is shifted from the plaintiff 
to the employer. Id. at V35-V36. Meeting that burden “may be

pattern or practice of nepotism existed because there was no preference for 
relatives” (Pet. App. HI56). The district court’s rejection of the challenge to 
the pnglish-language requirement was not challenged on appeal ( id .  at III46 
n.5). Without elaboration, the court of appeals found the challenges concern­
ing the rehire preference and termination of Alaska Natives to be without 
merit ( id .  at II156).

16 See also Pet. App. V34-V35 (“the plaintiffs must prove that a specific 
business practice has a ‘significantly discriminatory impact,’ ” citing C o n n e c ­

t i c u t  v .  T e a l .  457 U.S. 440, 446 (1982)); id .  at V35 (“plaintiffs’ prima facie case 
consists of a showing of significant disparate impact on a protected class, 
caused by specific, identified, employment practices or selection criteria”).

568



10

an arduous task," the court noted (Pet. App. V38 (internal 
quotation marks omitted)), but the burden does not shift “until 
the plaintiff has shown a causal connection between the 
challenged practices and the impact on a protected class" 
(ibid.).11

5. On remand, the panel elaborated on the disparate impact 
standards set forth by the en banc court, applied those stand­
ards to respondents’ allegations, and remanded the case to the 
district court (Pet. App. VI1-VI44). With respect to the burden 
on the employer once a prima facie case has been made out, the 
panel explained that the burden in cases involving selection 
criteria may be met “by demonstrating that the selection criteria 
applied are essential to job safety or efficiency or correlated 
with success on the job (id. at VI6-VI7 (citations omitted)).1 * It 
also explained that, once the employer proves the business 
necessity of the challenged practices, the employees may “dem­
onstrate that other employment practices or selection devices

17 Judge Sneed, joined by three other judges, concurred separately (Pet. 
App. V40-V75). After explaining that the causation requirement articulated by 
the majority demands that there be a significant number of members of the 
protected group at issue who are qualified for the positions at issue ( i d .  at 
V55-V56), he argued that disparate impact analysis should not apply to all 
types of employment practices. It should apply only when the plaintiff claims 
that “the employer has articulated an unnecessary practice that makes the 
plaintiffs true qualifications Irrelevant" — i . e . ,  that allows the employer not to 
ascertain the plaintiffs true qualifications ( i d .  at V59, V60). By contrast, 
Judge Sneed argued that disparate treatment analysis should apply when the 
plaintiff claims that the employer, knowing the plaintiffs qualifications, ig­
nores them because of the plaintiffs membership in a protected group ( i d .  at 
V59). Applying that distinction, Judge Sneed concluded that disparate impact 
analysis was applicable to the use of separate hiring channels and word-of- 
mouth recruitment for the noncannery jobs and to the rehire preference, 
which allowed petitioners to ignore respondents’ true qualifications, but not to 
the housing and dining practices, which did not allow rejection of prospective 
minority employees without considering their qualifications ( id .  at V65-V70, 
V72, V73-V75).

" When employment practices other than selection devices are at issue, the 
court continued, the practice must be supported by “more than a business pur­
pose”; it must " ‘substantially promote the proficient operation of the 
business’ ’’ (Pet. App. VI7-VI8, quoting C h r i s n e r  v. C o m p l e t e  A u t o  T r a n s i t ,  

I n c . ,  645 F.2d 1251, 1262 (6th Cir. 1981)).

569



II

could serve the employer’s needs with a lesser impact on the pro­
tected class” (id. at V19).

The court began its application of those standards by con­
cluding that respondents’ statistical showing of a disparity be­
tween the racial composition of the cannery jobs and that of the 
noncannery jobs was “sufficient to raise an inference that some 
practice or combination of practices has caused the distribution 
of employees by race and to place the burden on the employer to 
justify the business necessity of the practices identified by the 
plaintiffs” (Pet. App. VI18). In so concluding, the court re­
jected the district court’s finding that the same statistics were 
not probative for skilled noncannery positions because they did 
not reflect the pool of persons who had the required skills and 
were available for preseason work (id. at VI17). The court 
stated that that finding “was error because when job qualifica­
tions are themselves at issue, the burden is on the employer to 
prove that there are no qualified minority people for the at-issue 
jobs” and that “it is unrealistic to expect statistics to be cali­
brated to reflect preseason availability when the preseason starts 
only one month earlier than the season” (ibid.).

Reviewing the challenged employment practices, the court 
stated that, if there was nepotism, that is by definition a policy 
of preferring relatives, and such a policy would have to be justi­
fied by business necessity (Pet. App. VI20-VI22).19 The court 
next observed that, while the district court had found that there 
were in fact objective criteria for noncannery jobs, it had not 
found that those criteria “were actually applied by those who 
made hiring decisions”; as the court of appeals construed them, 
the district court’s findings showed only that skill and experi­
ence were the general qualifications looked for by the hiring of­
ficers (id. at VI22-V123). As to respondents* challenge to the

19 We read the court of appeals’ somewhat opaque discussion on this point 
to leave open on remand to the district court the question whether petitioners 
had a policy of preferring relatives to others. The court of appeals cited 
evidence showing that, if the incidence of relatives in the workforce is the 
result of such a policy, the policy has a significantly disparate impact on non­
whites in certain departments (Pet. App. V121), because petitioners’ hiring of 
ficers are predominantly white.

