6 Pattern-or-Practice Discrimination 6 Pattern-or-Practice Discrimination
(challenging discrimination that is systemic in the employer's hiring practice, rather than targetting discrimination in an individual case)
6.1 International Brotherhood of Teamsters v. United States 6.1 International Brotherhood of Teamsters v. United States
INTERNATIONAL BROTHERHOOD OF TEAMSTERS v. UNITED STATES et al.
No. 75-636.
Argued January 10, 1977
Decided May 31, 1977*
*327 L. N. D. Wells, Jr., argued the cause for petitioner in No. 75-636. With him on the briefs were David Previant and G. William Baab. Robert D. Shuler argued the cause for petitioner in No. 75-672. With him on the brief was John W. Ester.
Deputy Solicitor General Wallace argued the cause for the United States et al. in both cases. With him on the brief were Solicitor General Bork, Assistant Attorney General Pottinger, Thomas S. Martin, Brian K. Landsberg, David L. Rose, William B. Fenton, Jessica Dunsay Silver, and Abner W. Sibal. †
delivered the opinion of the Court.
This litigation brings here several important questions under Title VII of the Civil Rights Act of 1964, 78 Stat. 253, as amended, 42 U. S. C. § 2000e et seg. (1970 ed. and Supp. V). The issues grow out of alleged unlawful employment practices engaged in by an employer and a union. The employer is a common carrier of motor freight with nationwide operations, and the union represents a large group of its employees. The District Court and the Court of Appeals held that the employer had violated Title VII by engaging in a pattern and practice of employment discrimination against Negroes and Spanish-surnamed Americans, and that the union had violated the Act by agreeing with the employer to create and maintain a "seniority system that perpetuated the effects of past racial and ethnic discrimination. In addition to the basic questions presented by these two rulings, other subsidiary issues must be resolved if violations of Title VII occurred — issues concerning the nature of the relief to which aggrieved individuals may be entitled.
I
The United States brought an action in a Tennessee federal court against the petitioner T. I. M. E.-D. C., Inc. (company), pursuant to § 707 (a) of the Civil Rights Act of 1964, 42 U. S. C. § 2000e-6 (a).1 The complaint charged that the *329company had followed discriminatory hiring, assignment, and promotion policies against Negroes at its terminal in Nashville, Tenn.2 The Government brought a second action against the company almost three years later in a Federal District Court in Texas, charging a pattern and practice of employment discrimination against Negroes and Spanish-surnamed persons throughout the company’s transportation system. The petitioner International Brotherhood of Teamsters (union) was joined as a defendant in that suit. The two actions were consolidated for trial in the Northern District of Texas.
The central claim in both lawsuits was that the company had engaged in a pattern or practice of discriminating against minorities in hiring so-called line drivers. Those Negroes and Spanish-surnamed persons who had been hired, the Government alleged, were given lower paying, less desirable jobs as servicemen or local city drivers, and were thereafter discriminated against with respect to promotions and transfers.3 In *330this connection the complaint also challenged the seniority-system established by the collective-bargaining agreements between the employer and the union. The Government sought a general injunctive remedy and specific “make whole” relief for all individual discriminatees, which would allow them an opportunity to transfer to line-driver jobs with full company seniority for all purposes.
The cases went to trial4 and the District Court found that *331the Government had shown “by a preponderance of the evidence that T. I. M. E.-D. C. and its predecessor companies were engaged in a plan and practice of discrimination in violation of Title VII . . . .” 5 The court further found that the seniority system contained in the collective-bargaining contracts between the company and the union violated Title VII because it “operate [d] to impede the free transfer of minority groups into and within the company.” Both the company and the union were enjoined from committing further violations of Title VII.
With respect to individual relief the court accepted the Government’s basic contention that the “affected class” of discriminatees included all Negro and Spanish-surnamed incumbent employees who had been hired to fill city operations or serviceman jobs at every terminal that had a line-driver operation.6 All of these employees, whether hired before or after the effective date of Title VII, thereby became entitled to preference over all other applicants with respect to consideration for future vacancies in line-driver jobs.7 Finding that members of the affected class had been injured in different degrees, the court created three subclasses. Thirty persons who had produced “the most convincing evidence of discrimination and harm” were found to have suffered “severe injury.” The court ordered that they be offered the opportunity to fill line-driver jobs with competitive seniority dating back to July 2, *3321965, the effective date of Title VII.8 A second subclass included four persons who were “very possibly the objects of discrimination” and who “were likely harmed,” but as to whom there had been no specific evidence of discrimination and injury. The court decreed that these persons were entitled to fill vacancies in line-driving jobs with competitive seniority as of January 14, 1971, the date on which the Government had filed its systemwide lawsuit. Finally, there were over 300 remaining members of the affected class as to whom there was “no evidence to show that these individuals were either harmed or not harmed individually.” The court ordered that they be considered for line-driver jobs 9 ahead of any applicants from the general public but behind the two other subclasses. Those in the third subclass received no retroactive seniority; their competitive seniority as line drivers would begin with the date they were hired as line drivers. The court further decreed that the right of any class member to fill a line-driver vacancy was subject to the prior recall rights of laid-off line drivers, which under the collective-bargaining agreements then in effect extended for three years.10
*333The Court of Appeals for the Fifth Circuit agreed with the basic conclusions of the District Court: that the company had engaged in a pattern or practice of employment discrimination and that the seniority system in the collective-bargaining agreements violated Title VII as applied to victims of prior discrimination. 517 F. 2d 299. The appellate court held, however, that the relief ordered by the District Court was inadequate. Rejecting the District Court's attempt to trisect the affected class, the Court of Appeals held that all Negro and Spanish-surnamed incumbent employees were entitled to bid for future line-driver jobs on the basis of their company seniority, and that once a class member had filled a job, he could use his full company seniority — even if it predated the effective date of Title VII — for all purposes, including bidding and layoff. This award of retroactive seniority was to be limited only by a “qualification date” formula, under which seniority could not be awarded for periods prior to the date when (1) a line-driving position was vacant,11 and (2) the class member met (or would have met, given the opportunity) the qualifications for employment as a line driver.12 Finally, *334the Court of Appeals modified that part of the District Court's decree that had subjected the rights of class members to fill future vacancies- to the recall rights of laid-off employees. Holding that the three-year priority in favor of laid-qff workers “would unduly impede the eradication of past discrimination,'' id., at 322, the Court of Appeals ordered that class members be allowed to compete for vacancies with laid-off employees on the basis of the class members' retroactive seniority. Laid-off line drivers would retain their prior recall rights with respect only to “purely temporary'' vacancies. Ibid. 13
The Court of Appeals remanded the case to the District Court to hold the evidentiary hearings necessary to apply these remedial principles. We granted both the company's and the union's petitions for certiorari to consider the significant questions presented under the Civil Rights Act of 1964, 425 U. S. 990.
II
In this Court the company and the union contend that their conduct did not violate Title VII in any respect, asserting first that the evidence introduced at trial was insufficient to show that the company engaged in a “pattern or practice” of employment discrimination. The union further contends that the.seniority system contained in the collective-bargaining agreements in no way violated Title VII. If these contentions are correct, it is unnecessary, of course, to reach any of the issues concerning remedies that so occupied the attention of the Court of Appeals.
A
Consideration of the question whether the company engaged in a pattern or practice of discriminatory hiring prac*335tices involves controlling legal principles that are relatively clear. The Government’s theory of discrimination was simply that the company, in violation of § 703 (a) of Title VII,14 regularly and purposefully treated Negroes and Spanishsurnamed Americans less favorably than white persons. The disparity in treatment allegedly involved the refusal to recruit, hire, transfer, or promote minority group members on an equal basis with white jaeople, particularly with respect to line-driving positions. The ultimate factual issues are thus simply whether there was a pattern or practice of such disparate treatment and, if so, whether the differences were “racially premised.” McDonnell Douglas Corp. v. Green, 411 U. S. 792, 805 n. 18.15
*336As the plaintiff, the Government bore the initial burden of making out a prima facie case of discrimination. Albemarle Paper Co. v. Moody, 422 U. S. 405, 425; McDonnell Douglas Corp. v. Green, supra, at 802. And, because it alleged a systemwide pattern or practice of resistance to the full enjoyment of Title VII rights, the Government ultimately had to prove more than the mere occurrence of isolated or “accidental” or sporadic discriminatory acts. It had to establish by a preponderance of the evidence that racial discrimination was the company's standard operating procedure — the regular rather than the unusual practice.16
*337We agree with the District Court and the Court of Appeals that the Government carried its burden of proof. As of March 31, 1971, shortly after the Government filed its complaint alleging systemwide discrimination, the company had 6,472 employees. Of these, 314 (5%) were Negroes and 257 (4%) were Spanish-surnamed Americans. Of the 1,828 line drivers, however, there were only 8 (0.4%) Negroes and 5 (0.3%) Spanish-surnamed persons, and all of the Negroes had been hired after the litigation had commenced. With one exception — a man who worked as a line driver at the Chicago terminal from 1950 to 1959 — the company and its predecessors did not employ a Negro on a regular basis as a line driver until 1969. And, as the Government showed, even in 1971 there were terminals in areas of substantial Negro population where all of the company’s line drivers were white.17 A great majority of the Negroes (83%) and Spanish-surnamed Americans *338(78%) who did work for the company held the lower paying city operations and serviceman jobs,18 whereas only 39% of the nonminority employees held jobs in those categories.
The Government bolstered its statistical evidence with the testimony of individuals who recounted over 40 specific instances of discrimination. Upon the basis of this testimony the District Court found that “[njumerous qualified black and Spanish-surnamed American applicants who sought line driving jobs at the company over the years, either had their requests ignored, were given false or misleading information about requirements, opportunities, and application procedures, or were not considered and hired on the same basis that whites were considered and hired.” Minority employees who wanted to transfer to line-driver jobs met with similar difficulties.19
*339The company’s principal response to this evidence is that statistics can never in and of themselves prove the existence of a pattern or practice of discrimination, or even establish a prima facie case shifting to the employer the burden of rebutting the inference raised by the figures. But, as even our brief summary of the evidence shows, this was not a case in which the Government relied on “statistics alone.” The individuals who testified about their personal experiences with the company brought the cold numbers convincingly to life.
In any event, our cases make it unmistakably clear that “[statistical analyses have served and will continue to serve an important role” in cases in which the existence of discrimination is a disputed issue. Mayor of Philadelphia v. Educational Equality League, 415 U. S. 605, 620. See also McDonnell Douglas Corp. v. Green, 411 U. S., at 805. Cf. Washington v. Davis, 426 U. S. 229, 241-242. We have repeatedly approved the use of statistical proof, where it reached proportions comparable to those in this case, to establish a prima facie case of racial discrimination in jury selection cases, see, e. g., Turner v. Fouche, 396 U. S. 346; Hernandez v. Texas, 347 U. S. 475; Norris v. Alabama, 294 U. S. 587. Statistics are equally competent in proving employment discrimination.20 *340We caution only that statistics are not irrefutable; they come in infinite variety and, like any other kind of evidence, they may be rebutted. In short, their usefulness depends on all of the surrounding facts and circumstances. See, e. g., Hester v. Southern R. Co., 497 F. 2d 1374, 1379-1381 (CA5).
In addition to its general protest against the use of statistics in Title VII cases, the company claims that in this case the statistics revealing racial imbalance are misleading because they fail to take into account the company’s particular busi*341ness situation as of the effective date of Title VII. The company concedes that its line drivers were virtually all white in July 1965, but it claims that thereafter business conditions were such that its work force dropped. Its argument is that low personnel turnover, rather than post-Act discrimination, accounts for more recent statistical disparities. It points to substantial minority hiring in later years, especially after 1971, as showing that any pre-Act patterns of discrimination were broken.
The argument would be a forceful one if this were an employer who, at the time of suit, had done virtually no new hiring since the effective date of Title VII. But it is not. Although the company’s total number of employees apparently dropped somewhat during the late 1960’s, the record shows that many line drivers continued to be hired throughout this period, and that almost all of them were white.21 To be sure, there were improvements in the company’s hiring practices. The Court of Appeals commented that “T. I. M. E.-D. C.’s recent minority hiring progress stands as a laudable good faith effort to eradicate the effects of past discrimination in the area of hiring and initial assignment.” 22 517 F. 2d, at 316. But the District Court and the Court of Appeals found upon substantial evidence that the company had engaged in a course of discrimination that continued well after the effective date of Title VII. The company’s later changes in its hiring and *342promotion policies could be of little comfort to the victims of the earlier post-Act discrimination, and could not erase its previous illegal conduct or its obligation to afford relief to those who suffered because of it. Cf. Albemarle Paper Co. v. Moody, 422 U. S., at 413-423.23
The District Court and the Court of Appeals, on the basis of substantial evidence, held that the Government had proved a prima facie case of systematic and purposeful employment discrimination, continuing well beyond the effective date of Title VII. The company’s attempts to rebut that conclusion were held to be inadequate.24 For the reasons we have sum*343marized, there is no warrant for this Court to disturb the findings of the District Court and the Court of Appeals on this basic issue. See Blau v. Lehman, 368 U. S. 403, 408-409; Faulkner v. Gibbs, 338 U. S. 267, 268; United States v. Dickinson, 331 U. S. 745, 751; United States v. Commercial Credit Co., 286 U. S. 63, 67; United States v. Chemical Foundation, Inc., 272 U. S. 1, 14; Baker v. Schofield, 243 U. S. 114, 118; Towson v. Moore, 173 U. S. 17, 24.
B
The District Court and the Court of Appeals also found that the seniority system contained in the collective-bargaining agreements between the company and the union operated to violate Title VII of the Act.
For purposes of calculating benefits, such as vacations, pensions, and other fringe benefits, an employee’s seniority under this system runs from the date he joins the company, and takes into account his total service in all jobs and bargaining units. For competitive purposes, however, such as determining the order in which employees may bid for particular jobs, are laid off, or are recalled from layoff, it is bargaining-unit seniority that controls. Thus, a line driver’s seniority, *344for purposes of bidding for particular runs25 and protection against layoff, takes into account only the length of time he has been a line driver at a particular terminal.26 The practical effect is that a city driver or serviceman who transfers to a line-driver job must forfeit all the competitive seniority he has accumulated in his previous bargaining unit and start at the bottom of the line drivers’ “board.”
The vice of this arrangement, as found by the District Court and the Court of Appeals, was that it “locked” minority workers into inferior jobs and perpetuated prior discrimination by discouraging transfers to jobs as line drivers. While the disincentive applied to all workers, including whites, it was Negroes and Spanish-surnamed persons who, those courts found, suffered the most because many of them had been denied the equal opportunity to become line drivers when they were initially hired, whereas whites either had not sought or were refused line-driver positions for reasons unrelated to their race or national origin.
The linchpin of the theory embraced by the District Court and the Court of Appeals was that a discriminatee who must forfeit his competitive seniority in order finally to obtain a line-driver job will never be able to “catch up” to the seniority level of his contemporary who was not subject to discrimination.27 Accordingly, this continued, built-in disadvantage to *345the prior discriminatee who transfers to a line-driver job was held to constitute a continuing violation of Title VII, for which both the employer and the union who jointly created and maintain the seniority system were liable.
The union, while acknowledging that the seniority system may in some sense perpetuate the effects of prior discrimination, asserts that the system is immunized from a finding of illegality by reason of § 703 (h) of Title VII, 42 U. S. C. § 2000e-2 (h), which provides in part:
“Notwithstanding any other provision of this subchapter, it shall not be an unlawful employment practice for an employer to apply different standards of compensation, or different terms, conditions,' or privileges of employment pursuant to a bona fide seniority . . . system, . . . provided that such differences are not the result of an intention to discriminate because of race ... or national origin . . . .”
It argues that the seniority system in this case is “bona fide” within the meaning of § 703 (h) when judged in light of its history, intent, application, and all of the circumstances under which it was created and is maintained. More specifically, the union claims that the central purpose of § 703 (h) is to ensure that mere perpetuation of pre-Act discrimination is not unlawful under Title VII. And, whether or not § 703 (h) immunizes the perpetuation of post-Act discrimination, the union claims that the seniority system in this litigation has no such effect. Its position in this Court, as has been its position throughout this litigation, is that the seniority system presents no hurdle to post-Act discrim*346inatees who seek retroactive seniority to the date they would have become line drivers but for the company’s discrimination. Indeed, the union asserts that under its collective-bargaining agreements the union will itself take up the cause of the post-Act victim and attempt, through grievance procedures, to gain for him full “make whole” relief, including appropriate seniority.
The Government responds that a seniority system that perpetuates the effects of prior discrimination — pre-Act or post-Act — can never be “bona fide” under § 703 (h); at a minimum Title VII prohibits those applications of a seniority system that perpetuate the effects on incumbent employees of prior discriminatory job assignments. .
The issues thus joined are open ones in this Court.28 We considered § 703 (h) in Franks v. Bowman Transportation Co., 424 U. S. 747, but there decided only that § 703 (h) does not bar the award of retroactive seniority to job applicants who seek relief from an employer’s post-Act hiring discrimination. We stated that “the thrust of [§ 703 (h)] is directed toward *347defining what is and what is not an illegal discriminatory practice in instances in which the post-Act operation of a seniority system is challenged as perpetuating the effects of discrimination occurring prior to the effective date of the Act.” 424 U. S., at 761. Beyond noting the general purpose of the statute, however, we did not undertake the task of statutory construction required in this litigation.
