Wards Cove Packing Company, Inc. v. Atonio Brief Amicus Curiae in Support of Petitioners

    ■■ ■
No. 87-1387 |  | |

• i ;v :>

* • .1 .v*'
'a**.... ■ ■ ■ ■

• . V  .’3

3(n tljc j&uprrmc Court ot tl)r © nlttb s ta te s  j
October Term, 1988 * S  § f

f®
• • VJ-

Wards Cove Packing Company, Inc., et al., petitioners >

Frank Atonio, et al. . '
■ • •' \  '  • : ^  " ' V i n

_________________ _____________  -• '•" ‘ .r  • &*■'. * * •: > '/ a* > t
■ ■ :

/ i,'
• / .

ON WRIT OF CERTIORARI 
TO THE UNITED STA TES COURT OF APPEALS 

FOR THE NINTH CIRCUIT

BRIEF FOR THE UNITED STATES AS AMICUS CURIAE 
SUPPORTING PETITIONERS

■ ' - ' ' ' I  

'

-  • m

N..1Q
Charles Fried 

Solicitor General
Wm. Bradford Reynolds 

Assistant Attorney General 
R6ger Clego

Deputy Assistant Attorney General ' 7
Richard g . Taranto 

Assistant to the Solicitor General 
David K. Flynn a 
Lisa J. Stark

Attorneys .<

1

■ . .v5.ii■ ; . v >

f . >4
' rv • ^

V

a  :-i
Department of Justice 
Washington, D C. 20530 
(202) 633-2217 ■. ■ ’j. ■■■< •'

}



QUESTIONS PRESENTED

1 ,n fhis 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.

( 1)



TABLE OF CONTENTS

Page
Interest of the United States...................................................  1
Statement ............................................................................... 1
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

TABLE OF AUTHORITIES

Cases:
Aguilera v. Cook County Police <6 Corrections Merit

Board, 760 F.2d 844 (7th Cir.), cert, denied, 474 U.S.
907 (1985)......... .................................... ..................... 24

Albemarle Paper Co. v. Moody, 422 U.S. 405 (1975) . . . .  13, 14,
17, 23, 25, 28

Board of Trustees v. Sweeney, 439 U.S. 24 (1978)........... 26
Burwell v. Eastern Air Lines, Inc., 633 F.2d 361 (4th Cir.

1980), cert, denied, 450 U.S. 965 (1981)........................ 24
Chrisner v. Complete Auto Transit, Inc., 645 F.2d 1251

(6th Cir. 1981)..............................................................  10, 24
Connecticut v. Teal, 457 U.S. 440(1982)........................ 9, 13,

14, 22, 23
Dothard\. Rawlinson, 433 U.S. 321 (1977)....................  13, 17,

23, 25
EEOCv. Rath Packing Co., 787 F.2d 318 (8th Cir. 1986), 

cert, denied, No. 86-67 (Oct. 14, 1986)........................ 17

(III)



IV

Cases —Continued: Page
Eubanks v. Pickens-Bond Conslr. 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
Hammon 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. NLRB, 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. Gulf 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
NLRB v. Transportation Mgmt. Corp., 462 U.S. 393

(1983)........................................................................... 26
Parson v. Kaiser Aluminum A 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

V

Cases —Continued: Page
Teamstersv. United States, 431 U.S. 324(1977)..............  13, 17,

18, 28
Texas Dep’t o f Community Affairs v. Burdine, 450 U.S.

248 (1981) .................................................................... 26, 27
United States v. County o f Fairfax, 629 F.2d 932 (4th Cir.

1980), cert, denied, 449 U.S. 1078(1981)...................... 17
United Steelworkers o f America v. Weber, 443 U.S. 193

(1979) .............................................. .................... 22, 25, 27
Wambheim v. J.C. Penney Co., 705 F.2d 1492 (9th Cir.

1983), cert, denied, 467 U.S. 1255 (1984)...................... 24
Washington v. Davis, 426 U.S. 229 (1976)...................... 23, 25
Watson v. Fort Worth Bank <5 Trust, No. 86-6139 (June

29, 1988) ...................................................................... passim
Wheeler v. City o f Columbus, 686 F.2d 1144 (5th Cir.

1982) ...........................................................................  18
Williams v. Colorado Springs School Dist. No. 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. VII, 42 U.S.C. 2000e

el seq.............................................................................. 1
42 U.S.C. 2000e-2(a)(l)............................................  13
42 U.S.C. 2000e-2(a)(2)............................................  13
42 U.S.C. 2000e-2(j).................................................  18

29C.F.R. Pt. 1607 ........................................................... 17
Fed. R. Evid. 301 . . .'.......................................................  26

Miscellaneous:
E. Cleary, McCormick on Evidence (2d ed. 1972)........... 26
44 Fed. Reg. 11998 (1979) .. '............................................  17
Restatement (Second) of Torts (1965).............................  26 '



3)n tfjc Supreme Court of tlje Umteb states
O ctober T erm , 1988

No. 87-1387

Wards Cove P acking Company, Inc., et al., petitioners

v.