570



12

subjective hiring process, therefore, the court of appeals ruled 
that the district court must determine whether the identified job 
qualifications “were actually applied in a non-discriminatory 
manner” (id. at VI25), bearing in mind that “the burden is on 
the employer to prove the lack of qualified people in the non­
white group” (id. at VI26). In addition, the court ruled, the 
district court must make findings as to the job-relatedness of the 
criteria actually applied (id. at VI27). With respect to peti­
tioners’ use of word-of-mouth recruitment for the noncannery 
jobs rather than the hiring channels used for cannery jobs, the 
court of appeals found that there was some evidence-and that 
logic suggested-that some of the cannery workers had the skills 
for the noncannery jobs; therefore, the court held, petitioners 
must prove the business necessity of that hiring practice (id. at 
VI27-VI31).20 The court further concluded that the district 
court’s finding of business necessity for the rehire preference 
was supported by the evidence (id. at VI32-VI33).1'

INTRODUCTION AND SUMMARY OF ARGUMENT
Title VII of the Civil Rights Act of 1964, as applied to em­

ployee selection procedures, makes it unlawful for covered em­
ployers not to hire an individual “because of such individual’s * 11

18 The court also stated that there was insufficient evidence to support the 
district court’s finding that the people available through the channels for 
cannery-worker hiring (Local 37 and local Alaska populations) were not 
available for the preseason (Pet. App. VI3I-V132).

11 With respect to the challenged practices other than selection criteria, the 
court ruled that the district court must determine whether the race labeling 
that was found “operates as a headwind to minority advancement” (Pet. App. 
VI33), that the efficiency justification for the housing practices was not suffi­
cient to sustain a finding of business necessity unless “the companies substan­
tiate that these measures are clearly necessary to promote the proficient opera­
tion of the business” ( i d .  at VI37), that the dining practices must be analyzed 
anew under the disparate impact theory ( id .  at VI36), that individuals claims 
(and the defense that the individuals failed to file timely formal applications) 
must be evaluated after the district court completes its disparate impact 
analysis of petitioners’ process for selecting noncannery workers ( id .  at V139- 
VI43).

571



13

race” (42 U.S.C. 2000e-2(a)(I) and (2)).22 Intentional dis­
crimination based on race is the primary way in which an em­
ployer can act unlawfully “because o f ’ race. As the legislative 
history of the 1964 Act makes clear and as this Court has said, 
“(u)ndoubtedly disparate treatment was the most obvious evil 
Congress had in mind when it enacted Title VII.” Teamsters v. 
United States, 431 U.S. 324, 335 n.15 (1977).

Based on the assumption that certain other exclusionary prac­
tices are “functionally equivalent to intentional discrimination” 
(Watson v. Fort Worth Bank & Trust, No. 86-6139 (June 29,
1988), slip op. 6), this Court in Griggs v. Duke Power Co., 401 
U.S. 424 (1971), identified a second way in which an employer’s 
nonhiring decision might be found to be “because o f ’ race. 
Under that theory, known as the “disparate impact” theory, a 
selection practice can be found unlawful even in the absence of 
a subjective intent to discriminate, if the practice has a signifi­
cantly disproportionate impact on a protected group and has no 
“manifest relationship to the employment in question” (id. at 
432). See Connecticut v. Teal, 457 U.S. 440, 446-447 (1982); 
New York Transit Authority v. Beazer, 440 U.S. 568, 584, 587
(1979); Dothard v. Rawlinson, 433 U.S. 321, 329 (1977); 
Albemarle Paper Co. v. Moody, 422 U.S. 405, 425 (1975). Once 
the challenged selection practice is identified, the disparate im­
pact theory does not focus on the historical fact of what the 
employer intended, as the disparate treatment theory does; 
rather, it aims at removing "artificial, arbitrary, and unneces­
sary barriers” to employment (Griggs, 401 U.S. at 431) by 11

11 Petitioners have challenged the court of appeals’ decisions in this Court 
only insofar as those decisions concern employee selection procedures. The 
housing, dining, and other nonselection employment practices are not 
separately at issue in this Court. Accordingly, we limit our discussion of 
disparate impact analysis to selection devices. We note that this Court has not 
applied disparate impact analysis to nonselection employment practices and 
that, if such application is proper at all, it would require, at a minimum, refor­
mulation of the standards that have been articulated to date.

Because this case involves only racial discrimination, we limit our discussion 
to "race,” although the statute prohibits hiring decisions because of “race, 
color, religion, sex, or national origin” (42 U.S.C. 2(XX)e 2(a)(1) and (2)).

572



14

focusing on the racial impact and the business justification for 
the device.23

This Court’s decisions have established a three-part structure 
for analysis of disparate impact claims: the first stage requires 
proof of disparate impact caused by an identified selection 
device (the prima facie case); the second requires a showing of 
job-relatedness by the employer; the third provides the plaintiff 
with an opportunity to demonstrate that there are effective 
alternatives to the challenged practice that have a less severe 
racial impact. That structure reflects the fact that Title VII was 
not designed to force employers to justify every selection prac­
tice. Hence, the most fundamental and well-established element 
of the structure is the principle that judicial inquiry into 
business justification in a disparate impact case is not called for 
until the complaining party proves a disparate impact that is 
caused by the challenged selection practice. See Teal, 457 U.S. 
at 446-447; Albemarle Paper Co., 422 U.S. at 425 (business 
justification is demanded “only after the complaining party or 
class has made out a prima facie case of discrimination”). The 
precise contours of the other elements of disparate impact 
analysis are less well settled, as this Court’s decision in Watson 
shows.