(1)
Because the company discriminated both before and after the enactment of Title VII, the seniority system is said to have operated to perpetuate the effects of both pre- and post-Act discrimination. Post-Act discriminatees, however, may obtain full “make whole” relief, including retroactive seniority under Franks v. Bowman, supra, without attacking the legality of the seniority system as applied to them. Franks made clear and the union acknowledges that retroactive seniority may be awarded as relief from an employer’s discriminatory hiring and assignment policies even if the seniority system agreement itself makes no provision for such relief.29 424 U. S., at 778-779. Here the Government has proved that the company engaged in a post-Act pattern of discriminatory hiring, assignment, transfer, and promotion policies. Any Negro or Spanish-surnamed American injured by those policies *348may receive all appropriate relief as a direct remedy for this discrimination.30
^
What remains for review is the judgment that the seniority system Unlawfully perpetuated the effects of pre-Act discrimination. We must decide, in short, whether § 703 (h) validates otherwise bona fide seniority systems that afford no constructive seniority to victims discriminated against prior to the effective date of Title VII, and it is to that issue that we now turn.
The primary purpose of Title VII was “to assure equality of employment opportunities and to eliminate those discriminatory practices and devices which have fostered racially stratified job environments to the disadvantage of minority citizens.” McDonnell Douglas Corp. v. Green, 411 U. S., at 800.31 See also Albemarle Paper Co. v. Moody, 422 U. S., at *349417-418; Alexander v. Gardner-Denver Co., 415 U. S. 36, 44; Griggs v. Duke Power Co., 401 U. S., at 429-431. To achieve this purpose, Congress “proscribe [d] not only overt discrimination but also practices that are fair in form, but discriminatory in operation.” Id., at 431. Thus, the Court has repeatedly held that a prima facie Title VII violation may be established by policies or practices that are neutral on their face and in intent but that nonetheless discriminate in effect against a particular group. General Electric Co. v. Gilbert, 429 U. S. 125, 137; Washington v. Davis, 426 U. S., at 246-247; Albemarle Paper Co. v. Moody, supra, at 422, 425; McDonnell Douglas Corp. v. Green, supra, at 802 n. 14; Griggs v. Duke Power Co., supra.
One kind of practice “fair in form, but discriminatory in operation” is that which perpetuates the effects of prior discrimination.32 As the Court held in Griggs: “Under the Act, practices, 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 Ü. S., at 430.
Were it not for § 703 (h), the seniority system in this case would seem to fall under the Griggs rationale. The heart of the system is its allocation of the choicest jobs, the greatest protection against layoffs, and other advantages to those employees who have been line drivers for the longest time. Where, because of the employer’s prior intentional discrim*350ination, the line drivers with the longest tenure are without exception white, the advantages of the seniority system flow disproportionately to them and away from Negro and Spanish-surnamed employees who might by now have enjoyed those advantages had not the employer discriminated before the passage of the Act. This disproportionate distribution of advantages does in a very real sense “operate to 'freeze’ the status quo of prior discriminatory employment practices.” But both the literal terms of § 703 (h) and the legislative history of Title VII demonstrate that Congress considered this very effect of many seniority systems and extended a measure of immunity to them.
Throughout the initial consideration of H. R. 7152, later enacted as the Civil Rights Act of 1964, critics of the bill charged that it would destroy existing seniority rights.33 The consistent response of Title VII’s congressional proponents and of the Justice Department was that seniority rights would not be affected, even where the employer had discriminated prior to the Act.34 An interpretive memorandum placed in the Congressional Record by Senators Clark and Case stated:
“Title VII would have no effect on established seniority rights. Its effect is prospective and not retrospective. Thus, for example, if a business has been discriminating in the past and as a result has an all-white working force, when the title comes into effect the employer’s obligation would be simply to fill future vacancies on a nondiscriminatory basis. He would not be obliged — or in*351deed, permitted — -to fire whites in order to hire Negroes, or to prefer Negroes for future vacancies, or, once Negroes are hired, to give them special seniority rights at the expense of the white workers hired earlier.” 110 Cong. Rec. 7213 (1964) (emphasis added).35
A Justice Department statement concerning Title VII, placed in the Congressional Record by Senator Clark, voiced the same conclusion:
“Title VII would have no effect on seniority rights existing at the time it takes effect. If, for example, a collective bargaining contract provides that in the event of layoffs, those who were hired last must be laid off first, such a provision would not be affected in the least by title VII. This would be true even in the case where owing to discrimination prior to the effective date of the title, white workers had more seniority than Negroes.” Id., at 7207 (emphasis added).36
*352While these statements were made before § 703 (h) was added to Title VII, they are authoritative indicators of that section’s purpose. Section 703 (h) was enacted as part of the Mansfield-Dirksen compromise substitute bill that cleared the way for the passage of Title VII.37 The drafters of the compromise bill stated that one of its principal goals was to resolve the ambiguities in the House-passed version of H. R. 7152. See, e. g., 110 Cong. Rec. 11935-11937 (1964) (remarks of Sen. Dirksen); id., at 12707 (remarks of Sen. Humphrey). As the debates indicate, one of those ambiguities concerned Title VII’s impact on existing collectively bargained seniority rights. It is apparent that § 703 (h) was drafted with an eye toward meeting the earlier criticism on this issue with an explicit provision embodying the understanding and assurances of the Act’s proponents, namely, that Title VII would not outlaw such differences in treatment among employees as flowed from a bona fide seniority system that allowed for full exercise of seniority accumulated before the effective date of the Act. It is inconceivable that § 703 (h), as part of a compromise bill, was intended to vitiate the earlier representations of the Act’s supporters by increasing Title VII’s impact on seniority systems. The statement of Senator Humphrey, noted in Franks, 424 U. S., at 761, confirms that the addition of § 703 (h) “merely clarifies [Title VII’s] present intent and effect.” 110 Cong. Rec. 12723 (1964).
In sum, the unmistakable purpose of § 703 (h) was to make clear that the routine application of a bona fide seniority system would not be unlawful under Title VII. As the legislative history shows, this was the intended result even where the employer’s pre-Act discrimination resulted in whites having greater existing seniority rights than Negroes. Although a seniority system inevitably tends to perpetuate the effects of *353pre-Act discrimination in such cases, the congressional judgment was that Title VII should not outlaw the use of existing seniority lists and thereby destroy or water down the vested seniority rights of employees simply because their employer had engaged in discrimination prior to the passage of the Act.
To be sure, § 703 (h) does not immunize all seniority systems. It refers only to “bona fide” systems, and a proviso requires that any differences in treatment not be “the result of an intention to discriminate because of race ... or national origin . . . .” But our reading of the legislative history compels us to reject the Government’s broad argument that no seniority system that tends to perpetuate pre-Act discrimination can be “bona fide.” To accept the argument would require us to hold that a seniority system becomes illegal simply because it allows the full exercise of the pre-Act seniority rights of employees of a company that discriminated before Title VII was enacted. It would place an affirmative obligation on the parties to the seniority agreement to subordinate those rights in favor of the claims of pre-Act discriminatees without seniority. The consequence would be a perversion of the congressional purpose. We cannot accept the invitation to disembowel § 703 (h) by reading the words “bona fide” as the Government would have us do.38 Accordingly, we hold that an otherwise neutral, legitimate seniority system does not become unlawful under Title VII simply because it may per*354petuate pre-Act discrimination. Congress did not intend to make it illegal for employees with vested seniority rights to continue to exercise those rights, even at the expense of preAct discriminatees.39
That conclusion is inescapable even in a case, such as this one, where the pre-Act discriminatees are incumbent employees who accumulated seniority in other bargaining units. Although there seems to be no explicit reference in the legislative history to pre-Act discriminatees already employed in less desirable jobs, there can be no rational basis for distinguishing their claims from those of persons initially denied any job but hired later with less seniority than they might have had in the absence of pre-Act discrimination.40 We rejected any such *355distinction in Franks, finding that it had “no support anywhere in Title VII or its legislative history,” 424 U. S., at 768. As discussed above, Congress in 1964 made clear that, a seniority system is not unlawful because it honors employees’ existing rights, even where the employer has engaged in pre-Act discriminatory hiring or promotion practices. It would be as contrary to that mandate to forbid the exercise of seniority rights with respect to discriminatees who held inferior jobs as with respect to later hired minority employees who previously were denied any job. If anything, the latter group is the more disadvantaged. As in Franks, “ fit would indeed be surprising if Congress gave a remedy for the one [group] which it denied for the other.’ ” Ibid., quoting Phelps Dodge Corp. v. NLRB, 313 U. S. 177, 187.41
(3)
The seniority system in this litigation is entirely bona fide. It applies equally to all races and ethnic groups. To the extent that it “locks” employees into non-line-driver jobs, it *356does so for all. The city drivers and servicemen who are discouraged from transferring to line-driver jobs are not all Negroes or Spanish-surnamed Americans; to the contrary, the overwhelming majority are white. The placing of line drivers in a separate bargaining unit from other employees is rational, in accord with the industry practice, and consistent with National Labor Relation Board precedents.42 It is conceded that the seniority system did not have its genesis in racial discrimination, and that it was negotiated and has been maintained free from any illegal purpose. In these circumstances, the single fact that the system extends no retroactive seniority to pre-Act discriminatees does not make it unlawful.
Because the seniority system was protected by § 703 (h), the union’s conduct in agreeing to and maintaining the system did not violate Title VII. On remand, the District Court’s injunction against the union must be vacated.43
Ill
Our conclusion that the seniority system does not violate Title VII will necessarily affect the remedy granted to individual employees on remand of this litigation to the District Court. Those employees who suffered only pre-Act discrimination are not entitled to relief, and no person may *357be given retroactive seniority to a date earlier than the effective date of the Act. Several other questions relating to the appropriate measure of individual relief remain, however, for our consideration.
The petitioners argue generally that the trial court did not err in tailoring the remedy to the “degree of injury” suffered by each individual employee, and that the Court of Appeals’ “qualification date” formula sweeps with too broad a brush by granting a remedy to employees who were not shown to be actual victims of unlawful discrimination. Specifically, the petitioners assert that no employee should be entitled to relief until the Government demonstrates that he was an actual victim of the company’s discriminatory practices; that no employee who did not apply for a line-driver job should be granted retroactive competitive seniority; and that no employee should be elevated to a line-driver job ahead of any current line driver on layoff status. We consider each of these contentions separately.
A
The petitioners’ first contention is in substance that the Government’s burden of proof in a pattern-or-practice case must be equivalent to that outlined in McDonnell Douglas v. Green. Since the Government introduced specific evidence of company discrimination against only some 40 employees, they argue that the District Court properly refused to award retroactive seniority to the remainder of the class of minority incumbent employees.
In McDonnell Douglas the Court considered “the order and allocation of proof in a private, non-class action challenging employment discrimination.” 411 U. S., at 800. We held that an individual Title VII complainant must carry the initial burden of proof by establishing a prima facie case of racial discrimination. On the specific facts there involved, we concluded that this burden was met by showing that a *358qualified applicant, who was a member of a racial minority-group, had unsuccessfully sought a job for which there was a vacancy and for which the employer continued thereafter to seek applicants with similar qualifications. This initial showing justified the inference that the minority applicant was denied an employment opportunity for reasons prohibited by Title VII, and therefore shifted the burden to the employer to rebut that inference by offering some legitimate, nondiscriminatory reason for the rejection. Id., at 802.
The company and union seize upon the McDonnell Douglas pattern as the only means of establishing a prima facie case of individual discrimination. Our decision in that case, however, did not purport to create an inflexible formulation. We expressly noted that “[t]he facts necessarily will vary in Title VII cases, and the specification ... of the prima facie proof required from [a plaintiff] is not necessarily applicable in every respect to differing factual situations.” Id., at 802 n. 13. The importance of McDonnell Douglas lies, not in its specification of the discrete elements of proof there required, but in its recognition of the general principle that any Title VII plaintiff must carry the initial burden of offering evidence adequate to create an inference that an employment decision was based on a discriminatory criterion illegal under the Act.44
In Franks v. Bowman Transportation Co., the Court applied *359this principle in the context of a class action. The Franks plaintiffs proved, to the satisfaction of a District Court, that Bowman Transportation Co. “had engaged in a pattern of racial discrimination in various company policies, including the hiring, transfer, and discharge of employees.” 424 U. S., at 751. Despite this showing, the trial court denied seniority relief to certain members of the class of discriminatees because not every individual had shown that he was qualified for the job he sought and that a vacancy had been available. We held that the trial court had erred in placing this burden on the individual plaintiffs. By “demonstrating the existence of a discriminatory hiring pattern and practice” the plaintiffs had made out a prima facie case of discrimination against the individual class members; the burden therefore shifted to the employer “to prove that individuals who reapply were not in fact victims of previous hiring discrimination.” Id., at 772. The Franks case thus illustrates another means by which a Title VII plaintiff’s initial burden of proof can be met. The class there alleged a broad-based policy of employment discrimination ; upon proof of that allegation there were reasonable grounds to infer that individual hiring decisions were made in pursuit of the discriminatory policy and to require the employer to come forth with evidence dispelling that inference.45
*360Although not all class actions will necessarily follow the Franks model, the nature of a pattern-or-practice suit brings it squarely within our holding in Franks. The plaintiff in a pattern-or-practice action is the Government, and its initial burden is to demonstrate that unlawful discrimination has been a regular procedure or policy followed by an employer or group of employers. See supra, at 336, and n. 16. At the initial, “liability” stage of a pattern-or-practice suit the Government is not required to offer evidence that each person for whom it will ultimately seek relief was a victim of the employer’s discriminatory policy. Its burden is to establish a prima facie case that such a policy existed. The burden then shifts to the employer to defeat the prima facie showing of a pattern or practice by demonstrating that the Government’s proof is either inaccurate or insignificant. An employer might show, for example, that the claimed discriminatory pattern is a product of pre-Act hiring rather than unlawful post-Act discrimination, or that during the period it is alleged to have pursued a discriminatory policy it made too few employment decisions to justify the inference that it had engaged in a regular practice of discrimination46
*361If an employer fails to rebut the inference that arises from the Government’s prima facie case, a trial court may then conclude that a violation has occurred and determine the appropriate remedy. Without any further evidence from the Government, a court’s finding of a pattern or practice justifies an award of prospective relief. Such relief might take the form of an injunctive order against continuation of the discriminatory practice, an order that the employer keep records of its future employment decisions and file periodic reports with the court, or any other order “necessary to ensure the full enjoyment of the rights” protected by Title VII.47
When the Government seeks individual relief for the victims of the discriminatory practice, a district court must usually conduct additional proceedings after the liability phase of the trial to determine the scope of individual relief. The petitioners’ contention in this case is that if the Government has not, in the course of proving a pattern or practice, already brought forth specific evidence that each individual was discriminatorily denied an employment opportunity, it must carry that burden at the second, “remedial” stage of trial. That basic contention was rejected in the Franks case. As was true of the particular facts in Franks, and as is typical of Title VII pattern-or-practice suits, the question of individual relief does not arise until it has been proved that the employer has followed an employment policy of unlawful discrimination. The force of that proof does not dissipate at the remedial stage *362of the trial. • The employer cannot, therefore, claim that there is no reason to believe that its individual employment decisions were discriminatorily based; it has already been shown to have maintained a policy of discriminatory decisionmaking.
The proof of the pattern or practice supports an inference that any particular employment decision, during the period in which the discriminatory policy was in force, was made in pursuit of that policy. The Government need only show that an alleged individual discriminatee unsuccessfully applied for a job 48 and therefore was a potential victim of the proved discrimination. As in Franks, the burden then rests on the employer to demonstrate that the individual applicant was denied an employment opportunity for lawful reasons. See 424 U. S., at 773 n. 32.
In Part II-A, supra, we have held that the District Court and Court of Appeals were not in error in finding that the Government had proved a systemwide pattern and practice of racial and ethnic discrimination on the part of the company. On remand, therefore, every post-Act minority group applicant 49 for a line-driver position will be presumptively entitled to relief, subject to a showing by the company that its earlier refusal to place the applicant in a line-driver job was not based on its policy of discrimination.50
B
The Court of Appeals’ “qualification date” formula for relief did not distinguish between incumbent employees who *363had applied for line-driver jobs and those who had not. The appellate court held that where there has been a showing of classwide discriminatory practices coupled with a seniority system that perpetuates the effects of that discrimination, an individual member of the class need not show that he unsuccessfully applied for the position from which the class had been excluded. In support of its award of relief to all non-applicants, the Court suggested that “as a practical matter . . . a member of the affected class may well have concluded that an application for transfer to an all White position such as [line driver] was not worth the candle.” 517 F. 2d, at 320.
The company contends that a grant of retroactive seniority to these nonapplicants is inconsistent with the make-whole purpose of a Title VII remedy and impermissibly will require the company to give preferential treatment to employees solely because of their race. The thrust of the company’s contention is that unless a minority-group employee actually applied for a line-driver job, either for initial hire or for transfer, he has suffered no injury from whatever discrimination might have been involved in the refusal of such jobs to those who actually applied for them.
The Government argues in response that there should be no “immutable rule” that nonapplicants are nonvictims, and contends that a determination whether nonapplicants have suffered from unlawful discrimination will necessarily vary depending on the circumstances of each particular case. The Government further asserts that under the specific facts of this case, the Court of Appeals correctly determined that all qualified nonapplicants were likely victims and were therefore presumptively entitled to relief.
The question whether seniority relief may be awarded to nonapplicants was left open by our decision in Franks, since the class at issue in that case was limited to “identifiable applicants who were denied employment . . . after the effective date ... of Title VII.” 424 U. S., at 750. We now *364decide that an incumbent employee’s failure to apply for a job is not an inexorable bar to an award of retroactive seniority. Individual nonapplicants must be given an opportunity to undertake their difficult task of proving that they should be treated as applicants and therefore are presumptively entitled to relief accordingly.