Frank Atonio, et al.

ON WRIT OF CER TIORARI 
TO THE UNITED ST A TES COURT OF A P REALS 

FOR THE NINTH CIRCUIT

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 VII of the Civil Rights Act of 1964, 
42 U.S.C. 2000e el seq. The Attorney General has significant 
Title VII enforcement responsibilities. The United States, as the 
nation’s largest employer, is also subject to Title VII 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. at 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 1113-1114). 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). 1

( 1)



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).1

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.).1 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. 11-1130).

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 At issue in this case are the jobs other than those on the 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% (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.

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).* 1 * 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).

J The court found (Pet. App. 140): “There has been a general lack of interest 
by cannery workers in applying for noncannery workers jobs.”

* 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



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, I55-I76).1 * * * * 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 13?).

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 “[t]he 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 (id. 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 
(id. 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 experience 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 (id. at 175-176; see also id. at 112-113 (correcting list)).

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.8 * 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 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 
(id. at 184-194); found that petitioners’ dining practices originated in peti­
tioners’ deference to the leadership of Local 37 (id. 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 “(r)elatives 
of whites and particular! | nonwhites appear in high incidence at the canneries” 
(id. 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” (ibid ). The court also noted
that “numerous white persons who ‘knew’ someone were not hired due to inex­
perience” (id. at 1122).



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.9

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 I I10-1111). 
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.

9 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) departmental^ because the various departments worked the same 
shifts" (id. at 1125), and that “lilt 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 
1127-1129) 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).

7

at I I 11-1112). 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 I I13).10 *

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 The court observed that respondents “were general(l)y 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 evidence 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 industryl, as 
respondents) appeared to have understood” (Pel. App. I I15-1116). 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 1116-1II8).



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).12 In addition, few respondents 
made timely and proper applications for jobs.13 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).14

3. A three-judge panel of the court of appeals affirmed 
(Pet. App. Ill 1-II156). 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 of 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 III43-III54).15

12 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” (ibid.).

15 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

9

4. The court of appeals granted rehearing en banc and 
vacated the initial panel opinion (Pet. App. IV1-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 prelerence for 
relatives” (Pet. App. III56). The district court’s rejection of the challenge to 
the English-language requirement was not challenged on appeal (id. at 11146 
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 Connec­
ticut v. Teat. 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”).



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. VU-V144). 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-V17 (citations omitted))." 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 (id. 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 (id. 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 (id. 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. V17-V18, quoting Chrisner v. Complete Auto Transit. 
Inc., 645 F.2d 1251, 1262 (6th Cir. 1981)).

II

could serve the employer’s needs with a lesser impact on the pro­
tected class” (id. at VI9).

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. V118). 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).17 * 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 V122-VI23). As to respondents’ challenge to the

>» 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. VI21), because petitioners’ hiring of­
ficers are predominantly white.



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).21

INTRODUCTION AND SUMMARY OF ARGUMENT

Title Vll 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

20 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-VI32).

21 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” (id. 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).

13

race” (42 U.S.C. 2000e-2(a)(l) and (2)).22 Intentional dis­
crimination based on race is the primary way in which an em­
ployer can act unlawfully “because oF race. As the legislative 
history of the 1964 Act makes clear and as this Court has said, 
“(ujndoubtedly disparate treatment was the most obvious evil 
Congress had in mind when it enacted Title Vll.” 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” 
(IVatson 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

22 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. 2000e-2(a)(l) and (2)).



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 o r ’ race, is different' from the meaning given by the disparate 
treatment theory. Whereas the latter asks whether race actually motivated the 
nonhiring 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 oF’ race.

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-workforce 
stratification shown by respondents (i.e., 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.



16

ARGUMENT

T. 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 o f ’ 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

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).23 In some cases, however,

JJ See Metrocare v. Washington Metro. Area Transit Authority, 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”); Lewis v. NLRB. 750 F.2d 1266, 1275 (5th Cir. 1985); Grano v. 
Dep't o f Development, 637 F.2d 1073, 1078 (6th Cir. 1980); Piva v. Xerox 
Corp., 654 F.2d 591, 595 (9th Cir. 1981).