Petitioners’ first question involves the well-established re­
quirements of a prima facie case. The court of appeals ruled 
that respondents’ statistics were sufficient to carry their burden 
of proving a disparate impact in petitioners’ selection of 
employees for noncannery jobs. That conclusion is incorrect:

»  The meaning given by the disparate impact theory to the statutory con­
cept, “because of” race, is different from the meaning given by the disparate 
treatment theory. Whereas the latter asks whether race actually motivated the 
nqnhiring decision, the former makes an inquiry more analogous to the 
statistical inquiry of what factors account for, or explain, a particular 
phenomenon. Moreover, rather than looking at all possible nondiscriminatory 
explanations, disparate impact theory narrows the focus to possible business 
justifications. If a selection device is found to have a disparate impact on a 
particular group, and no explanation for the selection device (and hence the 
employer’s hiring decisions) can be found among sound business justifica­
tions, the only explanation remaining is race, and the nonhiring, in the terms 
of the statute, is therefore “because o f ’ race.

573



15

there was no basis for the court’s finding that petitioners’ selec­
tion practices had a disparate impact on minorities within the 
pool of applicants, or persons qualified, for those jobs. Indeed, 
the district court’s findings strongly suggest that there was no 
such disparate impact. Most notably, the intra-work force 
stratification shown by respondents (/.*?., the statistical disparity 
between the number of minorities in the cannery jobs and the 
number in the noncannery jobs) is explained by the use of Local 
37 for hiring in the cannery jobs; and because minorities are for 
that reason overrepresented in the cannery jobs, the stratifica­
tion does not suggest that exclusionary practices cause any 
underrepresentation in the noncannery jobs that are at issue.

Petitioners’ other questions are broad enough to encompass a 
challenge to the court of appeals’ definition of the structure of 
proof in a disparate impact case. We address four aspects of 
that structure that the Court could appropriately address to 
clarify the proper functioning of disparate impact analysis. 
First, a plaintiff who challenges a nonselection decision must, as 
part of the prima facie case, identify the actual mechanism used 
for the particular selection decision at issue. It is that selection 
mechanism that is the proper subject of disparate impact 
analysis when the plaintiff has alleged discrimination in hiring. 
The court should not focus on various practices that are not 
shown to have been part of the hiring decision, let alone prac­
tices that were concededly not part of the selection mechanism 
at all. Second, after a prima facie case has been made out, the 
question should be whether legitimate business goals are 
significantly served by the use of the selection device at issue. 
Third, the employer should have the burden of production on 
that issue, but not the burden of persuasion. Fourth, the plain­
tiff may prevail either by disproving the employer’s assertion 
that the selection device significantly serves legitimate business 
goals or by showing that alternatives exist that equally serve 
those goals but that have a lesser racial impact. In short, we urge 
the Court to adopt a framework based on the plurality opinion 
in Watson.

574



16

ARGUMENT
;I. THE COURT OF APPEALS INCORRECTLY HELD THAT RESPONDENTS’ STATISTICS MADE OUT A PRIMA 

FACIE CASE OF DISPARATE IMPACT
A. As the en banc court of appeals recognized (Pet. App. 

V19-V20), a prima facie case of disparate impact in selection for 
particular jobs requires that members of a protected group 
demonstrate that the selection mechanism caused a disparate 
impact on that group. That requires the plaintiffs to “offer 
statistical evidence of a kind and degree sufficient to show that 
the practice in question has caused the exclusion of applicants 
for jobs * * * because of their membership in a protected 
group.” Watson, slip op. 14 (plurality opinion); see also id. at 2 
n.2 (Blackmun, J., concurring in the judgment). Where, as 
here, minorities put forth statistics to show underrepresentation 
in the jobs at issue by comparing the number of minorities ac­
tually selected to the number of minorities in some larger pool, 
the definition of the pool must take account of the qualifica­
tions (including availability and interest) for the jobs at issue. A 
pool that is defined without reference to such qualifications 
cannot provide the basis for a prima facie case, because it does 
not support the inference that nonhiring of minorities was 
“because of” race rather than because of lack of qualifications.

Because the strength of statistical proof is subject to infinite 
gradations, the question whether particular statistics are “suffi­
ciently substantial that they raise * * * an inference of causa­
tion” ( Watson, slip op. 14 (plurality opinion)) calls for case-by­
case analysis. Id. at 14-15 n.3. But the common theme reflected 
in this Court’s decisions is that the comparison must be made by 
reference to a pool of individuals who are in the relevant labor 
market and are at least minimally qualified for the jobs at issue. 
Id. at 16 (“statistics based on an applicant pool containing in­
dividuals lacking minimal qualifications for the job would be of 
little probative value”); Hazelwoood School Dist. v. United 
States, 433 U.S. 299, 308 (1977) (footnote omitted) (“proper 
comparison was between the racial composition of [the 
employer’s] teaching staff and the racial composition of the

575



17

qualified public school teacher population in the relevant labor 
market’'); see Beazer, 440 U.S. at 586 n.29 (citation omitted) 
(“ ‘qualified job applicants’ ”); Teamsters, 431 U.S. at 338-340 
n.20 (same); Dothard, 433 U.S. at 330 (“otherwise qualified 
people”).24 A comparison with a pool that is too small (because 
it excludes substantial parts of the qualified labor pool) or too 
large (because it includes a substantial number of unqualified 
persons) does not support an inference that there are barriers to 
employment opportunities for minorities that are not present 
for others.