(1)
Analysis of this problem must begin with the premise that the scope of a district court’s remedial powers under Title VII is determined by the purposes of the Act. Albemarle Paper Co. v. Moody, 422 U. S., at 417. In Griggs v. Duke Power Co., and again in Albemarle, the Court noted that a primary objective of Title VII is prophylactic: to achieve equal employment opportunity and to remove the barriers that have operated to favor white male employees over other employees. 401 U. S., at 429-430; 422 U. S., at 417. The prospect of retroactive relief for victims of discrimination serves this purpose by providing the “ 'spur or catalyst which causes employers and unions to self-examine and to self-evaluate their employment practices and to endeavor to eliminate, so far as possible, the last vestiges’ ” of their discriminatory practices. Id., at 417-418. An equally important purpose of the Act is “to make persons whole for injuries suffered on account of unlawful employment discrimination.” Id., at 418. In determining the specific remedies to be afforded, a district court is “to fashion such relief as the particular circumstances of a case may require to effect restitution.” Franks, 424 U. S., at 764.
Thus, the Court has held that the purpose of Congress in vesting broad equitable powers in Title VII courts was “to make possible the 'fashion [ing] [of] the most complete relief possible,’ ” and that the district courts have “ 'not merely the power but the duty to render a decree which will so far as possible eliminate the discriminatory effects of the past as well as bar like discrimination in the future.’ ” Albemarle, *365 supra, at 421, 418. More specifically, in Franks we decided that a court must ordinarily award a seniority remedy unless there exist reasons for denying relief “ 'which, if applied generally, would not frustrate the central statutory purposes of eradicating discrimination . . . and making persons whole for injuries suffered.’ ” 424 U. S., at 771, quoting Albemarle, supra, at 421.
Measured against these standards, the company’s assertion that a person who has not actually applied for a job can never be awarded seniority relief cannot prevail. The effects of and the injuries suffered from discriminatory employment practices are not always confined to those who were expressly denied a requested employment opportunity. A consistently enforced discriminatory policy can surely deter job applications from those who are aware of it and are unwilling to subject themselves to the humiliation of explicit and certain rejection.
If an employer should announce his policy of discrimination by a sign reading “Whites Only” on the hiring-office door, his victims would not be limited to the few who ignored the sign and subjected themselves to personal rebuffs. The same message can be communicated to potential applicants more subtly but just as clearly by an employer’s actual practices— by his consistent discriminatory treatment of actual applicants, by the manner in which he 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 work force from which he has discriminatorily excluded members of minority groups.51 When a person’s *366desire for a job is not translated into a formal application solely because of his unwillingness to engage in a futile gesture he is as much a victim of discrimination as is he who goes through the motions of submitting an application.
In cases decided under the National Labor Relations Act, the model for Title VIPs remedial provisions, Albemarle, supra, at 419; Franks, supra, at 769, the National Labor Relations Board, and the courts in enforcing its orders, have recognized that the failure to submit a futile application does not bar an award of relief to a person claiming that he was denied employment because of union affiliation or activity. In NLRB v. Nevada Consolidated Copper Corp., 316 U. S. 105, this Court enforced an order of the Board directing an employer to hire, with retroactive benefits, former employees who had not applied for newly available jobs because of the employer’s well-known policy of refusing to hire union members. See In re Nevada Consolidated Copper Corp., 26 N. L. R. B. 1182, 1208, 1231. Similarly, when an application would have been no more than a vain gesture in light of employer discrimination, the Courts of Appeals have enforced Board orders reinstating striking workers despite the failure of individual strikers to apply for reinstatement when the strike ended. E. g., NLRB v. Park Edge Sheridan Meats, Inc., 323 P. 2d 956 (CA2); NLRB v. Valley Die Cast Corp., 303 F. 2d 64 (CA6); Eagle-Picher Mining & Smelting Co. v. NLRB, 119 F. 2d 903 (CA8). See also Piasecki Aircraft Corp. v. NLRB, 280 F. 2d 575 (CA3); NLRB v. Anchor Rome Mills, *367228 F. 2d 775 (CA5); NLRB v. Lummus Co., 210 F. 2d 377 (CA5). Consistent-with the NLRA model, several Courts of Appeals have held in Title VII cases that a nonapplicant can be a victim of unlawful discrimination entitled to make-whole relief when an application would have been a useless act serving only to confirm a discriminatee’s knowledge that the job he wanted was unavailable to him. Acha v. Beame, 531 F. 2d 648, 656 (CA2); Hairston v. McLean Trucking Co., 520 F. 2d 226, 231-233 (CA4); Bing v. Roadway Express, Inc., 485 F. 2d 441, 451 (CA5); United States v. N. L. Industries, Inc., 479 F. 2d 354, 369 (CA8).
The denial of Title VII relief on the ground that the claimant had not formally applied for the job could exclude from the Act’s coverage the victims of the most entrenched forms of discrimination. Victims of gross and pervasive discrimination could be denied relief precisely because the unlawful practices had been so successful as totally to deter job applications from members of minority groups. A per se prohibition of relief to nonapplicants could thus put beyond the reach of equity the most invidious effects of employment discrimination — those that extend to the very hope of self-realization. Such a per se limitation on the equitable powers granted to courts by Title VII would be manifestly inconsistent with the “historic purpose of equity to 'secur[e] complete justice’ ” and with the duty of courts in Title VII cases “ 'to render a decree which will so far as possible eliminate the discriminatory effects of the past.’ ” Albemarle Paper Co. v. Moody, 422 U. S., at 418.
(2)
To conclude that a person’s failure to submit an application ' for a job does not inevitably and forever foreclose his entitlement to seniority relief under Title VII is a far cry, however, from holding that nonapplicants are always entitled to such relief. A nonapplicant must show that he was a potential victim of unlawful discrimination. Because he is necessarily *368claiming that he was deterred from applying for the job by the employer’s discriminatory practices, his is the not always easy burden of proving that he would have applied for the job had it not been for those practices. Cf. Mt. Healthy City Board of Education v. Doyle, 429 U. S. 274. When this burden is met, the nonapplicant is in a position analogous to that of an applicant and is entitled to the presumption discussed in Part III-A, supra.
The Government contends that the evidence it presented in this case at the liability stage of the trial identified all non-applicants as victims of unlawful discrimination “with a fair degree of specificity,” and that the Court of Appeals’ determination that qualified nonapplicants are presumptively entitled to an award of seniority should accordingly be affirmed. In support of this contention the Government cites its proof of an extended pattern and practice of discrimination as evidence that an application from a minority employee for a line-driver job would have been a vain and useless act. It further argues that since the class of nonapplicant discriminatees is limited to incumbent employees, it is likely that every class member was aware of the futility of seeking a line-driver job and was therefore deterred from filing both an initial and a followup application.52
*369We cannot agree. While the scope and duration of the company’s discriminatory policy can leave little doubt that the futility of seeking line-driver jobs was communicated to the company’s minority employees, that in itself is insufficient. The known prospect of discriminatory rejection shows only that employees who wanted line-driving jobs may have been deterred from applying for them. It does not show which of the nonapplicants actually wanted such jobs, or which possessed the requisite qualifications.53 There are differences between city- and line-driving jobs.54 for example, but the desirability of the latter is not so self-evident as to warrant a conclusion that all employees would prefer to be line drivers if given a free choice.55 Indeed, a substantial number of white *370city drivers who were not subjected to the company's discriminatory practices were apparently content to retain their city jobs.56
In order to fill this evidentiary gap, the Government argues that a nonapplicant's current willingness to transfer into a line-driver position confirms his past desire for the job. An employee's response to the court-ordered notice of his entitlement to relief57 demonstrates, according to this argument, that *371the employee would have sought a line-driver job when he first became qualified to fill one, but for his knowledge of the company’s discriminatory policy.
This assumption falls short of satisfying the appropriate burden of proof. An employee who transfers into a line-driver unit is normally placed at the bottom of the seniority “board.” He is thus in jeopardy of being laid off and must, at best, suffer through an initial period of bidding on only the least desirable runs. See supra, at 343-344, and n. 25. Non-applicants who chose to accept the appellate court’s post hoc invitation, however, would enter the line-driving unit with retroactive seniority dating from the time they were first qualified. A willingness to accept the job security and bidding power afforded by retroactive seniority says little about what choice an employee would have made had he previously been given the opportunity freely to choose a starting line-driver job. While it may be true that many of the nonapplicant employees desired and would have applied for line-driver jobs but for their knowledge of the company’s policy of discrimination, the Government must carry its burden of proof, with respect to each specific individual, at the remedial hearings to be conducted by the District Court on remand.58
C
The task remaining for the District Court on remand will not be a simple one. Initially, the court will have to make a substantial number of individual determinations in deciding which of the minority employees were actual victims *372of the company’s discriminatory practices. After the victims have been identified, the court must, as nearly as possible, “ 'recreate the conditions and relationships that would have been had there been no’ ” unlawful discrimination. Franks, 424 U. S., at 769. This process of recreating the past will necessarily involve a degree of approximation and imprecision. Because the class of victims may include some who did not apply for line-driver jobs as well as those who did, and because more than one minority employee may have been denied each line-driver vacancy, the court will be required to balance the equities of each minority employee’s situation in allocating the limited number of vacancies that were discriminatorily refused to class members.
Moreover, after the victims have been identified and their rightful place determined, the District Court will again be faced with the delicate task of adjusting the remedial interests of discriminatees and the legitimate expectations of other employees innocent of any wrongdoing. In the prejudgment consent decree, see n. 4, supra, the company and the Government agreed that minority employees would assume line-driver positions that had been discriminatorily denied to them by exercising a first-priority right to job vacancies at the company’s terminals. The decree did not determine what constituted a vacancy, but in its final order the trial court defined “vacancy” to exclude any position that became available while there were laid-off employees awaiting an opportunity to return to work. Employees on layoff were given a preference to fill whatever openings might occur at their terminals during a three-year period after they were laid off.59 *373The Court of Appeals rejected the preference and held that all but “purely temporary” vacancies were to be filled according to an employee’s seniority, whether as a member of the class *374discriminated against or as an incumbent line driver on layoff. 517 F. 2d, at 322-323.
As their final contention concerning the remedy, the company and the union argue that the trial court correctly made the adjustment between the competing interests of discriminatees and other employees by granting a preference to laid-off employees, and that the Court of Appeals erred in disturbing it. The petitioners therefore urge the reinstatement of that part of the trial court’s final order pertaining to the rate at which victims will assume their rightful places in the line-driver hierarchy.60
Although not directly controlled by the Act,61 the extent to *375which the legitimate expectations of nonvictim employees should determine when victims are restored to their rightful place is limited by basic principles of equity. .In devising and implementing remedies under Title VII, no less than in formulating any equitable decree, a court must draw on the “qualities of mercy and practicality [that] have made equity the instrument for nice adjustment and reconciliation between the public interest and private needs as well as between competing private claims.” Hecht Co. v. Bowles, 321 U. S. 321, 329-330. Cf. Phelps Dodge Corp. v. NLRB, 313 U. S., at 195-196, modifying 113 F. 2d 202 (CA2); 19 N. L. R. B. 547, 600; Franks, 424 U. S., at 798-799 (Powell, J., concurring in part and dissenting in part). Especially when immediate implementation of an equitable remedy threatens to impinge upon the expectations of innocent parties, the courts must “look to the practical realities and necessities inescapably involved in reconciling competing interests,” in order to determine the “special blend of what is necessary, what is fair, and what is workable.” Lemon v. Kurtzman, 411 U. S. 192, 200-201 (opinion of Burger, C. J.).
Because of the limited facts now in the record, we decline to strike the balance in this Court. The District Court did not explain why it subordinated the interests of class members to the contractual recall expectations of other employees on layoff. When it made that determination, however, it was considering a class of more than 400 minority employees, all of whom had been granted some preference in filling line-driver vacancies. The overwhelming majority of these were in the District Court's subclass three, composed of those employees with respect to whom neither the Government nor the company had presented any specific evidence on the question of unlawful discrimination. Thus, when the court considered the problem of what constituted a line-driver “vacancy” *376to be offered to class members, it may have been influenced by the relatively small number of proved victims and the large number of minority employees about whom it had no information. On the other hand, the Court of Appeals redefined “vacancy” in the context of what it believed to be a class of more than 400 employees who had actually suffered from discrimination at the behest of both the company and the union, and its determination may well have been influenced by that understanding. For the reasons discussed in this opinion, neither court's concept was completely valid.
After the evidentiary hearings to be conducted on remand, both the size and the composition of the class of minority employees entitled to relief may be altered substantially. Until those hearings have been conducted and both the number of identifiable victims and the consequent extent of necessary relief have been determined, it is not possible to evaluate abstract claims concerning the equitable balance that should be struck between the statutory rights of victims and the contractual rights of nonvictim employees. That determination is best left, in the first instance, to the sound equitable discretion of the trial court.62 See Franks v. Bowman Transportation Co., supra, at 779; Albemarle Paper Co. v. Moody, 422 U. S., at 416. We observe only that when the court exercises its discretion in dealing with the problem of laid-off employees in light of the facts developed at the hearings on remand, it should clearly state its reasons so that meaningful review may be had on appeal. See Franks, supra, at 774; Albemarle Paper Co. v. Moody, supra, at 421 n. 14.
For all the reasons we have discussed, the judgment of the Court of Appeals is vacated, and the cases are remanded to the *377District Court for further proceedings consistent with this opinion.
It is so ordered.
with whom Mr. Justice Brennan joins, concurring in part and dissenting in part.
I agree with the Court that the United States proved that petitioner T. I. M. E.-D. C. was guilty of a pattern or practice of discriminating against blacks and Spanish-surnamed Americans in hiring line drivers. I also agree that incumbent minority-group employees who show that they applied for a line-driving job or that they would have applied but for the company’s unlawful acts are presumptively entitled to the full measure of relief set forth in our decision last Term in Franks v. Bowman Transportation Co., 424 U. S. 747 (1976).1 But I do not agree that Title VII permits petitioners to treat Negro and Spanish-surnamed line drivers differently from other drivers who were hired by the company at the same time simply because the former drivers were prevented by the company from acquiring seniority over the road. I therefore dissent *378from that aspect of the Court’s holding, and from the limitations on the scope of the remedy that follow from it.
As the Court quite properly acknowledges, ante, at 349-350, the seniority provision at issue here clearly would violate Title VII absent § 703 (h), 42 U. S. C. § 2000e-2 (h), which exempts at least some seniority systems from the reach of the Act. Title VII prohibits an employer from “classify [ing] his employees ... 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) (2) (1970 ed., Supp. V). “Under the Act, practices, 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.” Griggs v. Duke Power Co., 401 U. S. 424, 430 (1971) (emphasis added). Petitioners’ seniority system does precisely that: it awards the choicest jobs and other benefits to those possessing a credential — seniority—which, due to past discrimination, blacks and Spanish-surnamed employees were prevented from acquiring. Consequently, “[e]very time a Negro worker hired under the old segregated system bids against a white worker in his job slot, the old racial classification reasserts itself, and the Negro suffers anew for his employer’s previous bias.” Local 189, United Papermakers & Paperworkers v. United States, 416 F. 2d 980, 988 (CA5 1969) (Wisdom, J.), cert. denied, 397 U. S. 919 (1970).
As the Court also concedes, with a touch of understatement, “the view that § 703 (h) does not immunize seniority systems that perpetuate the effects of prior discrimination has much support.” Ante, at 346 n. 28. Without a single dissent, six Courts of Appeals have so held in over 30 cases,2 and two *379other Courts of Appeals have indicated their agreement, also without dissent.3 In an unbroken line of cases, the Equal Employment Opportunity Commission has reached the same *380conclusion.4 And the overwhelming weight of scholarly-opinion is in accord.5 Yet for the second time this Term, see General Electric Co. v. Gilbert, 429 U. S. 125 (1976), a majority of this Court overturns the unanimous conclusion of the Courts of Appeals and the EEOC concerning the scope of Title VII. Once again, I respectfully disagree.
*381I
Initially, it is important to bear in mind that Title VII is a remedial statute designed to eradicate certain invidious employment practices. The evils against which it is aimed are defined broadly: “to fail ... to hire or to discharge ... or otherwise to discriminate . . . with respect to . . . compensation, terms, conditions, or privileges of employment,” and “to limit, segregate, or classify . . . in any way which would deprive or tend to deprive any individual of employment opportunities or otherwise adversely affect his status.” 42 U. S. C. § 2000e-2 (a) (1970 ed., Supp. V) (emphasis added). Section 703 (h) carves out an exemption from these broad prohibitions. Accordingly, under longstanding principles of statutory construction, the Act should “be given a liberal interpretation . . . [and] exemptions from its sweep should be narrowed and limited to effect the remedy intended.” Piedmont & Northern R. Co. v. ICC, 286 U. S. 299, 311-312 (1932); see also Spokane & Inland R. Co. v. United States, 241 U. S. 344, 350 (1916); United States v. Dickson, 15 Pet. 141, 165 (1841) (Story, J.). Unless a seniority system that perpetuates discrimination falls “plainly and unmistakably within [the] terms and spirit" of § 703 (h), A. H. Phillips, Inc. v. Walling, 324 U. S. 490, 493 (1945), the system should be deemed unprotected. I submit that whatever else may be true of the section, its applicability to systems that perpetuate past discrimination is not “plainly and unmistakably” clear.
The language of § 703 (h) provides anything but clear support for the Court’s holding. That section provides, in pertinent part:
“[I]t shall not be an unlawful employment practice for an employer to apply different standards of compensation, or different terms, conditions or privileges of employment pursuant to a bona fide seniority . . . system . . . provided that such differences are not the result of an intention to *382 discriminate because of race, color, religion, sex, or national origin . . . (Emphasis added.)
In this case, however, the different “privileges of employment” ■for Negroes and Spanish-surnamed Americans, on the one hand, and for all others, on the other hand, produced by petitioners’ seniority system are precisely the result of prior, intentional discrimination in assigning jobs; but for that discrimination, Negroes and Spanish-surnamed Americans would not be disadvantaged by the system. Thus, if the proviso is read literally, the instant case falls squarely within it, thereby rendering § 703 (h) inapplicable. To avoid this result the Court is compelled to reconstruct the proviso to read: provided that such a seniority system “did not have its genesis in racial discrimination, and that it was negotiated and has been maintained free from any illegal purpose.” Ante, at 356.