“  See Mammon v. Barry. 813 F.2d 412, 427 n.31 (D.C. Cir. 1987), cert, 
denied, No. 87-1150 (May 31, 1988); United States v. County o f Fairfax, 629 
F.2d 932, 940 (4th Cir. 1980), cert, denied, 449 U.S. 1078 (1981); Hester v. 
Southern Ry., 497 F.2d 1374, 1379 (5th Cir. 1974); Rowe v. Cleveland 
Pneumatic Co. Numerical Control, 690 F.2d 88, 93 (6th Cir. 1982); Mister v. 
Illinois Cent. Gulf R.R., 832 F.2d 1427, 1435 (7th Cir. 1987); EEOC v. Rath 
Packing Co.. 787 F.2d 318, 337 (8th Cir. 1986), cert, denied, No. 86-67 (Oct. 
14, 1986); Moore v. Hughes Helicopters,'Inc., 708 F.2d 475, 483 (9th Cir. 
1983).

Adverse impact under the Uniform Guidelines on Employee Selection Pro­
cedures, 29 C.F.R. Ft. 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).



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

26 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.

27 The district court was less than clear about precisely what decision rules
petitioners applied in selecting individuals for noncannery jobs — e.g., whether
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.

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.2'

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 for 
the unskilled jobs (id. at I I 19-1120).30

21 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. I I16-1118). 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, loo, that employees and nonemployees 
“were free to apply for any job for which they feel qualified" (id. at 133).

29 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.

70 The district court also correctly criticized respondents’ 
cuniruiiing foi ihe substantial number of persons who were

statistics for not 
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.* 11 12 
As to unskilled noncannery positions, the district court found 
that the best evidence of the relevant labor pool showed that 
90®7o 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, “(t]he 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.

11 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.

12 The district court did find respondents’ statistics sufficient to make out a 
prima facie case of disparate treatment as to unskilled jobs (Pet. App. 
III I II 12).

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



22

process that was actually used to make hiring decisions. See 
Watson, slip op. 13 (plurality opinion) (“[t]he 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. V19-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

V We 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. Connecticut v. Teal 
says that the plaintiff may 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.

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



24

guide judicial application.14 Most recently, in IVatson, 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 
Vll’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 Air 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. Kaiser Aluminum & Chemical Corp., 575 F.2d 1374, 1389 
(5th Cir. 1978) (citation and emphasis omitted) (practice must “ ‘foster safety 
and efficiency * * * [andj be essential to that goal' ”), cert, denied, 441 U.S. 
968 (1979); Chrisner v. Complete Auto 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. Cook County Police <4 Cor­
rections Merit Board. 760 F.2d 844, 847 (7th Cir.) (practice must be 
“reasonable” or “efficient”), cert, denied, 474 U.S. 907 (1985); Kirby v. Col­
ony Furniture Co.. 613 F.2d 696, 703 (8th Cir. 1980) (practice must be shown 
to be necessary to safe and efficient job performance); Wambheim v. J.C. 
Penney Co.. 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. Colorado Springs School Dist. No.
II, 641 F.2d 835, 842 (10th Cir. 1981) (“practice must be essential, the purpose 
compelling”).

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. 18 
(plurality opinion). Indirectly compelling those results in the 
name of Title Vll 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. 85-1 129 (Mar. 25, 1987), slip op. lln .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 “proved 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

>i 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" (Watson, slip op. 18 
(plurality opinion)). Hence, the business justification standard does not entail 
a requirement of formal validation. Watson, 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 (o a legitimate training pro­
gram. Washington v. Davis, 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’ ” (ibid., quoting Furnco Constr. Corp. v. If aters, 438 U.S. 
567, 578 (1978)).



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, McCormick on Evidence § 336, at 
783-784 (2d ed. 1972). For example, in NLRB v. Transportation 
Mgmt. Corp., 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” (Furnco Constr. Corp. v. Waters, 
438 U.S. 567, 577 (1978)) or to “show” (McDonnell Douglas 
Corp. v. Green, 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 (Texas Dep’t o f Community Affairs 
v. Burdine, 450 U.S. 248, 256-258 (1981); Board o f Trustees v. 
Sweeney, 439 U.S. 24, 24-26 (1978)).

We agree with the plurality in Watson (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” (Beazer, 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.36 In

>6 See Burdine, 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 NLRB v. Transportation 
Mgmt. Corp., 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

29

ion)). See Furnco 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.”

CONCLUSION

The judgment of the court of appeals should be vacated. 
Respectfully submitted.

C harles Fried 
Solicitor General 

W m . Bradford Reynolds 
Assistant Attorney General 

Roger C legg
Deputy Assistant Attorney General 

Richard  G. Taranto 
Assistant to the Solicitor Genera! 

David K. Flynn 
L isa J. Stark 

Attorneys

September 1988

J' As the liaison 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.”

w  U S GOVERNMENT PRINTING OFFICE: 1988-202 037/60644