Typically, the pool of actual applicants —or, better, of 
qualified applicants —provides the proper benchmark for 
measuring disparate impact. See, e.g., Beazer, 440 U.S. at 585 
(rejecting statistics because they told “nothing about the class of 
otherwise-qualified applicants and employees” who are 
members of the protected class); Dothard, 433 U.S. at 329; 
Albemarle Paper Co., 422 U.S. at 425; Griggs, 401 U.S. at 426. 
See also Hazelwood School Dist., 433 U.S. at 308 n.13 (appli­
cant flow data would be “very relevant” and employer should be 
permitted to introduce such data).25 In some cases, however,

24 See M e tr o c a r e v. W a s h in g to n  M e t r o .  A r e a  T r a n s i t  A u t h o r i t y , 679 F.2d 
922, 930 (D.C. Cir. 1982) (“statistics must compare the percentage of blacks 
hired for given jobs with the percentage of blacks qualified for those 
positions’’); L e w is v. N L R B , 750 F.2d 1266, 1275 (5th Cir. 1985); G r a n o  v. 
D e p ' t  o f  D e v e l o p m e n t , 637 F.2d 1073, 1078 (6th Cir. 1980); P iv a  v. X e r o x  

C o r p . , 654 F.2d 591, 595 (9th Cir. 1981).

15 See H a m m o n  v. B a r r y , 813 F.2d 412, 427 n.31 (D.C. Cir. 1987), cert, 
denied, No. 87-1150 (May 31, 1988); U n i t e d  S ta t e s v. C o u n t y  o f  F a ir fa x , 629 
F.2d 932, 940 (4th Cir. 1980), cert, denied, 449 U.S. 1078 (1981); H e s te r v. 
S o u th e r n  R y . , 497 F.2d 1374, 1379 (5th Cir. 1974); R o w e v. C le v e la n d  

P n e u m a t i c  C o .  N u m e r ic a l  C o n t r o l , 690 F.2d 88, 93 (6th Cir. 1982); M i s te r v. 
I l l in o is  C e n t .  G u l f  R . R . , 832 F.2d 1427, 1435 (7th Cir. 1987); E E O C v .  R a th  

P a c k in g  C o . , 787 F.2d 318, 337 (8th Cir. 1986), cert, denied, No. 86 67 (Oct. 
14, 1986); M o o r e v. H u g h e s  H e l ic o p te r s ,  I n c . , 708 F.2d 475, 483 (9th Cir. 
1983).
Adverse impact under the Uniform Guidelines on Employee Selection Pro­

cedures, 29 C.F.R. Pt. 1607, is determined primarily by reference to applicant 
pools. See 44 Fed. Reg. 11998 (1979) (questions and answers on the meaning 
of the Uniform Guidelines).

576



18

the applicant pool may not be good evidence of the relevant 
qualified labor pool, because, for example, the plaintiff can 
prove that applications were deterred by the employer’s con­
duct. See Dothard, 433 U.S. at 330 (no need to use applicant 
pool where “otherwise qualified people might be discouraged 
from applying” by the height and weight requirements); 
Teamsters, 431 U.S. at 365-367; Reynolds v. Sheet Metal 
Workers, Local 102, 702 F.2d 221, 225 (D.C. Cir. 1981); 
Wheeler v. City o f Columbus, 686 F.2d 1144, 1152 (5th Cir. 
1982); Eubanks v. Pickens-Bond Constr. Co., 635 F.2d 1341, 
1350 n.10 (8th Cir. 1980).26 Even general population statistics 
may be sufficient evidence for particular jobs, but when that is 
so, it is because those statistics “accurately reflect the pool of 
qualified job applicants” (Teamsters, 431 U.S. at 339-340 n.20). 
“When special qualifications are required to fill particular jobs, 
comparisons to the general population (rather than to the 
smaller group of individuals who possess the necessary 
qualifications) may have little probative value” (Hazelwood 
School Dist., 433 U.S. at 308 n.13). Cf. 42 U.S.C. 2000e-2(j).

B. The court of appeals in this case was wholly unjustified 
in concluding (Pet. App. VI18) that respondents’ statistics suf­
ficed to meet their burden of making out of a prima facie case of 
disparate impact.27 Those statistics did not compare the number 
of selected minorities to the number that applied for the non­
cannery jobs. And the court of appeals did not have before it

»  Similarly, if an employer has made special recruiting efforts to increase 
the number of applicants from a particular protected group, the applicant 
pool may not represent the qualified labor pool from that group, and a 
disparate applicant selection rate may not show a disparate impact on the 
qualified labor pool.

i i The district court was less than clear about precisely what decision rules 
petitioners applied in selecting individuals for noncannery j o b s w h e t h e r  
there was in fact a preference for relatives, and whether objective criteria were 
actually applied by the various hiring officers. But any uncertainties on that 
score are irrelevant to evaluating the court of appeals finding of a prima facie 
case of disparate impact in the selection of the noncannery workforce as a 
whole, which was based on the proportion of minorities in the entire pool of 
persons selected for noncannery positions, whatever the selection device 
employed.

577



19

any finding that a significant number of minorities had been 
deterred from applying or that, for any other reason, applicant 
pool data were unreliable.28

The court of appeals relied only on comparisons between, on 
the one hand, the number of minorities in petitioners’ noncan­
nery jobs and, on the other hand, the number in petitioners’ 
cannery jobs, the number in the Alaska salmon canning in­
dustry as a whole, and the number in petitioners’ workforce as a 
whole. Because cannery workers made up a large portion of the 
latter two pools, the sufficiency of the evidence turns on 
whether the pool of cannery workers fairly represented the rele­
vant labor pool for various noncannery jobs. The court of ap­
peals had before it no findings sufficient to conclude that it 
did.29

In particular, the court of appeals had before it no findings, 
and no basis to believe, that cannery workers made up more 
than a small portion of the entire relevant labor pool for un­
skilled potential noncannery positions, or that the cannery 
workers were representative of that pool. Nor did the court have 
before it any findings, or any basis to believe, that cannery 
workers were even part of the relevant qualified labor pool for 
skilled positions. To the contrary, the district court found that 
respondents’ statistics had “little probative value” for the skilled 
jobs (Pet. App. 1114) and that petitioners’ statistics had 
significantly more probative value than respondents’ even fof 
the unskilled jobs (id. at 1119-1120).J0

J* The district court did note that there was some evidence of individual in­
stances of respondents feeling deterred from applying for noncannery posi­
tions (Pet. App. II16-1118). Without reference to deterrence, the court also 
found that cannery workers generally showed little interest in applying for 
noncannery jobs ( id . at 140). It found, too, that employees and nonemployees 
“were free to apply for any job for which they feel qualified’’ ( id . at 133).