There are no explicit statements in the legislative history of Title VII that warrant this radical reconstruction of the proviso. The three documents placed in the Congressional Record by Senator Clark concerning seniority all were written many weeks before the Mansfield-Dirksen amendment containing § 703 (h) was introduced. Accordingly, they do not specifically discuss the meaning of the proviso.6 More im*383portantly, none of the documents addresses the general problem of seniority systems that perpetuate discrimination. Not surprisingly, Congress simply did not think of such subtleties in enacting a comprehensive, pathbreaking Civil Rights Act.7 To my mind, this is dispositive. Absent unambiguous statutory language or an authoritative statement in the legislative history legalizing seniority systems that continue past wrongs, I do not see how it can be said that the § 703 (h) exemption “plainly and unmistakably” applies.
II
Even if I were to agree that this case properly can be decided on the basis of inferences as to Congress’ intent, I still could not accept the Court’s holding. In my. view, the legislative history of the 1964 Civil Rights Act does not support the conclusion that Congress intended to legalize seniority systems that perpetuate discrimination, and administrative and legislative developments since 1964 positively refute that conclusion.
A
The Court’s decision to uphold seniority systems that perpetuate post-Act discrimination — that is, seniority systems that treat Negroes and Spanish-sumamed Americans who become line drivers as new employees even though, after the effective date of Title VII, these persons were discriminatorily assigned to city-driver jobs where they accumulated seniority — -is explained in a single footnote. Ante, at 348 n. 30. That footnote relies almost entirely on United Air Lines, Inc. *384v. Evans, post, p. 553. But like the instant decision, Evans is devoid of any analysis of the legislative history of § 703 (h); it simply asserts its conclusion in a single paragraph. For the Court to- base its decision here on the strength of Evans is sheer bootstrapping.
Had the Court objectively examined the legislative history, it would have been compelled to reach the opposite conclusion. As we stated just last Term, “it is apparent that the thrust of [§ 703 (h)] is directed toward defining what is and what is not an illegal discriminatory practice in instances in which the post-Act operation of a seniority system is challenged as perpetuating the effects of discrimination occurring prior to the effective date of the Act.”8 Franks v. Bowman Transportation Co., 424 U. S., at 761 (emphasis added). Congress was concerned with seniority expectations that had developed prior to the enactment of Title VII, not with expectations arising thereafter to the extent that those expectations were dependent on whites benefiting from unlawful discrimination. Thus, the paragraph of the Clark-Case Interpretive Memorandum dealing with seniority systems begins:
“Title VII would have no effect -on established seniority rights. Its effect is prospective and not retrospective.” 110 Cong. Rec. 7213 (1964) (emphasis added).
Similarly, the Justice Department memorandum that Senator Clark introduced explains:
“Title VII would have no effect on seniority rights existing at the time it takes effect. If, for example a collective bargaining contract provides that in the event of layoffs, those who were hired last must be laid off first, such a provision would not be affected ... by title VII. This *385would be true even in the case where owing to discrimination prior to the effective date of the title, white workers had more seniority than Negroes .... Any differences in treatment based on established seniority rights would not be based on race and would not be forbidden by the title.” Id., at 7207 (emphasis added).
Finally, Senator Clark’s prepared answers to questions propounded by Senator Dirksen stated:
“Question. If an employer is directed to abolish his employment list because of discrimination what happens to seniority?
“Answer. The bill is not retroactive, and it will not require an employer to change existing seniority lists.” Id., at 7217 (emphasis added).
For the Court to ignore this history while reaching a conclusion contrary to it is little short of remarkable.
B
The legislative history of § 703 (h) admittedly affords somewhat stronger support for the Court’s conclusion with respect to seniority systems that perpetuate pre-Act discrimination— that is, seniority systems that treat Negroes and Spanishsurnamed Americans who become line drivers as new employees even though these persons were discriminatorily assigned to city-driver jobs where they accumulated seniority before the effective date of Title VII. In enacting § 703 (h), Congress intended to extend at least some protection to seniority expectations that had developed prior to the effective date of the Act. But the legislative history is very clear that the only threat to these expectations that Congress was seeking to avert was nonremedial, fictional seniority. Congress did not want minority group members who were hired after the effective date of the Act to be given superseniority simply because they were members of minority groups, nor did it want the use of seniority to be invalidated whenever it had a disparate *386impact on newly hired minority employees. These are the evils — and the only evils — that the opponents of Title VII raised9 and that the Clark-Case Interpretive Memorandum addressed.10 As the Court acknowledges, “there seems to be no explicit reference in the legislative history to pre-Act discriminatees already employed in less desirable jobs.” Ante, at 354.
Our task, then, assuming still that the case properly can be decided on the basis of imputed legislative intent, is “to put to ourselves the question, which choice is it the more likely that Congress would have made,” Burnet v. Guggenheim, 288 *387U. S. 280, 285 (1933) (Cardozo, J.), had it focused on the problem: would it have validated or invalidated seniority systems that perpetuate pre-Act discrimination? To answer that question, the devastating impact of today’s holding validating such systems must be fully understood. Prior to 1965 blacks and Spanish-surnamed Americans who were able to find employment were assigned the lowest paid, most menial jobs in many industries throughout the Nation but especially in the South. In many factories, blacks were hired as laborers while whites were trained and given skilled positions; 11 in the transportation industry blacks could only become porters; 12 and in steel plants blacks were assigned to the coke ovens and blasting furnaces, "the hotter and dirtier” places of employment.13 The Court holds, in essence, that while after 1965 these incumbent employees are entitled to an equal opportunity to advance to more desirable jobs, to take advantage of that opportunity they must pay a price: they must surrender the seniority they have accumulated in their old jobs. Por many, the price will be too high, and they will be locked into their previous positions.14 Even those willing to pay the price will *388have to reconcile themselves to being forever behind subsequently hired whites who were not discriminatorily assigned. Thus equal opportunity will remain a distant dream for all incumbent employees.
I am aware of nothing in the legislative history of the 1964 Civil Rights Act to suggest that if Congress had focused on this fact it nonetheless would have decided to write off an entire generation of minority-group employees. Nor can I believe that the Congress that enacted Title VII would have agreed to postpone for one generation the achievement of economic equality. The backers of that Title viewed economic equality as both a practical necessity and a moral imperative.15 They were well aware of the corrosive impact employment discrimination has on its victims, and on society generally.16 They sought, therefore, “to eliminate those discriminatory practices and devices which have fostered racially stratified job environments to the disadvantage of minority citizens”; McDonnell Douglas Corp. v. Green, 411 U. S. 792, 800 (1973); see also Griggs v. Duke Power Co., 401 U. S., at 429-431; Alexander v. Gardner-Denver Co., 415 U. S. 36, 44 (1974); and “to make persons whole for injuries suffered on account of unlawful employment discrimination,” Albemarle Paper Co. v. Moody, 422 U. S. 405, 418 (1975). In *389short, Congress wanted to enable black workers to assume their rightful place in society.
It is, of course, true that Congress was not willing to invalidate seniority systems on a wholesale basis in pursuit of that goal.17 But the United States, as the plaintiff suing on behalf of the incumbent minority group employees here, does not seek to overturn petitioners’ seniority system. It seeks only to have the “time actually worked in [minority group] jobs [recognized] as the equal of [the majority group’s] time,” Local 189, United Papermakers & Paperworkers v. United States, 416 F. 2d, at 995, within the existing seniority system. Admittedly, such recognition would impinge on the seniority expectations white employees had developed prior to the effective date of the Act. But in enacting Title VII, Congress manifested a willingness to do precisely that. For example, the Clark-Case Interpretive Memorandum, see n. 6, supra, makes clear that Title VII prohibits unions and employers from using discriminatory waiting lists, developed prior to the effective date of the Title, in making selections for jobs or training programs after that date. 110 Cong. Rec. 7213 (1964). Such a prohibition necessarily would disrupt the expectations of those on the lists. More generally, the very fact that Congress made Title VII effective shortly after its enactment demonstrates that expectations developed prior to passage of the Act were not considered sacrosanct, since Title VII’s general ban on employment discrimination inevitably interfered with the pre-existing expectations of whites who anticipated benefiting from continued discrimination. Thus I am in complete agreement with Judge Butzner’s conclusion *390in his seminal decision in Quarles v. Philip Morris, Inc., 279 F. Supp. 505, 516 (ED Va. 1968): “It is . . . apparent that Congress did not intend to freeze an entire generation of Negro employees into discriminatory patterns that existed before the Act.” 18
C
If the legislative history .of § 703 (h) leaves any doubt concerning the section’s applicability to seniority systems that perpetuate either pre- or post-Act discrimination, that doubt is entirely dispelled by two subsequent developments. The Court all but ignores both developments; I submit they are critical.
First, in more than a score of decisions beginning at least as early as 1969, the Equal Employment Opportunity Commission has consistently held that seniority systems that perpetuate prior discrimination are unlawful.19 While the Court may have retreated, see General Electric Co. v. Gilbert, 429 U. S. 125, 141-142 (1976), from its prior view that the interpretations of the EEOC are “ ‘entitled to great deference,’ ” Albemarle Paper Co. v. Moody, supra, at 431, quoting Griggs *391v. Duke Power Co., supra, at 434, I have not. Before I would sweep aside the EEOC’s consistent interpretation of the statute it administers, I would require “ 'compelling indications that it is wrong.’ ” Espinoza v. Farah Mfg. Co., 414 U. S. 86, 94-95 (1973), quoting Red Lion Broadcasting Co. v. FCC, 395 U. S. 367, 381 (1969). I find no such indications in the Court’s opinion.
Second, in 1972 Congress enacted the Equal Employment’ Opportunity Act of 1972, Pub. L. 92-261, 86 Stat. 103, amending Title VII. In so doing, Congress made very clear that it approved of the lower court decisions invalidating seniority systems that perpetuate discrimination. That Congress was aware of such cases is evident from the Senate and House Committee Reports which cite the two leading decisions, as well as several prominent law review articles. S. Rep. No. 92-415, p. 5 n. 1 (1971); H. R. Rep. No. 92-238, p. 8 n. 2 (1971). Although Congress took action with respect to other lower court opinions with which it was dissatisfied,20 it made no attempt to overrule the seniority cases. To the contrary, both the Senate and House Reports expressed approval of the ''perpetuation principle” as applied to seniority systems21 and *392invoked the principle to justify the Committees’ recommendations to extend Title VII’s coverage to state and local government employees,22 and to expand the powers of the EEOC.23 Moreover, the Section-by-Section Analysis of the *393Conference Committee bill, which was prepared and placed in the Congressional Record by the floor managers of the bill, stated in “language that could hardly be more explicit,” Franks v. Bowman Transportation Co., 424 U. S., at 765 n. 21, that, “in any areas where a specific contrary intention is not indicated, it was assumed that the present case law . . . would continue to govern the applicability and construction of Title VII.” 118 Cong. Rec. 7166, 7564 (1972). And perhaps most important, in explaining the section of the 1972 Act that empowers the EEOC “to prevent any person from engaging in any unlawful employment practice as set forth in section 2000e-2 or 2000e-3,” 42 U. S. C. § 2000e-5 (a) (1970 ed., Supp. V), the Section-by-Section Analysis declared:
“The unlawful employment practices encompassed by sections 703 and 704 which were enumerated in 1964 by the original Act, and as defined and expanded by the courts, remain in effect.” 118 Cong. Rec. 7167, 7564 (1972) (emphasis added).24
We have repeatedly held: “When several acts of Congress are passed touching the same subject matter, subsequent legislation may be considered to assist in the interpretation of prior legislation upon the same subject.” Tiger v. Western Investment Co., 221 U. S. 286, 309 (1911); see NLRB v. Bell Aerospace Co., 416 U. S. 267, 275 (1974) (subsequent legisla*394tion entitled to “significant weight”); Red Lion Broadcasting Co. v. FCC, 395 U. S., at 380; United States v. Stafoff, 260 U. S. 477, 480 (1923) (Holmes, J.); New York & Norfolk R. Co. v. Peninsula Produce Exchange, 240 U. S. 34, 39 (1916) (Hughes, J.); United States v. Weeks, 5 Cranch 1, 8 (1809). Earlier this Term, we implicitly followed this canon in using a statute passed in 1976 to conclude that the Administrative Procedure Act, 5 U. S. C. §§ 701-706, enacted in 1946, was not intended as an independent grant of jurisdiction to the federal courts. Califano v. Sanders, 430 U. S. 99 (1977). The canon is particularly applicable here for two reasons. First, because there is no explicit legislative history discussing seniority systems that perpetuate discrimination, we are required to “'[seize] every thing from which aid can be derived Brown v. GSA, 425 U. S. 820, 825 (1976), quoting, United States v. Fisher, 2 Cranch 358, 386 (1805), if we are to reconstruct congressional intent. Second, because petitioners' seniority system was readopted in collective-bargaining agreements signed after the 1972 Act took effect, any retroactivity problems that ordinarily inhere in using a later Act to interpret an earlier one are not present here. Cf. Stockdale v. Insurance Cos., 20 Wall. 323, 331-332 (1874). Thus, the Court's bald assertion that the intent of the Congress that enacted the 1972 Act is “entitled to little if any weight,” ante, at 354 n. 39, in construing § 703 (h) is contrary to both principle and precedent.
Only last Term, we concluded that the legislative materials reviewed above “completely [answer] the argument that Congress somehow intended seniority relief to be less available” than backpay as a remedy for discrimination. Franks v. Bowman Transportation Co., supra, at 765 n. 21. If anything, the materials provide an even more complete answer to the argument that Congress somehow intended to immunize seniority systems that perpetuate past discrimination. To the extent that today’s decision grants immunity to such systems, I respectfully dissent.
6.2 Statistics Explainer 6.2 Statistics Explainer
Teamsters and the next case, Hazelwood, next raise the question whether a disparity alone is proof that the employer intentionally excluded minority candidates from being hired. The answer is generally no -- it’s possible to imagine an employer using genuinely neutral hiring criteria and methods, and ending up with a disparity, just by random chance – especially if the employer has only a small number of employees, and/or the number of minorities that were available for hiring is also small. But if the employer has a large number of employees overall and if we know that abundant minority candidates were available to be hired, might it be possible to infer that the employer must have had discriminatory motives when that employer still ends up with almost no racial minorities among its employees?
In such cases, plaintiffs might be able to use statistics to demonstrate to the court’s satisfaction that the disparity is so great that it could not have been caused by random chance, and must have therefore been intended by the employer. Since many of us do not have a background in statistics, I’ve created this explainer to help you navigate the the statistical analysis in this and the next case.
Standard deviation, a/k/a sigma (σ) is a number that you can calculate that can help you determine how much a disparity within a given population differs from a population of normal distribution.
Normal distribution is a way to visualize the frequency of particular experimental outcomes. As an example, let’s consider an experiment of flipping coins. Imagine that you and 99 other people each flipped a coin 100 times, and each one of you counts the number of time the coins landed on heads (we will call this number “H”). Now everyone else reports their H to you and you have to plot it on a graph. Your x-axis (horizontal line) is for plotting all of the potential Hs, and the y-axis (the vertical line) is the frequency of each H (the number of people who got that same result). What shape do you expect this plot to take? Which values of H are going to be the most common within your group of 100? Which will be the most rare?
The plot you’ve created is called a normal distribution because it will look something like the plot below. If it doesn’t, you simply need to increase the number of flips and/or the number of flippers, and you will get this plot, statistically guaranteed!
Of course, you will expect that because the chance of heads is 50% on any given flip, the most likely outcome is that 50 of 100 flips will be heads (H = 50). 50 is therefore our “mean.” (On the graph below, it is the line in the middle that is labeled 0, but you can imagine it says 50 instead). We would expect H values that are close to 50 to also occur with relatively high frequency, while H values that are far away from 50 to occur less frequently. We would expect a lot of 47s, 52s, 49s, and 53s, but it would be exceedingly rare for anyone to report an H of 100, or an H of 0, right?

Now, consider the possibility that someone in your group of 100 flippers reports to you a really low or a really high H. One that is so far away (in either direction) from 50 that it makes you suspicious that something is wrong with the coin! Is there a number like that that you can imagine the odds of getting it randomly are just too low to be random chance? Ask yourself what that number might be. Then, we will calculate what that is, using two pieces of information. One piece of information that is generally known and applicable across all standard deviation problems, and another piece of information that we have to calculate using the equation for standard deviation.
First, the piece of information is known is the reference points for standard deviation on a normal distribution. They are shown on the diagram above: 95 percent of outcomes (in our experiment, values of H – the number of heads in 100 coin flips) are within 2 standard deviations from the mean. The diagram also tells us that when you get out to three standard deviations from the mean, more than 99 percent of outcomes are within those boundaries. Therefore, the value of H you are worried about is more than 3 standard deviations from the mean, you are right to be suspicious because the odds of that happening randomly are less than 1%.
That brings us to the second piece of information, the value of one standard deviation. This we have to calculate for our situation, using the formula for standard deviation that the book gives on page 220:
N in the case of our coin flip calculation is 100, the total number of coin flips.
P is the percentage of flips that we would expect to be heads, or 0.50
Q is the percentage of flips that we would expect to not be heads, which is again 0.50.
Multiply N * P * Q, or 100 x 0.50 x. 0.50. You get 25. Now take the square root of 25, which is 5.
Thus, one standard deviation = 5. Two standard deviations = 10. Three standard deviations = 15, etc.
In other words, if we get an H that is more than 3 standard deviations from the mean, we are highly suspicious of this coin, since it is really unlikely (less than 1% likely) that this was due to random chance. And what are those cutoff numbers? 50 ± 15. I.e., 35 or lower, and 65 and higher.