19 The importance of the requirement of such findings, and of the overall 
adequacy of the statistical case, is highlighted in cases where disparate impact 
analysis is applied to subjective selection processes, as the nature of such proc­
esses may make the judicial inquiry into business justification particularly dif­
ficult.

,0 The district court also correctly criticized respondents’ statistics for not 
controlling for the substantial number of persons who were hired under the



20

In fact, the district court’s findings strongly suggest the op­
posite conclusions from those drawn by the court of appeals.31 
As to unskilled noncannery positions, the district court found 
that the best evidence of the relevant labor pool showed that 
90% of that pool was white and only 1% was either Filipino or 
Alaska Native (Pet. App. 136-137, 1110-1111, 1119-1120). Those 
figures show that the cannery workers, most of whom were 
Filipinos or Alaska Natives, were not representative of the rele­
vant unskilled labor pool; and it is apparently not contended, in 
light of the actual minority representation in the noncannery 
workforce (see Pet. 4; Br. in Opp. 1-2; Pet. App. 143-145), that 
those figures would sustain a prima facie case of disparate im­
pact.32 As to skilled noncannery jobs, the district court found 
that such jobs required skills (or preseason availability) not 
possessed or readily acquirable or acquired on the job by can­
nery workers and, indeed, that cannery workers and laborers do 
not make up a labor pool for other jobs {id. at 130, 135-136, 
139-141, 147, 1107-1109, 1113).

Notably, the district court found that there was an obvious 
explanation for the disparity disclosed by respondents’ statistics 
(that is, a significant difference in racial composition of the can­
nery and noncannery workforces). Simply put, “(tjhe institu­
tional factor of Local 37’s overrepresentation of non-whites ac­
counts for this statistic” (Pet. App. 142; id. at 135-136, 1110). 
That obvious cause of minority overrepresentation in the can­
nery jobs explains the disparities to which respondents point,

rehire preference (Pet. App. 1120-1122). That preference was upheld by the 
lower courts, a ruling that is not challenged in this Court.

ii The district court did not apply disparate impact analysis to the selection 
of noncannery workers generally, and there is therefore no finding that 
respondents’ statistics did not make out a prima facie case under the disparate 
impact model. For that reason, and because there is some confusion in the 
district court’s findings, we suggest that this Court should remand on that 
issue for the court of appeals to determine whether a further remand to the 
district court is needed for additional factual findings.

>i The district court did find respondents’ statistics sufficient to make out a 
prima facie case of disparate t r e a tm e n t as to unskilled jobs (Pet. App. 
1111 1112).

579



21

without suggesting that there is underrepresentation, or that 
there are exclusionary practices, in the noncannery jobs. In­
deed, according to the district court’s findings, if petitioners 
ceased using Local 37 as a hiring channel for cannery jobs, the 
intra-workforce stratification would apparently disappear or 
dwindle to insignificance —and with it the presence of large 
numbers of Filipinos and Alaska Natives in the industry 
workforce as a whole —even if there were no change whatever in 
the methods of selecting noncannery workers. If there would be 
no case of disparate impact alleging exclusion from the noncan­
nery jobs in that circumstance, surely there should be no liabili­
ty simply because petitioners have hired disproportionately 
large numbers of minorities for the cannery jobs.

II. AFTER A PLAINTIFF MAKES OUT A PRIMA FACIE 
CASE SHOWING THAT AN IDENTIFIED SELECTION 
MECHANISM CAUSES A DISPARATE IMPACT, THE 
EMPLOYER HAS THE BURDEN OF PRODUCING 
ENOUGH EVIDENCE TO SUSTAIN A JUDGMENT IN 
ITS FAVOR THAT THE CHALLENGED MECHANISM 
SIGNIFICANTLY SERVES LEGITIMATE BUSINESS 
GOALS, AND THE PLAINTIFF MAY THEN PREVAIL BY 
PROVING THE CONTRARY OR BY SHOWING THAT 
AN ALTERNATIVE PRACTICE WITH A LESS 
DISPARATE IMPACT EQUALLY SERVES THOSE 
GOALS

A holding that respondents failed to make out a prima facie 
case would make unnecessary any further analysis of the 
disparate impact challenge to the selection of the noncannery 
workforce as a whole. If this case is to be remanded, however, 
as we suggest (see note 31, supra), it would be appropriate for 
the Court to address some of the questions about disparate im­
pact analysis that the Watson case left unresolved. The plurality 
opinion in Watson furnishes a proper framework for answering 
those questions.

A. In addition to making a statistical case of disparate im­
pact in selection, a plaintiffs prima facie case challenging an 
employer’s adverse selection decision must identify the decision

580



22

process that was actually used to make hiring decisions. See 
Watson, slip op. 13 (plurality opinion) (“[tjhe plaintiff must 
begin by * * * isolating and identifying the specific employment 
practices that are allegedly responsible for any observed 
statistical disparities”); see also id. at 2 n.2 (Blackmun, J., con­
curring in the judgment); Pet. App. VI9-V20. Thus, as part of 
their prima facie case, respondents had to identify the process 
for selecting noncannery workers —whether by subjective judg­
ment by individual hiring officers or by the application of objec­
tive criteria or a policy of nepotism. To be sure, certain 
nonselection practices may be relevant to determining the rele­
vant labor pools —for example, if certain on-the-job practices 
such as segregated housing deter applicants, the applicant pool 
may not be a proper measure of disparate impact. But practices 
that are not part of the selection mechanism (including the 
failure to use proposed alternatives) are not themselves properly 
subject to disparate impact analysis in a selection case.