Now. What does this have to do with employment discrimination?
Instead of a coin flip experiment, we have a workplace. Instead of a coin flipper, we have an employer who makes hiring decisions, like the employer in the Teamsters case. Every decision to hire someone is has two possible outcomes, like a coin flip. The person hired will either be black, or not black (for example) (I know it’s awkward to categorize people as either black or not-black, and I am only doing this for the purpose of applying the statistical analysis).
Unlike a coin flip, however, we do not expect 50/50 black versus not-black. The ratio of black to not-black people in our society and in the relevant workplace population is not 50/50. So you need to work a little harder than in the coin flip example to figure out what is the expected percentage of black employees the employer would have if discrimination was not at issue. In Teamsters, the job category in question was line driver. Only 8 out of 1828 line drivers (just 0.4%) were black. If you looked at the company overall, though, you’d see that 5% of the employees were black. Therefore, the plaintiffs argued, we should expect that what we would expect most of the time, absent discrimination, is that 5% of the 1828 line drivers (= 91) should be black …. Just as we would expect that we’d get 50 heads if we flipped a coin 100 times.
So now you can imagine a normal distribution chart where 91 (instead of 50) is the number in the middle of the curve, the mean. You can have numbers above and below 91 and not be suspicious, but at some point a number that is too far away from 91 is going to be so improbable that we can infer it was not the result of random chance, it must instead be because of discrimination.
You know how to calculate this cutoff, from the work we did with the standard deviation equation, above!
N in the case of our coin flip calculation was 100, the total number of coin flips. Now, it is the number of time the employer made employment decisions for line drivers: 1828
P used to be the percentage of flips that we would expect to be heads, or 0.50. Now, P is the percentage of line drivers that we would expect to be black, or 0.05. (Where we got this 5% was discussed above – again, it’s the percentage of black employees in the company in general).
Q is always 100% minus P, expressed in decimals that correspond to percent. Above, it was the percentage of flips that we would expect to not be heads, which was also 0.50. Here, though it is 0.95, because that is 1.0 – 0.5.
Multiply N * P * Q, or 1828 x 0.05 x. 0.95. You get 86.83. Now take the square root of 86.83 and you get 9.32. The value of one standard deviation, σ, is 9.32. Two standard deviations is 19 (rounding). Three standard deviations is 28. What does this mean? This means that the odds of having a number of black line drivers that is higher than 91 + 28 = 119, or lower than 91 – 28 – 63, is less than 1%. The odds of that happening for any reason other than intentional is so small, that the Court inferred it had to be intentional discrimination. Now, remember how many black line drivers there were? Only 8! That is considerably less than 63. In fact, you can think of it this way: 63 is 3 standard deviations from the mean of 91; while 8 is (91-8)/9.32 = 9 (rounding) standard deviations away! If three standard deviations from the mean is suspicious, 9 is super-duper suspicious!
After you read the next case, Hazelwood, try running a statistical analysis to see for yourself whether the underrepresentation of Black teachers is significant enough to meet the burden for intentional discrimination (for simplicity, just do the numbers for the 1972-73 school year). As you read the case, find the actual number of black teachers that school year, compared to N, the total number of teachers that year. Ask what is the appropriate value for P, what percentage you would expect to be black in a normal distribution of teachers? See if you can understand why the government (for the plaintiffs) is arguing for one P, and the school district, for another P (see footnote 14 on p. 215). You can run the equation using each of the proposed P’s, to see if it makes a difference to the outcome. Calculate σ, and then use it to figure out whether the actual number of black teachers is far enough from the expected number (i.e., more than 3σ) to warrant a conclusion that the school district must have intended the disparity.
6.3 Hazelwood School District v. United States 6.3 Hazelwood School District v. United States
HAZELWOOD SCHOOL DISTRICT et al. v. UNITED STATES
No. 76-255.
Argued April 27, 1977
Decided June 27, 1977
*300Stewart, J., delivered the opinion of the Court, in which Bueger, C. J., and BrenNan, White, Marshall, BlacicmtiN, Powell, and RehNquist, JJ., joined. BrennaN, J., post, p. 313, and White, J., post, p. 347, filed concurring opinions. SteveNs, J., filed a dissenting opinion, post, p. 314.
William H. Allen argued the cause for petitioners. With him on the briefs were Coleman S. Hicks and Don O. Russell.
Deputy Solicitor General Wallace argued the cause for the United States. With him on the brief were Solicitor General McCree, Assistant Attorney General Days, Thomas S. Martin, Brian K. Landsberg, Walter W. Barnett, and Cynthia L. Attwood. *
delivered the opinion of the Court.
The petitioner Hazelwood School District covers 78 square miles in the northern part of St. Louis County, Mo. In 1973 the Attorney General brought this lawsuit against Hazelwood and various of its officials, alleging that they were engaged in a “pattern or practice” of employment discrimination in violation of Title VII of the Civil Rights Act of 1964, 78 Stat. 253, as amended, 42 U. S. C. § 2000e et seg. (1970 ed. and Supp. V).1 The complaint asked for an injunction requiring Hazelwood to cease its discriminatory practices, to take affirmative steps to obtain qualified Negro faculty members, and to offer employment and give backpay to victims of past illegal discrimination.
Hazelwood was formed from 13 rural school districts between 1949 and 1951 by a process of annexation. By the 1967-1968 school year, 17,550 students were enrolled in the •¡district, of whom only 59 were Negro; the number of Negro 'pupils increased to 576 of 25,166 in 1972-1973, a total of ■just over 2%.
From the beginning, Hazelwood followed relatively unstructured procedures in hiring its teachers. Every person requesting an application for a teaching position was sent one, and completed applications were submitted to a central per*302sonnel office, where they were kept on file.2 During the early 1960's the personnel office notified all applicants whenever a teaching position became available, but as the number of applications on file increased in the late 1960’s and early 1970’s, this practice was no longer considered feasible. The personnel office thus began the practice of selecting anywhere from 3 to 10 applicants for interviews at'the school where the vacancy existed. The personnel office did not substantively screen the applicants in determining which of them to send for interviews, other than to ascertain that each applicant, if selected, would be eligible for state certification by the time he began the job. Generally, those who had most recently submitted applications were jmost likely to be chosen for interviews.3
Interviews were conducted by a department chairman, program coordinator, or the principal at the school where the teaching vacancy existed. Although those conducting the interviews did fill out forms rating the applicants in a number of respects, it ^undisputed that each school principal pos-., sessed virtually, unlimited discretion in hiring teachers for his .school. The only general guidance given to the principals was to ' hire' the “most competent” person available, and such intangibles as “personality, disposition, appearance, poise, voice, articulation, and ability to deal with people” counted heavily. The principal’s choice was routinely honored by Hazelwood’s Superintendent and the Board of Education.
In the early 1960’s Hazelwood found it necessary to recruit new teachers, and for that purpose members of its staff visited a number of colleges and universities in Missouri and bordering States. All the institutions visited were predominantly white, and Hazelwood did not seriously recruit at either of the *303two predominantly Negro four-year colleges in Missouri.4 As a buyer’s market began to develop for public school teachers, Hazelwood curtailed its recruiting efforts. For the 1971-1972 school year, 3,127 persons applied for only 234 teaching vacancies; for the 1972-1973 school year, there were 2,373 applications for 282 vacancies. A number of the applicants who were not hired were Negroes.5 /
v Hazelwood hired its first Negro teacher in 1969. The number of Negro faculty members gradually increased in successive years: 6 of 957 in the 1970 school year; 16 of 1,107 by the end of the 1972 school year; 22 of 1,231 in the 1973 school year. By comparison, according to 1970 census figures, of more than 19,000 teachers employed in that year in the St. Louis' area, 15.4% were Negro. That percentage figure in-eluded the St. Louis City School District, which in recent years has followed a policy of attempting to maintain a 50% Negro teaching staff. Apart from that school district, 5.7% of the teachers in the county were Negro in 1970.
Drawing upon these historic facts, the Government mounted its “pattern or practice” attack in the District Court upon four different fronts. It adduced evidence of (1) a history of alleged racially discriminatory practices, (2) statistical disparities in hiring, (3) the standardless and largely subjective hiring procedures, and (4) specific instances of alleged discrimination against 55 unsuccessful Negro applicants for teaching jobs. Hazelwood offered virtually no additional evidence in response, relying instead on evidence introduced by the Government, perceived deficiencies in the Government’s case, and its own officially promulgated policy “to hire all *304teachers on the basis of training, preparation and recommendations, regardless of race, color or creed.” 6
The District Court ruled that the Government had failed to establish a pattern or practice of discrimination. The court was unpersuaded by -'the alleged history of discrimination, noting that no dual school system had ever existed in Hazel-woody, The statistics showing that relatively small numbers of Negroes were employed as teachers were found nonproba-tive, on the ground that the percentage of Negro pupils in Hazelwood was similarly small. The court found nothing illegal or suspect in the teacher-hiring procedures that Hazel-wood had followed. Finally, the court reviewed the evidence in the 55 cases of alleged individual discrimination, and after stating that the., burden of proving intentional discrimination was on the Government, it found that this burden had not been sustained in a single, instance. Hence, the court entered judgment for the defendants. 392 F. Supp. 1276 (ED Mo.).
The Court of Appeals for the Eighth Circuit reversed. 534 F. 2d 805. After suggesting that the District Court had assigned inadequate weight to evidence of discriminatory conduct on the part of Hazelwood before the effective date of Title VII/-7.the Court of Appeals rejected the trial court’s *305analysis of the statistical data as resting on an irrelevant comparison of Negro teachers to Negro pupils in Hazelwood. The proper comparison, in the appellate court's view, was one between Negro teachers in Hazelwood and Negro teachers in the relevant labor market area. Selecting St. Louis County and St. Louis City as the relevant area,8 the Court of Appeals compared the 1970 census figures, showing that 15.4% of teachers in that area were Negro, to the racial composition of Hazelwood’s teaching staff. In the 1972-1973 and 1973-1974 school years, only 1.4% and 1.8%, respectively, of Hazelwood’s teachers were Negroes. This statistical disparity, particularly when viewed against the background of the teacher-hiring procedures that Hazelwood had followed, was held to constitute a prima facie case of a pattern or practice of racial discrimination.
In addition, the Court of Appeals reasoned that the trial court had erred in failing to measure the 55 instances in which Negro applicants were denied jobs against the four-part standard for establishing a prima facie case of individual discrimination set out in this Court’s opinion in McDonnell Douglas Corp. v. Oreen, 411 U. S. 792, 802.9 Applying that *306standard, the appellate court found 16 cases of individual discrimination,10 which “buttressed” the statistical proof. Because Hazelwood had not rebutted the Government's prima facie case of a pattern or practice of racial discrimination, the Court of Appeals directed judgment for the Government and prescribed the remedial order to be entered.11
We granted certiorari, 429 U. S. 1037, to consider a substantial question affecting the enforcement of a pervasive federal law.
The petitioners primarily attack the judgment of the Court of Appeals for its reliance on “undifferentiated work force statistics to find an unrebutted prima facie case of employment discrimination.” 12 The question they raise, in short, is *307whether a basic component in the Court of Appeals’ finding of a pattern or practice of discrimination — the comparatively small percentage of Negro employees on Hazelwood’s teaching staff — was lacking in probative force.
This Court’s recent consideration in Teamsters v. United States, 431 U. S. 324, of the role of statistics in pattern-or-practice suits under Title VII provides substantial guidance in evaluating the arguments advanced by the petitioners. In that case we stated that it is the Government’s burden to “establish by a preponderance of the evidence that racial discrimination was the [employer’s] standard operating procedure — the regular rather than the unusual practice.” Id., at 336. We also noted that statistics can be an important source of proof in employment discrimination cases, since
“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. Evidence of long-lasting and gross disparity between the composition of a work force and that of the general population thus may be significant even though § 703 (j) makes clear that Title VII imposes no requirement that a work force mirror the general population.” Id., at 340 n. 20.
See also Arlington Heights v. Metropolitan Housing Dev. Corp., 429 U. S. 252, 266; Washington v. Davis, 426 U. S. 229, 241-242. Where gross statistical disparities can be shown, they alone may in a proper case constitute prima facie proof *308of a pattern or practice of discrimination. Teamsters, supra, at 339.
There can be no doubt, in light of the Teamsters case, that the District Court’s comparison of Hazelwood’s teacher work force to its student population fundamentally misconceived the role of statistics in employment discrimination cases. The Court of Appeals was correct in the view that a proper comparison was between the racial composition of Hazelwood’s teaching staff and the racial composition of the qualified public school teacher population in the relevant labor market.13 See Teamsters, supra, at 337-338, and n. 17. The percentage of Negroes on Hazelwood’s teaching staff in 1972-1973 was 1.4%, and in 1973-1974 it was 1.8%. By contrast, the percentage of qualified Negro teachers in the area was, according to the 1970 census, at least 5.7%.14 Although these differ-*309enees were on their face substantial, the Court of Appeals erred in substituting its judgment for that of the District Court and holding that the Government had conclusively proved its "pattern or practice” lawsuit.
The Court of Appeals totally disregarded the possibility that this prima facie statistical proof in the record might at the trial court level be rebutted by statistics dealing with Hazelwood’s hiring after it became subject to Title VII. Racial discrimination by public employers was not made illegal under Title VII until March 24, 1972. A public employer who from that date forward made all its employment decisions in a wholly nondiscriminatory way would not violate Title VII even if it had formerly maintained an all-white work force by purposefully excluding Negroes.15 For this rea*310son, the Court cautioned in the Teamsters opinion that once a prima facie case has been established by statistical workforce disparities, the employer must be given an opportunity to show that “the claimed discriminatory pattern is a product of pre-Act hiring rather than unlawful post-Act discrimination.” 431 U. S., at 360.
The record in this case showed that for the 1972-1973 school year, Hazelwood hired 282 new teachers, 10 of whom (3-.5%) were Negroes; for the following school year it hired 123 new teachers, 5 of whom (4.1%) were Negroes. Over the two-year period, Negroes constituted a total of 15 of the 405 new teachers hired (3.7%). Although the Court of Appeals briefly mentioned these data in reciting the facts, it wholly ignored them in discussing whether the Government had shown a pattern or practice of discrimination. And it gave no consideration at all to the possibility that post-Act data as to the number of Negroes hired compared to the total number of Negro applicants might tell a totally different story.16
What the hiring figures prove obviously depends upon the figures to which they are compared. The Court of Appeals accepted the Government’s argument that the relevant comparison was to the labor market area of St. Louis County and the city of St. Louis, in which, according to the 1970 census, 15.4% of all teachers were Negro. The propriety of that comparison was vigorously disputed by the petitioners, who urged that because the city of St. Louis has made special attempts to maintain a 50% Negro teaching staff, inclusion of *311that school district in the relevant market area distorts the comparison. Were that argument accepted, the percentage of Negro teachers in the relevant labor market area (St. Louis County alone) as shown in the 1970 census would be 5.7% rather than 15.4%.
The difference between these figures may well be important; the disparity between 3.7% (the percentage of Negro teachers hired by Hazelwood in 1972-1973 and 1973-1974) and 5.7% may be sufficiently small to weaken the Government's other proof, while the disparity between 3;7% and 15.4% may be sufficiently large to reinforce it.17 In determining *312which of the two figures — or, very possibly, what intermediate figure — provides the most accurate basis for comparison to the hiring figures at Hazelwood, it will be necessary to evaluate such considerations as (i) whether the racially based hiring policies of the St. Louis City School District were in effect as far back as 1970, the year in which the census figures were taken;18 (ii) to what extent those policies have changed the racial composition of that district's teaching staff from what it would otherwise have been; (iii) to what extent St. Louis’ recruitment policies have diverted to the city, teachers who might otherwise have applied to Hazelwood;19 (iv) to what extent Negro teachers employed by the city would prefer employment in other districts such as Hazelwood; and (v) what the experience in other school districts in St. Louis County indicates about the validity of excluding the City School District from the relevant labor market.
It is thus clear that a determination of the appropriate comparative figures in this case will depend upon further evaluation by the trial court. As this Court admonished in Teamsters: “[Statistics . . . come in infinite variety . . . . [T]heir usefulness depends on all of the surrounding facts and circumstances.” 431 U. S., at 340. Only the trial court is in a position to make the appropriate determination after further findings. And only after such a determination is made can a foundation be established for deciding whether or not Hazelwood engaged in a pattern or practice of racial *313discrimination in its employment practices in violation of the law.20
We hold, therefore, that the Court of Appeals erred in disregarding the post-Act hiring statistics in the record, and that it should have remanded the case to the District Court for further findings as to the relevant labor market area and for an ultimate determination of whether Hazelwood engaged in a pattern or practice of employment discrimination after March 24, 1972. 21 Accordingly, the judgment is vacated, and the case is remanded to the District Court for further proceedings consistent with this opinion.
It is so ordered.
[For concurring opinion of Mr. Justice White, see post, p. 347.]
concurring.
I join the Court’s opinion. Similarly to our decision in Dayton Board of Education v. Brinkman, post, p. 406, today’s opinion revolves around the relative factfinding roles of district courts and courts of appeals. It should be plain, however, that the liberal substantive standards for establishing a Title VII violation, including the usefulness of statistical proof, are reconfirmed.
In the present case, the District Court had adopted a wholly inappropriate legal standard of discrimination, and therefore *314did not evaluate the factual record before it in a meaningful way. This remand in effect orders it to do- so. It is my understanding, as apparently it is MR. Justice Stevens', post, at 318 n. 5, that the statistical inquiry mentioned by the Court, ante, at 311 n. 17, and accompanying text, can be of no help to the Hazelwood School Board in rebutting the Government’s evidence of discrimination. Indeed, even if the relative comparison market is found to be 5.7% rather than 15.4% black, the applicable statistical analysis at most will not serve to bolster the Government’s case. This obviously is of no aid to Hazelwood in meeting its burden of proof. Nonetheless I think that the remand directed by the Court is appropriate and will allow the parties to address these figures and calculations with greater care and precision. I also agree that given the misapplication of governing legal principles by the District Court, Hazelwood reasonably should be given the opportunity to come forward with more focused and specific applicant-flow data in the hope of answering the Government’s prima facie case. If, as presently seems likely, reliable applicant data are found to be lacking, the conclusion reached by my Brother Stevens will inevitably be forthcoming.
dissenting.
The basic framework in a pattern-or-practice suit brought by the Government under Title VII of the Civil Rights Act of 1964 is the same as that in any other lawsuit. The plaintiff has the burden of proving a prima facie case; if he does so, the burden of rebutting that case shifts to the defendant.1 In this *315case, since neither party complains that any relevant evidence was excluded, our task is to decide (1) whether the Government’s evidence established a prima facie case; and (2), if so, whether the remaining evidence is sufficient to carry Hazel-wood’s burden of rebutting that prima facie case.
I
The first question is clearly answered by the Government’s statistical evidence, its historical evidence, and its evidence relating to specific acts of discrimination.
One-third of the teachers hired by Hazelwood resided in the city of St. Louis at the time of their initial employment. As Mr. Justice Clark explained in his opinion for the Court of Appeals, it was therefore appropriate to treat the city, as well as the county, as part of the relevant labor market.2 *316In that market, 15% of the teachers were black. In the Hazelwood District at the time of trial less than 2% of the teachers were black. An even more telling statistic is that after Title VII became applicable to it, only 3.7% of the new teachers hired by Hazelwood were black. Proof of these gross disparities was in itself sufficient to make out a prima facie case of discrimination. See Teamsters v. United States, 431 U. S. 324, 339; Castaneda v. Partida, 430 U. S. 482, 494-498.
As a matter of history, Hazelwood employed no black teachers until 1969. Both before and after the 1972 amendment making the statute applicable to public school districts, Hazelwood used a standardless and largely subjective hiring procedure. Since “relevant aspects of the decisionmaking process had undergone little change,” it is proper to infer that the pre-Act policy of preferring white teachers continued to influence Hazelwood’s hiring practices.3
The inference of discrimination was corroborated by post-Act evidence that Hazelwood had refused to hire 16 qualified black applicants for racial reasons. Taking the Government’s evidence as a whole, there can be no doubt about the sufficiency of its prima facie case.
*317II
Hazelwood “offered virtually no additional evidence in response,” ante, at 303. It challenges the Government’s statistical analysis by claiming that the city of St. Louis should be excluded from the relevant market and pointing out that only 5.7% of the teachers in the county (excluding the city) were black. It further argues that the city’s policy of trying to maintain a 50% black teaching staff diverted teachers from the county to the city. There are two separate reasons why these arguments are insufficient: they are not supported by the evidence; even if true, they do not overcome the Government’s case.
The petitioners offered no evidence concerning wage differentials, commuting problems, or the relative advantages of teaching in an inner-city school as opposed to a suburban school. Without any such evidence in the record, it is difficult to understand why the simple fact that the city was the source of a third of Hazelwood’s faculty should not be sufficient to demonstrate that it is a part of the relevant market. The city’s policy of attempting to maintain a 50/50 ratio clearly does not undermine that conclusion, particularly when the record reveals no shortage of qualified black applicants in either Hazelwood or other suburban school districts.4 Surely not all of the 2,000 black teachers employed by the city were unavailable for employment in Hazelwood at the time of their initial hire.
But even if it were proper to exclude the city of St. Louis from the market, the statistical evidence would still tend to prove discrimination. With the city excluded, 5.7% of the teachers in the remaining market were black. On the basis of a random selection, one would therefore expect 5.7% of *318the 405 teachers hired by Hazelwood in the 1972-1973 and 1973-1974 school years to have been black. But instead of 23 black teachers, Hazelwood hired only 15, less than two-thirds of the expected number. Without the benefit of expert testimony, I would hesitate to infer that the disparity between 23 and 15 is great enough, in itself, to prove discrimination.5 It is perfectly clear, however, that whatever probative force this disparity has, it tends to prove discrimination and does absolutely nothing in the way of carrying Hazelwood’s burden of overcoming the Government’s prima facie case.
Absolute precision in the analysis of market data is too much to expect. We may fairly assume that a nondiscriminatory selection process would have resulted in the hiring of somewhere between the 15% suggested by the Government and the 5.7% suggested by petitioners, or perhaps 30 or 40 black teachers, instead of the 15 actually hired.6 On that assumption, the Court of Appeals’ determination that there were 16 individual cases of discriminatory refusal to hire black applicants in the post-1972 period seems remarkably accurate.
In sum, the Government is entitled to prevail on the present record. It proved a prima facie case, which Hazelwood failed to rebut. Why, then, should we burden a busy federal court with another trial? Hazelwood had an opportunity to offer evidence to dispute the 16 examples of racially motivated refusals to hire; but as the Court notes, the Court of Appeals has already “held that none of the 16 prima facie cases of *319individual discrimination had been rebutted by the petitioners. See 534 F. 2d 805, 814 (CA8).” Ante, at 306. n. 10. Hazelwood also had an opportunity to offer any evidence it could muster to show a change in hiring practices or to contradict the fair inference to be drawn from the statistical evidence. Instead, it “offered virtually no additional evidence in response,” ante, at 303.
Perhaps “a totally different story” might be told by other statistical evidence that was never presented, ante, at 310. No lawsuit has ever been tried in which the losing party could not have pointed to a similar possibility.7 It is always possible to imagine more evidence which could have been offered, but at some point litigation must come to an end.8
*320Rather than depart from well-established rules of procedure, I would affirm the judgment of the Court of Appeals.9 Since that judgment reflected a correct appraisal of the record, I see no reason to prolong this litigation with a remand neither side requested.10
6.4 Wal-Mart Stores, Inc. v. Dukes 6.4 Wal-Mart Stores, Inc. v. Dukes
WAL-MART STORES, INC. v. DUKES et al.
No. 10-277.
Argued March 29, 2011
Decided June 20, 2011
*340Theodore J. Boutrous, Jr., argued the cause for petitioner. With him on the briefs were Rachel S. Brass, Theane Evan-gelis Kapur, Theodore B. Olson, Mark A. Perry, and Amir C. Tayrani.
*341Joseph M. Sellers argued the cause for respondents. With him on the brief were Brad Seligman, Jocelyn D. Larkin, Christine E. Webber, Jenny R. Yang, Kalpana Kota-gal, Steven Stemerman, Elizabeth A. Lawrence, Arcelia Hurtado, Sheila Y. Thomas, Stephen Tinkler, and Merit Bennett *
delivered the opinion of the Court.
We are presented with one of the most expansive class actions ever. The District Court and the Court of Appeals approved the certification of a class comprising about one and a half million plaintiffs, current and former female employees of petitioner Wal-Mart who allege that the discretion exercised by their local supervisors over pay and promotion matters violates Title VII by discriminating against women. In addition to injunctive and declaratory relief, the plaintiffs seek an award of backpay. We consider whether the certification of the plaintiff class was consistent with Federal Rules of Civil Procedure 23(a) and (b)(2).
I
A
Petitioner Wal-Mart is the Nation's largest private employer. It operates four types of retail stores throughout the country: Discount Stores, Supercenters, Neighborhood Markets, and Sam’s Clubs. Those stores are divided into seven nationwide divisions, which in turn comprise 41 regions of 80 to 85 stores apiece. Each store has between 40 and 53 separate departments and 80 to 500 staff positions. In all, Wal-Mart operates approximately 3,400 stores and employs more than 1 million people.
*343Pay and promotion decisions at Wal-Mart are generally committed to local managers’ broad discretion, which is exercised “in a largely subjective manner.” 222 F. R. D. 137, 145 (ND Cal. 2004). Local store managers may increase the wages of hourly employees (within limits) with only limited corporate oversight. As for salaried employees, such as store managers and their deputies, higher corporate authorities have discretion to set their pay within preestablished ranges.
Promotions work in a similar fashion. Wal-Mart permits store managers to apply their own subjective criteria when selecting candidates as “support managers,” which is the first step on the path to management. Admission to Wal-Mart’s management training program, however, does require that a candidate meet certain objective criteria, including an above-average performance rating, at least one year’s tenure in the applicant’s current position, and a willingness to relocate. But except for those requirements, regional and district managers have discretion to use their own judgment when selecting candidates for management training. Promotion to higher office — e. g., assistant manager, co-manager, or store manager — is similarly at the discretion of the employee’s superiors after prescribed objective factors are satisfied.
B
The named plaintiffs in this lawsuit, representing the 1.5 million members of the certified class, are three current or former Wal-Mart employees who allege that the company discriminated against them on the basis of their sex by denying them equal pay or promotions, in violation of Title VII of the Civil Rights Act of 1964, 78 Stat. 255, as amended, 42 U. S. C. § 2000e~l et seq.1
*344Betty Dukes began working at a Pittsburg, California, Wal-Mart in 1994. She started as a cashier, but later sought and received a promotion to customer service manager. After a series of disciplinary violations, however, Dukes was demoted back to cashier and then to greeter. Dukes concedes she violated company policy, but contends that the disciplinary actions were in fact retaliation for invoking internal complaint procedures and that male employees have not been disciplined for similar infractions. Dukes also claims two male greeters in the Pittsburg store are paid more than she is.
Christine Kwapnoski has worked at Sam’s Club stores in Missouri and California for most of her adult life. She has held a number of positions, including a supervisory position. She claims that a male manager yelled at her frequently and screamed at female employees, but not at men. The manager in question “told [her] to ‘doll up,’ to wear some makeup, and to dress a little better.” App. 1003a.
The final named plaintiff, Edith Arana, worked at a Wal-Mart store in Duarte, California, from 1995 to 2001. In 2000, she approached the store manager on more than one occasion about management training, but was brushed off. Arana concluded she was being denied opportunity for advancement because of her sex. She initiated internal complaint procedures, whereupon she was told to apply directly to the district manager if she thought her store manager was being unfair. Arana, however, decided against that and never applied for management training again. In 2001, she was fired for failure to comply with Wal-Mart’s timekeeping policy.
These plaintiffs, respondents here, do not allege that Wal-Mart has any express corporate policy against the advancement of women. Rather, they claim that their local managers’ discretion over pay and promotions is exercised disproportionately in favor of men, leading to an unlawful disparate impact on female employees, see 42 U. S. C. *345§ 2000e-2(k). And, respondents say, because Wal-Mart is aware of this effect, its refusal to cabin its managers’ authority amounts to disparate treatment, see §2000e-2(a). Their complaint seeks injunctive and declaratory relief, punitive damages, and backpay. It does not ask for compensatory damages.
Importantly for our purposes, respondents claim that the discrimination to which they have been subjected is common to all Wal-Mart’s female employees. The basic theory of their case is that a strong and uniform “corporate culture” permits bias against women to infect, perhaps subconsciously, the discretionary decisionmaking of each one of Wal-Mart’s thousands of managers — thereby making every woman at the company the victim of one common discriminatory practice. Respondents therefore wish to litigate the Title VII claims of all female employees at Wal-Mart’s stores in a nationwide class action.
C
Class certification is governed by Federal Rule of Civil Procedure 23. Under Rule 23(a), the party seeking certification must demonstrate, first, that
“(1) the class is so numerous that joinder of all members is impracticable;
“(2) there are questions of law or fact common to the class;
“(3) the claims or defenses of the representative parties are typical of the claims or defenses of the class; and
“(4) the representative parties will fairly and adequately protect the interests of the class.”
Second, the proposed class must satisfy at least one of the three requirements listed in Rule 23(b). Respondents rely on Rule 23(b)(2), which applies when “the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or correspond*346ing declaratory relief is appropriate respecting the class as a whole.”2
Invoking these provisions, respondents moved the District Court to certify a plaintiff class consisting of “ ‘[a]ll women employed at any Wal-Mart domestic retail store at any time since December 26,1998 who have been or may be subjected to Wal-Mart’s challenged pay and management track promotions policies and practices.’” 222 F. R. D., at 141-142 (quoting Plaintiff’s Motion for Class Certification in Case No. 3:01-cv-02252-CRB (ND Cal), Doc. 99, p. 37). As evidence that there were indeed “questions of law or fact common to” all the women of Wal-Mart, as Rule 23(a)(2) requires, respondents relied chiefly on three forms of proof: statistical evidence about pay and promotion disparities between men and women at the company, anecdotal reports of discrimination from about 120 of Wal-Mart’s female employees, and the testimony of a sociologist, Dr. William Bielby, who conducted a “social framework analysis” of Wal-Mart’s “culture” and personnel practices, and concluded that the company was “vulnerable” to gender discrimination. 603 F. 3d 671, 601 (CA9 2010) (en banc).
Wal-Mart unsuccessfully moved to strike much of this evidence. It also offered its own countervailing statistical and other proof in an effort to defeat Rule 23(a)’s requirements *347of commonality, typicality, and adequate representation. Wal-Mart further contended that respondents’ monetary claims for backpay could not be certified under Rule 23(b)(2), first because that Rule refers only to injunctive and declaratory relief, and second because the backpay claims could not be manageably tried as a class without depriving Wal-Mart of its right to present certain statutory defenses. With one limitation not relevant here, the District Court granted respondents’ motion and certified their proposed class.3
D
A divided en banc Court of Appeals substantially affirmed the District Court’s certification order. 603 F. 3d 571. The majority concluded that respondents’ evidence of commonality was sufficient to “raise the common question whether Wal-Mart’s female employees nationwide were subjected to a single set of corporate policies (not merely a number of independent discriminatory acts) that may have worked to unlawfully discriminate against them in violation of Title VII.” Id., at 612 (emphasis deleted). It also agreed with the District Court that the named plaintiffs’ claims were sufficiently typical of the class as a whole to satisfy Rule 23(a)(3), and that they could serve as adequate class representatives, see Rule 23(a)(4). Id., at 614-615. With respect to the Rule 23(b)(2) question, the Ninth Circuit held that respondents’ backpay claims could be certified as part of a (b)(2) class because they did not “predominare]” over the requests for declaratory and injunctive relief, meaning they were not “superior in strength, influence, or authority” to *348the nonmonetary claims. Id., at 616 (internal quotation marks and brackets omitted).4
Finally, the Court of Appeals determined that the action could be manageably tried as a class action because the District Court could adopt the approach the Ninth Circuit approved in Hilao v. Estate of Marcos, 103 F. 3d 767, 782-787 (1996). There compensatory damages for some 9,541 class members were calculated by selecting 137 claims at random, referring those claims to a special master for valuation, and then extrapolating the validity and value of the untested claims from the sample set. See 603 F. 3d, at 625-626. The Court of Appeals “s[aw] no reason why a similar procedure to that used in Hilao could not be employed in this case.” Id., at 627. It would allow Wal-Mart “to present individual defenses in the randomly selected 'sample cases,’ thus revealing the approximate percentage of class members whose unequal pay or nonpromotion was due to something other than gender discrimination.” Ibid., n. 56 (emphasis deleted).
We granted certiorari. 562 U. S. 1091 (2010).
1 — <
The class action is “an exception to the usual rule that litigation is conducted by and on behalf of the individual named parties only.” Califano v. Yamasaki, 442 U. S. 682, 700-701 (1979). In order to justify a departure from that rule, “a class representative must be part of the class and 'possess the same interest and suffer the same injury’ as the *349class members.” East Tex. Motor Freight System, Inc. v. Rodriguez, 431 U. S. 395, 403 (1977) (quoting Schlesinger v. Reservists Comm. to Stop the War, 418 U. S. 208, 216 (1974)). Rule 23(a) ensures that the named plaintiffs are appropriate representatives of the class 'whose claims they wish to litigate. The Rule's four requirements — numerosity, commonality, typicality, and adequate representation — “effectively ‘limit the class claims to those fairly encompassed by the named plaintiff’s claims.’ ” General Telephone Co. of Southwest v. Falcon, 457 U. S. 147, 156 (1982) (quoting General Telephone Co. of Northwest v. EEOC, 446 U. S. 318, 330 (1980)).
A
The crux of this case is commonality — the rule requiring a plaintiff to show that “there are questions of law or fact common to the class.” Rule 23(á)(2).5 That language is easy to misread, since “[a]ny competently crafted class complaint literally raises common ‘questions.’ ” Naga-reda, Class Certification in the Age of Aggregate Proof, 84 N. Y. U. L. Rev. 97, 131-132 (2009). For example: Do all of us plaintiffs indeed work for Wal-Mart? Do our managers have discretion over pay? Is that an unlawful employment practice? What remedies should we get? Reciting these questions is not sufficient to obtain class certification. Com*350monality requires the plaintiff to demonstrate that the class members “have suffered the same injury,” Falcon, supra, at 157. This does not mean merely that they have all suffered a violation of the same provision of law. Title VII, for example, can be violated in many ways — by intentional discrimination, or by hiring and promotion criteria that result in disparate impact, and by the use of these practices on the part of many different superiors in a single company. Quite obviously, the mere claim by employees of the same company that they have suffered a Title VII injury, or even a disparate-impact Title VII injury, gives no cause to believe that all their claims can productively be litigated at once. Their claims must depend upon a common contention — for example, the assertion of discriminatory bias on the part of the same supervisor. That common contention, moreover, must be of such a nature that it is capable of classwide resolution— which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.
“What matters to class certification ... is not the raising of common ‘questions’ — even in droves — but, rather, the capacity of a class-wide proceeding to generate common answers apt to drive the resolution of the litigation. Dissimilarities within the proposed class are what have the potential to impede the generation of common answers.” Nagareda, supra, at 132.