Of course, a decision rule for selection may be complex: it 
may, for example, involve consideration of multiple factors. 
And certainly if the factors combine to produce a single ultimate 
selection decision and it is not possible to challenge each one, 
that decision may be challenged (and defended) as a whole.” 
But disparate impact analysis is designed to root out “ ‘built-in 
headwinds’ ” and “barriers” to selection (Griggs, 401 U.S. at 
432; see also Teal, 457 U.S. at 440), and not otherwise needless­
ly to intrude upon employer practices (see United Steelworkers 
o f America v. Weber, 443 U.S. 193, 206 (1979)). Hence, in its 
disparate impact decisions, this Court has properly focused on 
the specific devices or processes, including subjective ones, that

»  W e  do not here address whether, If an employer uses a multifactor deci­
sion process and the plaintiff proves disparate impact of the entire process, the 
plaintiff is required, in order to make out a prima facie case, also to test each 
component for disparate impact where that is possible. C o n n e c t ic u t v. T e a l  

says that the plaintiff m a y do so in a multistage process. Whatever the scope of 
the third question in the petition, we limit ourselves to the point that only an 
employer’s selection device or devices are subject to challenge in a disparate 
impact selection case. Other employer practices, if subject to challenge, must 
be separately challenged.

581



23

the employer uses to select employees, not on the employer’s 
overall employment policies, including nonselection practices 
Albemarle Paper Co. (employment tests and seniority systems); 
Washington v. Davis, 426 U.S. 229 (1976) (aptitude tests); 
Dothard (height and weight requirements); Teal (written ex­
amination); Beazer (methadone user exclusion); Watson (sub­
jective judgment by supervisor).

B. Once a prima facie case has been made out, judicial in­
quiry into the ultimate question whether the challenged nonhir­
ing was “because o f’ race moves to the next two stages of 
disparate impact analysis —the first focusing on the justification 
for the selection device that produced the adverse selection deci­
sion, the second focusing on the availability of alternatives to 
the challenged practice that have lesser racial impact. 
Analytically, the two stages are closely related: they are both in­
gredients of the Title VII concept of business justification, 
because a challenged practice that causes a disparate racial im­
pact is not justifiable —even if it is well-supported by business 
reasons —if there are equally good alternatives to the practice 
that cause a lesser impact. We discuss three aspects of the in­
quiry into business justification in an effort to identify a fair 
and workable approach to the inquiry.

1. This Court’s decisions have used different formulations 
of the substantive standard governing the first stage of the in­
quiry into business justification that is required once a plaintiff 
makes out a prime facie case. See, e.g., Teal, 457 U.S. at 446 
(citation omitted) (“ ‘manifest relationship to the em­
ployment’ ”); Albemarle Paper Co., 422 U.S. at 425 (“ ‘job 
related’ ”); Dothard, 433 U.S. at 329, 331-332 & n.14 (“job 
related”; “necessary to safe and efficient job performance”; 
“essential to good job performance”); Griggs, 401 U.S. at 431, 
432 (“business necessity”; “manifest relationship to the employ­
ment”). Because those varying formulations suggest either 
higher or lower thresholds of justification, it would be useful 
for this Court to adopt a single governing formulation to

582



24

guide judicial application.34 Most recently, in Watson, the 
plurality indicated (slip op. 17, 18) that “legitimate business 
reasons” would suffice to show a “ ‘manifest relationship to the 
employment.’ ” We think that the emphasis on reasonableness 
that is reflected in that approach was usefully encapsulated in 
the formulation this Court used when it found sufficient 
business justification in New York Transit Authority v. Beazer, 
440 U.S. at 587 n.31, where the Court stated: the employer’s 
“legitimate employment goals of safety and efficiency * * * are 
significantly served by —even if they do not require-(the 
challenged selection rule].”

That standard does not permit a justification based on a non­
business reason or on a negligible contribution to a business 
purpose. So low a standard would threaten to undermine Title 
VIl’s concern to “promote hiring on the basis of job qualifica­
tions” (Griggs, 401 U.S. at 431, 434) and its use to root out ex­
clusionary practices that are “functionally equivalent to inten­
tional discrimination,” even though intent cannot be proved

14 The different terms used by this Court have led the courts of appeals to 
articulate different standards as well. See, e g . ,  Kinsey v. First Regional 
Securities, Inc., 557 F.2d 830, 837 (D.C. Cir. 1977) (citation omitted) (practice 
must have an “ ‘overriding legitimate business purpose such that the practice is 
necessary to the safe and efficient operation of the business’ ”); Burwell v. 
Eastern A ir  Lines, Inc., 633 F.2d 361,370 (4th Cir. 1980) (practice must bear a 
“manifest relation to the * * * employment"), cert, denied, 450 U.S. 965 
(1981); Parson  v. K aiser A lum inum  A Chem ical C orp., 575 F.2d 1374, 1389 
(5th Cir. 1978) (citation and emphasis omitted) (practice must “ ‘foster safety 
and efficiency * * * (and) be essential to that goal’ ”), cert, denied, 441 U.S. 
968 (1979); Chrisner v. C om plete  A u to  Transit, Inc., 645 F.2d at 1262 (“in­
dispensability is not the touchstone"; “practice must substantially promote the 
proficient operation of the business’’); Aguilera v. C ook C ounty Police A C or­
rections M erit Board, 760 F.2d 844, 847 (7th Cir.) (practice must be 
“reasonable” or “efficient”), cert, denied, 474 U.S. 907 (1985); Kirby v. C ol­
o n y  Furniture C o ., 613 F.2d 696, 703 (8th Cir. 1980) (practice must be shown 
to be necessary to safe and efficient job performance); W ambheim  v. J.C. 
P enney C o ., 705 F.2d 1492, 1495 (9th Cir. 1983) (citation omitted) (practice 
must have “ ‘legitimate and overriding business considerations' ”), cert, 
denied, 467 U.S. 1255 (1984); Williams v. C olorado Springs School Dist. No. 
I I ,  641 F.2d 835, 842 (10th Cir. 1981) ("practice must be essential, the purpose 
compelling”).

583



25

(Watson, slip op. 6). At the same time, the Beazer standard does 
not require that the selection mechanism be absolutely essential 
to the business. So high a standard would not only be virtually 
impossible to meet but would threaten to put pressure on 
employers to avoid disparate impact liability by adopting quotas 
or otherwise turning their attention away from job qualifica­
tions and toward numerical balance. See Watson, slip op. IS 
(plurality opinion). Indirectly compelling those results in the 
name of Title VII is not consistent with the statute, which does 
not contemplate so serious an intrusion on managerial 
prerogatives. See id. at 12; Johnson  v. Transportation Agency, 
No. 85-1129 (Mar. 25, 1987), slip op. 11 n.7; Weber, 443 U.S. at 
204-207 & n.7; Albemarle Paper Co., 422 U.S. at 449 
(Blackmun, J., concurring in the judgment); Griggs, 401 U.S. at 
431, 434. The Beazer standard strikes a reasonable balance.15

2. As Justice Blackmun explained in his concurring opinion 
in Watson, slip op. 2, many of this Court’s decisions in 
disparate impact cases use language that can be and often has 
been read to mean that the employer assumes the burden of per­
suasion on the question of business justification once a prima 
facie case has been made out. See Albemarle Paper Co., 422 
U.S. at 425 (employer must “meet the burden of proving that its 
tests are ‘job related’ ”); Dothard, 433 U.S. at 329 (employer 
must “provel] that the challenged requirements are job 
related”); Griggs, 401 U.S. at 432 (employer has "the burden of 
showing” a manifest relationship to the job). But that language

15 That standard should, of course, be applied with an appreciation of the 
problems of proving the precise contribution of particular selection devices to 
discerning important qualifications, especially “personal qualities that have 
never been considered amenable to standardized testing” ( W a t s o n ,  slip op. 18 
(plurality opinion)). Hence, the business justification standard does not entail 
a requirement of formal validation. W a ts o n , slip op. 17 (plurality opinion); 
see id . at 7-8 (Blackmun, J., concurring in the judgment). In addition, the 
standard may be satisfied somewhat indirectly —for example, by a sufficient 
relationship, not directly to the job at issue, but to a legitimate training pro­
gram. W a s h in g to n  v. D a v is , 426 U.S. at 250-252. It should also “be borne in 
mind that ‘(cjourts are generally less competent than employers to restructure 
business practices’ ” ( i b i d . ,  quoting F tir n c o  C o n s t r .  C a r p .  v. W a te r s , 438 U.S. 
567, 578 (1978)).

584



26

is ambiguous, as “burden of proof’ and “showing” may be used 
to refer either to a burden of persuasion or to a burden of pro­
duction. See E. Cleary, M c C o r m ic k  o n  Evidence § 336, at 
783-784 (2d ed. 1972). For example, in N L R B  v. T ra n sp o r ta tio n  
M g m t. C o r p . ,  462 U.S. 393, 404 n.7 (1983), the Court ruled that 
“burden of proof’ in the Administrative Procedure Act, 5 
U.S.C. 556(d) (§ 7(c)), meant only “the burden of going for­
ward, not the burden of persuasion.” And in the disparate treat­
ment context, several of the Court’s decisions referred to the de­
fendant’s burden to “prove” (F u rn co  C o n s tr . C o rp . v. W a te rs , 
438 U.S. 567, 577 (1978)) or to “show” (M c D o n n e ll D o u g la s  
C o r p .  v. G re e n , 411 U.S. 792, 804 (1973)) a nondiscriminatory 
reason for a challenged employment decision, but the Court 
then made it clear that the employer’s burden was one of pro­
duction, not of persuasion (T exas D e p 't  o f  C o m m u n ity  A f fa ir s  
v. B u rd in e , 450 U.S. 248, 256-258 (1981); B o a rd  o f  T ru stees  v. 
S w e e n e y , 439 U.S. 24, 24-26 (1978)).

We agree with the plurality in W a tso n  (slip op. 17) that the 
same result should apply in the disparate impact context. Leav­
ing the burden of persuasion on the plaintiff is consistent with 
the general rule (see Restatement (Second) of Torts § 433B 
(1965)) that a plaintiff at all times bears the burden of per­
suading the trier of fact on the basic causation element of a 
violation —here, that the nonhiring was “because o f ’ race rather 
than for a sound business reason. Lack of business justification 
is a fundamental element of the violation under the disparate 
impact theory of Title VII liability, and a plaintiff alleging 
disparate impact has the “ultimate burden of proving a violation 
of Title VII” (B e a ze r , 440 U.S. at 587 n.31). Moreover, impos­
ing only a burden of production keeps the disparate impact 
proof scheme in accord with the norm recognized in Fed. R. 
Evid. 301, which states that, unless otherwise provided by 
statute or rule, a “presumption” (here, a presumption of 
discrimination that arises from a prime facie case) shifts only 
the burden of going forward, not the burden of persuasion.56 In

>« Sec B u r d in e . 450 U.S. at 255 n.8. That Fed. R. Evid. 301 is relevant does 
not mean that it is controlling. The Court stated in N L R B  v. T r a n s p o r ta t io n  
M g m t .  C o rp ., 462 U.S. at 404 n.7, that the rule “in no way restricts the 
authority of a court or an agency to change the customary burdens of persua­
sion in a manner that otherwise would be permissible.”