Rule 23 does not set forth a mere pleading standard. A party seeking class certification must affirmatively demonstrate his compliance with the Rule — that is, he must be prepared to prove that there are in fact sufficiently numerous parties, common questions of law or fact, etc. We recognized in Falcon that “sometimes it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question,” 457 U. S., at 160, and that certification is proper only if “the trial court is satisfied, after a *351rigorous analysis, that the prerequisites of Rule 23(a) have been satisfied,” id., at 161; see id., at 160 (“[AJctual, not presumed, conformance with Rule 23(a) remains ... indispensable”). Frequently that “rigorous analysis” will entail some overlap with the merits of the''plaintiff’s underlying claim. That cannot be helped. “ ‘[T]he class determination generally involves considerations that are enmeshed in the factual and legal issues comprising the plaintiff’s cause of action.’” Id., at 160 (quoting Coopers & Lybrand v. Livesay, 437 U. S. 463, 469 (1978); some internal quotation marks omitted).6 Nor is there anything unusual about that consequence: The necessity of touching aspects of the merits in order to resolve *352preliminary matters, e. g., jurisdiction and venue, is a familiar feature of litigation. See Szabo v. Bridgeport Machines, Inc., 249 F. 3d 672, 676-677 (CA7 2001) (Easterbrook, J.).
In this case, proof of commonality necessarily overlaps with respondents’ merits contention that Wal-Mart engages in a pattern or practice of discrimination.7 That is so because, in resolving an individual’s Title VII claim, the crux of the inquiry is “the reason for a particular employment decision,” Cooper v. Federal Reserve Bank of Richmond, 467 U. S. 867, 876 (1984). Here respondents wish to sue about literally millions of employment decisions at once. Without some glue holding the alleged reasons for all those decisions together, it will be impossible to say that examination of all the class members’ claims for relief will produce a common answer to the crucial question why was I disfavored.
B
This Court’s opinion in Falcon describes how the commonality issue must be approached. There an employee who claimed that he was deliberately denied a promotion on account of race obtained certification of a class comprising all employees wrongfully denied promotions and all applicants wrongfully denied jobs. 457 U. S., at 152. We rejected that composite class for lack of commonality and typicality, explaining:
“Conceptually, there is a wide gap between (a) an indi- . vidual’s claim that he has been denied a promotion [or *353higher pay] on discriminatory grounds, and his otherwise unsupported allegation that the company has a policy of discrimination, and (b) the existence of a class of persons who have suffered the same injury as that individual, such that the individual’s claim and the class claims will share common questions of law or fact and that the individual's claim will be typical of the class claims.” Id., at 157-158.
Falcon suggested two ways in which that conceptual gap might be bridged. First, if the employer “used a biased testing procedure to evaluate both applicants for employment and incumbent employees, a class action on behalf of every applicant or employee who might have been prejudiced by the test clearly would satisfy the commonality and typicality requirements of Rule 23(a).” Id., at 159, n. 15. Second, “[significant proof that an employer operated under a general policy of discrimination conceivably could justify a class of both applicants and employees if the discrimination manifested itself in hiring and promotion practices in the same general fashion, such as through entirely subjective de-cisionmaking processes.” Ibid. We think that statement precisely describes respondents’ burden in this case. The first manner of bridging the gap obviously has no application here; Wal-Mart has no testing procedure or other company-wide evaluation method that can be charged with bias. The whole point of permitting discretionary decisionmaking is to avoid evaluating employees under a common standard.
The second manner of bridging the gap requires “[significant proof” that Wal-Mart “operated under a general policy of discrimination.” That is entirely absent here. Wal-Mart’s announced policy forbids sex discrimination, see App. 1567a-1596a, and as the District Court recognized the company imposes penalties for denials. of equal employment opportunity, 222 F. R. D., at 154. The only evidence of a “general policy of discrimination” respondents produced was the testimony of Dr. William Bielby, their sociological *354expert. Relying on “social framework” analysis, Bielby testified that Wal-Mart has a “strong corporate culture,” that makes it “vulnerable” to “gender bias.” Id., at 152. He could not, however, “determine with any specificity how regularly stereotypes play a meaningful role in employment decisions at Wal-Mart. At his deposition ... Dr. Bielby conceded that he could not calculate whether 0.5 percent or 95 percent of the employment decisions at Wal-Mart might be determined by stereotyped thinking.” 222 F. R. D. 189,192 (ND Cal. 2004). The parties dispute whether Bielby’s testimony even met the standards for the admission of expert testimony under Federal Rule of Evidence 702 and our Dau-bert case, see Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U. S. 579 (1993).8 The District Court concluded that Daubert did not apply to expert testimony at the certification stage of class-action proceedings. 222 F. R. D., at 191. We doubt that is so, but even if properly considered, Bielby’s testimony does nothing to advance respondents’ case. “[Wjhether 0.5 percent or 95 percent of the employment decisions at Wal-Mart might be determined by stereotyped thinking” is the essential question on which respondents’ theory of commonality depends. If Bielby admittedly has no answer to that question, we can safely disregard *355what he has to say. It is worlds away from “[significant proof” that Wal-Mart “operated under a general policy of discrimination.”
C
The only corporate policy that the plaintiffs’ evidence convincingly establishes is Wal-Mart’s “policy” of allowing discretion by local supervisors over employment matters. On its face, of course, that is just the opposite of a uniform employment practice that would provide the commonality needed for a class action; it is a policy against having uniform employment practices. It is also a very common and presumptively reasonable way of doing business — one that we have said “should itself raise no inference of discriminatory conduct,” Watson v. Fort Worth Bank & Trust, 487 U. S. 977, 990 (1988).
To be sure, we have recognized that, “in appropriate cases,” giving discretion to lower-level supervisors can be the basis of Title VII liability under a disparate-impact theory — since “an employer's undisciplined system of subjective decisionmaking [can have] precisely the same effects as a system pervaded by impermissible intentional discrimination.” Id., at 990-991. But the recognition that this type of Title VII claim “can” exist does not lead to the conclusion that every employee in a company using a system of discretion has such a claim in common. To the contrary, left to their own devices most managers in any corporation — and surely most managers in a corporation that forbids sex discrimination — would select sex-neutral, performance-based criteria for hiring and promotion that produce no actionable disparity at all. Others may choose to reward various attributes that produce disparate impact — such as scores on general aptitude tests or educational achievements, see Griggs v. Duke Power Co., 401 U. S. 424, 431-432 (1971). And still other managers may be guilty of intentional discrimination that produces a sex-based disparity. In such a company, demonstrating the invalidity of one manager’s use *356of discretion will do nothing to demonstrate the invalidity of another's. A party seeking to certify a nationwide class will be unable to show that all the employees’ Title VII claims will in fact depend on the answers to common questions.
Respondents have not identified a common mode of exercising discretion that pervades the entire company — aside from their reliance on Dr. Bielby’s social-framework analysis that we have rejected. In a company of Wal-Mart’s size and geographical scope, it is quite unbelievable that all managers would exercise their discretion in a common way without some common direction. Respondents attempt to make that showing by means of statistical and anecdotal evidence, but their evidence falls well short.
The statistical evidence consists primarily of regression analyses performed by Dr. Richard Drogin, a statistician, and Dr. Marc Bendick, a labor economist. Drogin conducted his analysis region by region, comparing the number of women promoted into management positions with the percentage of women in the available pool of hourly workers. After considering regional and national data, Drogin concluded that “there are statistically significant disparities between men and women at Wal-Mart . . . [and] these disparities . . . can be explained only by gender discrimination.” 603 F. 3d, at 604 (internal quotation marks omitted). Bendick compared work-force data from Wal-Mart and competitive retailers and concluded that Wal-Mart “promotes a lower percentage of women than its competitors.” Ibid.
Even if they are taken at face value, these studies are insufficient to establish that respondents’ theory can be proved on a elasswide basis. In Falcon, we held that one named plaintiff’s experience of discrimination was insufficient to infer that “discriminatory treatment is typical of [the employer’s employment] practices.” 457 U. S., at 158. A similar failure of inference arises here. As Judge Ikuta observed in her dissent, “[information about disparities at the regional and national level does not establish the existence *357of disparities at individual stores, let alone raise the inference that a company-wide policy of discrimination is implemented by discretionary decisions at the store and district level.” 603 F. 3d, at 637. A regional pay disparity, for example, may be attributable to only a small set of Wal-Mart stores, and cannot by itself establish the uniform, store-by-store disparity upon which the plaintiffs’ theory of commonality depends.
There is another, more fundamental, respect in which respondents’ statistical proof fails. Even if it established (as it does not) a pay or promotion pattern that differs from the nationwide figures or the regional figures in all' of Wal-Mart’s 3,400 stores, that would still not demonstrate that commonality of issue exists. Some managers will claim that the availability of women, or qualified women, or interested women, in their stores’ area does not mirror the national or regional statistics. And almost all of them will claim to have been applying some sex-neutral, performance-based criteria — whose nature and effects will differ from store to store. In the landmark case of ours which held that giving discretion to lower-level supervisors can be the basis of Title VII liability under a disparate-impact theory, the plurality opinion conditioned that holding ón the corollary that merely proving that the discretionary system has produced a racial or sexual disparity is not enough. “The plaintiff must begin by identifying the specific employment practice that is challenged.” Watson, supra, at 994; accord, Wards Cove Packing Co. v. Atonio, 490 U. S. 642, 656 (1989) (approving that statement), superseded by statute on other grounds, 42 U. S. C. § 2000e-2(k). That is all the more necessary when a class of plaintiffs is sought to be certified. Other than the bare existence of delegated discretion, respondents have identified no “specific employment practice” — much less one that ties all their 1.5 million claims together. Merely showing that Wal-Mart’s policy of discretion has produced an overall sex-based disparity does not suffice.
*358Respondents’ anecdotal evidence suffers from the same defects, and in addition is too weak to raise any inference that all the individual, discretionary personnel decisions are discriminatory. In Teamsters v. United States, 431 U. S. 324 (1977), in addition to substantial statistical evidence of companywide discrimination, the Government (as plaintiff) produced about 40 specific accounts of racial discrimination from particular individuals. See id., at 338. That number was significant because the company involved had only 6,472 employees, of whom 571 were minorities, id., at 337, and the class itself consisted of around 334 persons, United States v. T. I. M. E.-D. C., Inc., 517 F. 2d 299, 308 (CA5 1975), overruled on other grounds, Teamsters, supra. The 40 anecdotes thus represented roughly one account for every eight members of the class. Moreover, the Court of Appeals noted that the anecdotes came from individuals “spread throughout” the company who “for the most part” worked at the company’s operational centers that employed the largest numbers of the class members. 517 F. 2d, at 315, and n. 30. Here, by contrast, respondents filed some 120 affidavits reporting experiences of discrimination — about 1 for every 12,500 class members — relating to only some 235 out of Wal-Mart’s 3,400 stores. 603 F. 3d, at 634 (Ikuta, J., dissenting). More than half of these reports are concentrated in only 6 States (Alabama, California, Florida, Missouri, Texas, and Wisconsin); half of all States have only one or two anecdotes; and 14 States have no anecdotes about Wal-Mart’s operations at all. Id., at 634-635, and n. 10. Even if every single one of these accounts is true, that would not demonstrate that the entire company “operated] under a general policy of discrimination,” Falcon, 457 U. S., at 159, n. 15, which is what respondents must show to certify a companywide class.9
*359The dissent misunderstands the nature of the foregoing analysis. It criticizes our focus on the dissimilarities between the putative class members on the ground that we have “blend[ed]” Rule 23(a)(2)’s commonality requirement with. Rule 23(b)(3),s inquiry into whether common questions “predominate” over individual ones. See post, at 374-376 (Ginsburg, J., concurring in part and dissenting in part). That is not so. We quite agree that for purposes of Rule 23(a)(2) “ ‘[e]ven a single [common] question’ ” will do, post, at 376, n. 9 (quoting Nagareda, The Preexistence Principle and the Structure of the Class Action, 103 Colum. L. Rev. 149, 176, n. 110 (2003)). We consider dissimilarities not in order to determine (as Rule 23(b)(3) requires) whether common questions predominate, but in order to determine (as Rule 23(a)(2) requires) whether there is “[e]ven a single [common] question.” And there is not here. Because respondents provide no convincing proof of a companywide discriminatory pay and promotion policy, we have concluded that they have not established the existence of any common question.10
In sum, we agree with Chief Judge Kozinski that the members of the class
“held a multitude of jobs, at different levels of Wal-Mart’s hierarchy, for variable lengths of time, in 3,400 stores, sprinkled acroos 50 statec, with a kaleidoscope of supervisors (male and female), subject to a variety of *360regional policies that all differed .... Some thrived while others did poorly. They have little in common but their sex and this lawsuit.” 603 F. 3d, at 652 (dissenting opinion).
Ill
We also conclude that respondents’ claims for backpay were improperly certified under Federal Rule of Civil Procedure 23(b)(2). Our opinion in Ticor Title Ins. Co. v. Brown, 511 U. S. 117, 121 (1994) (per curiam), expressed serious doubt about whether claims for monetary relief may be certified under that provision. We now hold that they may not, at least where (as here) the monetary relief is not incidental to the injunctive or declaratory relief.
A
Rule 23(b)(2) allows class treatment when “the party opposing the class has acted or refused to act on grounds that apply generally to the class, so that final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole.” One possible reading of this provision is that it applies only to requests for such injunctive or declaratory relief and does not authorize the class certification of monetary claims at all. We need not reach that broader question in this case, because we think that, at a minimum, claims for individualized relief (like the backpay at issue here) do not satisfy the Rule. The key to the (b)(2) class is “the indivisible nature of the injunctive or declaratory remedy warranted — the notion that the conduct is such that it can be enjoined or declared unlawful only as to all of the class members or as to none of them.” Nagareda, 84 N. Y. U. L. Rev., at 132. In other words, Rule 23(b)(2) applies only when a single injunction or declaratory judgment would provide relief to each member of the class. It does not authorize class certification when each individual class member would be entitled to a different injunction or declaratory judgment against the defendant. Similarly, it does not *361authorize class certification when each class member would be entitled to an individualized award of monetary damages.
That interpretation accords with the history of the Rule. Because Rule 23 “stems from equity practice” that predated its codification, Amchem, Products, Inc. v. Windsor, 521 U. S. 591, 613 (1997), in determining its meaning we have previously looked to the historical models on which the Rule was based, Ortiz v. Fibreboard Corp., 527 U. S. 815, 841-845 (1999). As we observed in Amchem, “[c]ivil rights cases against parties charged with unlawful, class-based discrimination are prime examples” of what (b)(2) is meant to capture. 521 U. S., at 614. In particular, the Rule reflects a series of decisions involving challenges to racial segregation — conduct that was remedied by a single classwide order. In none of the cases cited by the Advisory Committee as examples of (b)(2)’s antecedents did the plaintiffs combine any claim for individualized relief with their classwide injunction. See Advisory Committee’s Note, 28 U. S. C. App., pp. 1260-1261 (1964 ed., Supp. II) (citing cases); e. g., Potts v. Flax, 313 F. 2d 284, 289, n. 5 (CA5 1963); Brunson v. Board of Trustees of School Dist. No. 1, Clarendon Cty., 311 F. 2d 107, 109 (CA4 1962) (per curiam); Frasier v. Board of Trustees of Univ. of N. C., 134 F. Supp. 589, 593 (MDNC 1955) (three-judge court), aff’d, 350 U. S. 979 (1956) (per curiam).
Permitting the combination of individualized and class-wide relief in a (b)(2) class is also inconsistent with the structure of Rule 23(b). Classes certified trader (b)(1) and (b)(2) share the most traditional justifications for class treatment— that individual adjudications would be impossible or unworkable, as in a (b)(1) class,11 or that the relief sought must per*362force affect the entire class at once, as in a (b)(2) class. For that reason these are also mandatory classes: The Rule provides no opportunity for (b)(1) or (b)(2) class members to opt out, and does not even oblige the District Court to afford them notice of the action. Rule 23(b)(3), by contrast, is an “adventuresome innovation” of the 1966 amendments, Am-chem, 521 U. S., at 614 (internal quotation marks omitted), framed for situations “in which 'class-action treatment is not as clearly called for/” id., at 615 (quoting Advisory Committee’s Notes, 28 U. S. C. Ápp., p. 697 (1994 ed.)). It allows class certification in. a much wider set of circumstances but with greater procedural protections. Its only prerequisites are that “the questions of law or fact common to class members predominate over any questions affecting only individual members, and that a class action is superior to other available methods for fairly and efficiently adjudicating the controversy.” Rule 23(b)(3). And unlike (b)(1) and (b)(2) classes, the (b)(3) class is not mandatory; class members are entitled to receive “the best notice that is practicable under the circumstances” and to withdraw from the class at their option. See Rule 23(c)(2)(B).
Given that structure, we think it clear that individualized monetary claims belong in Rule 23(b)(3). The procedural protections attending the (b)(3) class — predominance, superiority, mandatory notice, and the right to opt out — are missing from (b)(2) not because the Rule considers them unnecessary, but because it considers them unnecessary to a (b)(2) class. When a class seeks an indivisible injunction benefiting all its members at once, there is no reason to undertake a case-specific inquiry into whether class issues predominate or whether class action is a superior method of adjudicating *363the dispute. Predominance and superiority are self-evident. But with respect to each class member's individualized claim for money, that is not so — which is precisely why (b)(3) requires the judge to make findings about predominance and superiority before allowing the class. Similarly, (b)(2) does not require that class members be given notice and opt-out rights, presumably because it is thought (rightly or wrongly) that notice has no purpose when the class is mandatory, and that depriving people of their right to sue in this manner complies with the Due Process Clause. In the context of a class action predominantly for money damages we have held that absence of notice and opt out violates due process. See Phillips Petroleum Co. v. Shutts, 472 U. S. 797, 812 (1985). While we have never held that to be so where the monetary claims do not predominate, the serious possibility that it may be so provides an additional reason not to read Rule 23(b)(2) to include the monetary claims here.