27

addition, the “strong (congressional) desire to preserve 
managerial prerogatives” (Weber, 443 U.S. at 204-207) that is 
embodied in Title VII counsels against a rule that imposes 
liability unless an employer carries a burden of persuasion to 
justify its business practices.

Nothing about disparate impact cases justifies a departure 
from the model for litigating disparate treatment cases. In 
disparate impact cases, as in disparate treatment cases, the 
employer’s “explanation of its legitimate reasons must be clear 
and reasonably specific” (Burdine 450 U.S. at 258); the plaintiff 
has liberal access to discovery from the employer; and the 
employer has an incentive to persuade the trier of fact of the 
justification for its practice (which has already been shown to 
have a disparate impact). See ibid. Once the employer produces 
evidence of business justification, the plaintiff may, of course, 
introduce contrary evidence, including testimony by experts and 
by employees themselves, concerning what qualifications are 
truly related to job performance. If the risk of nonpersuasion as 
to the employer’s state of mind does not “unduly hinder” plain­
tiffs in disparate treatment cases (ibid.), neither should plain­
tiffs in disparate impact cases be unduly hindered by carrying 
the risk of nonpersuasion as to the business justification for the 
challenged selection device.

Finally, given an agreed-upon substantive standard for the 
first-stage inquiry into business justification, the plaintiffs 
bearing of the risk of nonpersuasion should tip the balance 
against the plaintiff only in a limited class of cases. As in 
disparate treatment cases, the burden of production requires 
that the employer put forth evidence that is “legally sufficient to 
justify a judgment for the defendant” (Burdine, 450 U.S. at 
255). The burden of persuasion requires more; but because the 
issue is governed by a preponderance-of-the-evidence standard, 
the allocation of the risk of nonpersuasion should alter the 
result only in marginal cases. Moreover, even if the plaintiff 
fails to persuade the trier of fact that the challenged practice 
does not meet the threshold business-justification standard, 
because it does not significantly serve legitimate business goals, 
the plaintiff may still prevail by showing that an alternative 
exists, as we discuss below.



28

In short, the Court should recognize a parallelism between 
disparate impact and disparate treatment analysis. The distinc­
tive questions presented in a disparate impact case “do not imply 
that the ultimate legal issue is different than in cases where 
disparate treatment analysis is used” ( Watson, slip op. 6). In­
deed, because the disparate impact concept of discrimination is 
an alternative to disparate treatment, which was the “most ob­
vious evil Congress had in mind when it enacted Title VII” 
( Teamsters, 431 U.S. at 335 n.15), it would be anomalous to 
shift the burden of persuasion on a critical issue in a disparate 
impact case when no such shifting occurs in a disparate treat­
ment case. As the Court explained in Watson (slip. op. 6), an 
employer should not be held “liable for unintentional 
discrimination on the basis of less evidence than is required to 
prove intentional discrimination.”

3. If the burden of persuasion on business justification re­
mains with the plaintiff, and the employer meets the.burden of 
production on business justification, the plaintiff may still 
prevail by putting forth sufficient evidence to persuade the trier 
of fact that the employer’s claim of business justification is un­
convincing. Even if the plaintiff does not overcome the 
employer’s claim of business justification, however, the plaintiff 
can still prevail by showing “ ‘that other tests or selection 
devices, without a similarly undesirable racial effect, would also 
serve the employer’s legitimate interest in efficient and trust­
worthy.workmanship’ ” (Watson, slip op. 17 (plurality opinion) 
(quoting Albemarle Paper Co., 422 U.S. at 425)).

To meet the plaintiffs burden on the issue of alternative 
devices, it should not suffice merely to establish that there is 
some alternative selection procedure that has a less disparate im­
pact. Rather, the plaintiff must show that the proposed alter­
native would serve the employer’s business goals as effectively as 
the selection mechanism under challenge. “Factors such as the 
cost or other burdens of proposed alternative selection devices 
are relevant in determining whether they would be equally as ef­
fective as the challenged practice in serving the employer’s 
legitimate business goals” (Watson, slip op. 17 (plurality opin­

587



29

ion)). SeeFurnco Constr. Corp. v. Waters, 430 U.S. at 577-578. 
But if the employer shows that a test or other selection device is 
job-related, the plaintiff should be allowed nonetheless to 
secure at least prospective relief by proffering a less 
discriminatory test or device that equally serves the employer’s 
purposes. The failure to use such an alternative demonstrates 
that the employer’s present practice is not truly justified in 
business terms.37

CONCLUSION
The judgment of the court of appeals should be vacated. 
Respectfully submitted.

C harles Fried 
S o lic ito r  G eneral

W m . Bradford  Reynolds 
A ssistan t A tto rn e y  G eneral

Roger Clegg
D ep u ty  A ssis tan t A tto rn e y  G eneral

Richard  G. T aranto  
A ssistan t to  the Solicitor G eneral

David  K. Flynn 
L isa J. Stark 

A tto rn e ys

September 1988

57 As the W a ts o n  plurality observed of factors such as cost (slip op. 17), the 
ready availability of equally effective alternatives “would also be relevant in 
determining whether the challenged practice has operated as the functional 
equivalent of a pretext for discriminatory treatment.”

588



Law Reprints
5442 30th St., NW 
Washington, DC 20015