B
Against that conclusion, respondents argue that their claims for backpay were appropriately certified as part of a class under Rule 23(b)(2) because those claims do not “predominate” over their requests for injunctive and declaratory relief. They rely upon the Advisory Committee’s statement that Rule 23(b)(2) “does not extend to cases in which the appropriate final relief relates exclusively or predominantly to money damages.” 28 U. S. C. App., p. 1260 (1964 ed., Supp. II) (emphasis added). The negative implication, they argue, is that it does extend to cases in which the appropriate final relief relates only partially and nonpredominantly to money damages. Of course it is the Rule itself, not the Advisory Committee’s description of it, that governs. And a mere negative inference does not in our view suffice to establish a disposition that has no basis in the Rule’s text, and that does obvious violence to the Rule’s structural features. The mere “predominance” of a proper (b)(2) injunctive claim *364does nothing to justify elimination of Rule 23(b)(3)’s procedural protections: It neither establishes the superiority of class adjudication over individual adjudication nor cures the notice and opt-out problems. We fail to see why the Rule should be read to nullify these protections whenever a plaintiff class, at its option, combines its monetary claims with a request — even a “predominating request” — for an injunction.
Respondents’ predominance test, moreover, creates perverse incentives for class representatives to place at risk potentially valid claims for monetary relief. In this case, for example, the named plaintiffs declined to include employees’ claims for compensatory damages in their complaint. That strategy of including only backpay claims made it more likely that monetary relief would not “predominate.” But it also created the possibility (if the predominance test were correct) that individual class members’ compensatory-damages claims would be precluded by litigation they had no power to hold themselves apart from. If it were determined, for example, that a particular class member is not entitled to backpay because her denial of increased pay or a promotion was not the product of discrimination, that employee might be collaterally estopped from independently seeking compensatory damages based on that same denial. That possibility underscores the need for plaintiffs with individual monetary claims to decide for themselves whether to tie their fates to the class representatives’ or go it alone — a choice Rule 23(b)(2) does not ensure that they have.
The predominance test would also require the District Court to reevaluate the roster of class members continually. The Ninth Circuit recognized the necessity for this when it concluded that those plaintiffs no longer employed by Wal-Mart lack standing to seek injunctive or declaratory relief against its employment practices. The Court of Appeals’ response to that difficulty, however, was not to eliminate all former employees from the certified class, but to eliminate only those who had left the company’s employ by the date *365the complaint was filed. That solution has no logical connection to the problem, since those who have left their Wal-Mart jobs since the complaint was filed have no more need for prospective relief than those who left beforehand. As a consequence, even though the validity of a (b)(2) class depends on whether “final injunctive relief or corresponding declaratory relief is appropriate respecting the class as a whole,” Rule 23(b)(2) (emphasis added), about half the members of the class approved by the Ninth Circuit have no claim for injunctive or declaratory relief at all. Of course, the alternative (and logical) solution of excising plaintiffs from the class as they leave their employment may have struck the Court of Appeals as wasteful of the District Court’s time. Which indeed it is, since if a backpay action were properly certified for class treatment under (b)(8), the ability to litigate a plaintiff’s backpay claim as part of the class would not turn on the irrelevant question whether she is still employed at Wal-Mart. What follows from this, however, is not that some arbitrary limitation on class membership should be imposed but that the backpay claims should not be certified under Rule 23(b)(2) at all.
Finally, respondents argue that their backpay claims are appropriate for a (b)(2) class action because a backpay award is equitable in nature. The latter may be true, but it is irrelevant. The Rule does not speak of “equitable” remedies generally but of injunctions and declaratory judgments. As Title VII itself makes pellucidly clear, backpay is neither. See 42 U. S. C. § 2000e-5(g)(2)(B)(i) and (ii) (distinguishing between declaratory and injunctive relief and the payment of “backpay,” see § 2000e~5(g)(2)(A)).
C
In Allison v. Citgo Petroleum Corp., 151 F. 3d 402, 415 (CA5 1998), the Fifth Circuit held that a (b)(2) class would permit the certification of monetary relief that is “incidental to requested injunctive or declaratory relief,” which it de*366fined as “damages that flow directly from liability to the class as a whole on the claims forming the basis of the injunc-tive or declaratory relief.” In that court’s view, such “incidental damages should not require additional hearings to resolve the disparate merits of each individual’s case; it should neither introduce new and substantial legal or factual issues, nor entail complex individualized determinations.” Ibid. We need not decide in this case whether there are any forms of “incidental” monetary relief that are consistent with the interpretation of Rule 23(b)(2) we have announced and that comply with the Due Process Clause. Respondents do not argue that they can satisfy this standard, and in any event they cannot.
Contrary to the Ninth Circuit’s view, Wal-Mart is entitled to individualized determinations of each employee’s eligibility for backpay. Title VII includes a detailed remedial scheme. If a plaintiff prevails in showing that an employer has discriminated against him in violation of the statute, the court “may enjoin the respondent from engaging in such unlawful employment practice, and order such affirmative action as.may be appropriate, [including] reinstatement or hiring of employees, with or without back pay ... or any other equitable relief as the court deems appropriate.” §2000e-5(g)(1). But if the employer can show that it took an adverse employment action against an employee for any reason other than discrimination, the court cannot order the “hiring, reinstatement, or promotion of an individual as an employee, or the payment to him of any back pay.” § 2000e-5(g)(2)(A).
We have established a procedure for trying pattern-or-practice cases that gives effect to these statutory requirements. When the plaintiff seeks individual relief such as reinstatement or backpay after establishing a pattern or practice of discrimination, “a district court must usually conduct additional proceedings ... to determine the scope of individual relief.” Teamsters, 431 U. S., at 361. At this phase, the burden of proof will shift to the company, but it *367will have the right to raise any individual affirmative defenses it may have, and to “demonstrate that the individual applicant was denied an employment opportunity for lawful reasons.” Id., at 362.
The Court of Appeals believed that it was possible to replace such proceedings with Trial by Formula. A sample set of the class members would be selected, as to whom liability for sex discrimination and the backpay owing as a result would be determined in depositions supervised by a master. The percentage of claims determined to be valid would then be applied to the entire remaining class, and the number of (presumptively) valid claims thus derived would be multiplied by the average backpay award in the sample set to arrive at the entire class recovery — without further individualized proceedings. 603 F. 3d, at 626-627. We disapprove that novel project. Because the Rules Enabling Act forbids interpreting Rule 23 to “abridge, enlarge or modify any substantive right,” 28 U. S. C. § 2072(b); see Ortiz, 627 U. S., at 845, a class cannot be certified on the premise that Wal-Mart will not be entitled to litigate its statutory defenses to individual claims. And because the necessity of that litigation will prevent backpay from being “incidental” to the classwide injunction, respondents’ class could not be certified even assuming, arguendo, that “incidental” monetary relief can be awarded to a 23(b)(2) class.
* * *
The judgment of the Court of Appeals is
Reversed.
with whom Justice Breyer, Justice Sotomayor, and Justice Kagan join,
concurring in part and dissenting in part.
The class in this case, I agree with the Court, should not have been certified under Federal Rule of Civil Procedure 23(b)(2). The plaintiffs, alleging discrimination in violation *368of Title VII, 42 U. S. C. § 2000e et seq., seek monetary relief that is not merely incidental to any injunctive or declaratory relief that might be available. See ante, at 360-367. A putative class of this type may be certifiable under Rule 23(b)(3), if the plaintiffs show that common class questions “predominate” over issues affecting individuals — e. g., qualification for, and the amount of, backpay or compensatory damages — and that a class action is “superior” to other modes of adjudication.
Whether the class the plaintiffs describe meets the specific requirements of Rule 23(b)(3) is not before the Court, and I would reserve that matter for consideration and decision on remand.1 The Court, however, disqualifies the class at the starting gate, holding that the plaintiffs cannot cross the “commonality” line set by Rule 23(a)(2). In so ruling, the Court imports into the Rule 23(a) determination concerns properly addressed in a Rule 23(b)(3) assessment.
I — I
<J
Rule 23(a)(2) establishes a preliminary requirement for maintaining a class action: “[T]here are questions of law or fact common to the class.” 2 The Rule “does not require that all questions of law or fact raised in the litigation be com*369mon,” 1 H. Newberg & A. Conte, Newberg on Class Actions §3.10, pp. 3-48 to 3-49 (3d ed. 1992); indeed, “[e]ven a single question of law or fact common to the members of the class will satisfy the commonality requirement,” Nagareda, The Preexistence Principle and the Structure of the Class Action, 103 Colum. L. Rev. 149, 176, n. 110 (2003). See Advisory Committee’s 1937 Notes on Fed. Rule Civ. Proc. 23, 28 U. S. C. App., p. 138 (citing with approval cases in which “there was only a question of law or fact common to” the class members).
A “question” is ordinarily understood to be “[a] subject or point open to controversy.” American Heritage Dictionary 1483 (3d ed. 1992). See also Black’s Law Dictionary 1366 (9th ed. 2009) (defining “question of fact” as “[a] disputed issue to be resolved . . . [at] trial” and “question of law” as “[a]n issue to be decided by the judge”). Thus, a “question” “common to the class” must be a dispute, either of fact or of law, the resolution of which will advance the determination of the class members’ claims.3
B
The District Court, recognizing that “one significant issue common to the class may be sufficient to warrant certification,” 222 F. R. D. 137, 145 (ND Cal. 2004), found that the plaintiffs easily met that test. Absent an error of law or an abuse of discretion, an appellate tribunal has no warrant to upset the District Court’s finding of commonality. See Califano v. Yamasaki, 442 U. S. 682, 703 (1979) (“[M]ost issues arising under Rule 23 ... [are] committed in the first instance to the discretion of the district court.”).
*370The District Court certified a class of “[a]ll women employed at any Wal-Mart domestic retail store at any time since December 26,1998.” 222 F. R. D., at 141-143 (internal quotation marks omitted). The named plaintiffs, led by Betty Dukes, propose to litigate, on behalf of the class, allegations that Wal-Mart discriminates on the basis of gender in pay and promotions. They allege that the company “[r]e-li[es] on gender stereotypes in making employment decisions such as . . . promotion^] [and] pay.” App. 55a. Wal-Mart permits those prejudices to infect personnel decisions, the plaintiffs contend, by leaving pay and promotions in the hands of “a nearly all male managerial workforce” using “arbitrary and subjective criteria.” Ibid. Further alleged barriers to the advancement of female employees include the company’s requirement, “as a condition of promotion to management jobs, that employees be willing to relocate.” Id., at 56a. Absent instruction otherwise, there is a risk that managers will act on the familiar assumption that women, because of their services to husband and children, are less mobile than men. See Dept, of Labor, Federal Glass Ceiling Commission, Good for Business: Making Full Use of the Nation’s Human Capital 151 (1995).
Women fill 70 percent of the hourly jobs in the retailer’s stores but make up only “33 percent of management employees.” 222 F. R. D., at 146. “[T]he higher one looks in the organization the lower the percentage of women.” Id., at 155. The plaintiffs’ “largely uncontested descriptive statistics” also show that women working in the company’s stores “are paid less than men in every region” and “that the salary gap widens over time even for men and women hired into the same jobs at the same time.” Ibid.] cf. Ledbetter v. Goodyear Tire & Rubber Co., 550 U. S. 618, 643 (2007) (Ginsburg, J., dissenting).
The District Court identified “systems for . . . promoting in-store employees” that were “sufficiently similar across regions and stores” to conclude that “the manner in which *371these systems affect the class raises issues that are common to all class members.” 222 P. R. D., at 149. The selection of employees for promotion to in-store management “is fairly characterized as a ‘tap on the shoulder’ process,” in which managers have discretion about whose shoulders to tap. Id., at 148. Vacancies are not regularly posted; from among those employees satisfying minimum qualifications, managers choose whom to promote on the basis of their own subjective impressions. Ibid.
Wal-Mart’s compensation policies also operate uniformly across stores, the District Court found. The retailer leaves open a $2 band for every position’s hourly pay rate. Wal-Mart provides- no standards or criteria for setting wages within that band, and thus does nothing to counter unconscious bias on the part of supervisors. See id., at 146-147.
Wal-Mart’s supervisors do not make their discretionary decisions in a vacuum. The District Court reviewed means Wal-Mart used to maintain a “carefully constructed . . . corporate culture,” such as frequent meetings to reinforce the common way of thinking, regular transfers of managers between stores to ensure uniformity throughout the company, monitoring of stores “on a close and constant basis,” and “Wal-Mart TV,” “broadcas[t]... into all stores.” Id., at 151— 153 (internal quotation marks omitted).
The plaintiffs’ evidence, including class members’ tales of their own experiences,4 suggests that gender bias suffused Wal-Mart’s company culture. Among illustrations, senior management often refer to female associates as “little Janie *372Qs.” Plaintiffs’ Motion for Class Certification in No. 3:01-cv-02252-CRB (ND Cal.), Doc. 99, p. 21 (internal quotation marks omitted). One manager told an employee that “[m]en are here to make a career and women aren’t.” 222 F. R. D., at 166 (internal quotation marks omitted). A committee of female Wal-Mart executives concluded that “[stereotypes limit the opportunities offered to women.” Plaintiffs’ Motion for Class Certification in No. 3:01-cv-02252-CRB (ND Cal.), Doc. 99, at 24 (internal quotation marks omitted).
Finally, the plaintiffs presented an expert’s appraisal to show that the pay and promotions disparities at Wal-Mart “can be explained only by gender discrimination and not by . . . neutral variables.” 222 F. R. D., at 155. Using regression analyses, their expert, Richard Drogin, controlled for factors including, inter alia, job performance, length of time with the company, and the store where an employee worked. Id., at 159.6 The results, the District Court found, were sufficient to raise an “inference of discrimination.” Id., at 155-160.
C
The District Court’s identification of a common question, whether Wal-Mart’s pay and promotions policies gave rise to unlawful discrimination, was hardly infirm. The practice of delegating to supervisors large discretion to make personnel decisions, uncontrolled by formal standards, has long been known to have the potential to produce disparate effects. Managers, like all humankind, may be prey to biases *373of which they are unaware.6 The risk of discrimination is heightened when those managers are predominantly of one sex, and are steeped in a corporate culture that perpetuates gender stereotypes.
The' plaintiffs5 allegations' resemble those in one of the prototypical cases in this area, Leisner v. New York Tel. Co., 358 F. Supp. 359, 364-365 (SDNY 1973). In deciding on promotions, supervisors in that case were to start with objective measures; but ultimately, they were to “look at the individual as a total individual.” Id., at 365 (internal quotation marks omitted). The final question they were to ask and answer: “Is this person going to be successful in our business?” Ibid, (internal quotation marks omitted). It is hardly surprising that for many managers, the ideal candidate was someone with characteristics similar to their own.
We have held that “discretionary employment practices” can give rise to Title VII claims, not only when such practices are motivated by discriminatory intent but also when they produce discriminatory results. See Watson v. Fort Worth Bank & Trust, 487 U. S. 977, 988, 991 (1988). But see ante, at 357 (“[Pjroving that [a] discretionary system has produced a . . . disparity is not enough.”). In Watson, as here, an employer had given its managers large authority over promotions. An employee sued the bank under Title VII, alleging that the “discretionary promotion system” *374caused a discriminatory effect based on race. 487 U. S., at 984 (internal quotation marks omitted). Four different supervisors had declined, on separate occasions, to promote the employee. Id., at 982. Their reasons were subjective and unknown. The employer, we noted, “had not developed precise and formal criteria for evaluating candidates”; “[i]t relied instead on the subjective judgment of supervisors.” Ibid.
Aware of “the problem of subconscious stereotypes and prejudices,” we held that the employer’s “undisciplined system of subjective decisionmaking” was an “employment practic[e]” that “may be analyzed under the disparate impact approach.” Id., at 990-991. See also Wards Cove Packing Co. v. Atonio, 490 U. S. 642, 657 (1989) (recognizing “the use of ‘subjective decision making’” as an “employment prac-tie[e]” subject to disparate-impact attack).
The plaintiffs’ allegations state claims of gender discrimination in the form of biased decisionmaking in both pay and promotions. The evidence reviewed by the District Court adequately demonstrated that resolving those claims would necessitate examination of particular policies and practices alleged to affect, adversely and globally, women employed at Wal-Mart’s stores. Rule 23(a)(2), setting a necessary but not a sufficient criterion for class-action certification, demands nothing further.
II
A'
The Court gives no credence to the key dispute common to the class: whether Wal-Mart’s discretionary pay and promotion policies are discriminatory. See ante, at 349 (“Reciting” questions like “Is [giving managers discretion over pay] an unlawful employment practice?” “is not sufficient to obtain class certification.”). “What matters,” the Court asserts, “is not the raising of common ‘questions,’” but whether there are “[dissimilarities within the proposed *375class” that “have the potential to impede the generation of common answers.” Ante, at 350 (quoting Nagareda, Class Certification in the Age of Aggregate Proof, 84 N. Y. U. L. Rev. 97, 132 (2009); some internal quotation marks omitted).
The Court blends Rule 23(a)(2)’s threshold criterion with the more demanding criteria of Rule 23(b)(3), and thereby elevates the (a)(2) inquiry so that it is no longer “easily satisfied,” 5 J. Moore et al., Moore’s Federal Practice §23.23[2], p. 23-72 (3d ed. 2011).7 Rule 23(b)(3) certification requires, in addition to the four 23(a) findings, determinations that “questions of law or fact common to class members predominate over any questions affecting only individual members” and that “a class action is superior to other available methods for . . . adjudicating the controversy.”8
*376The Court’s emphasis on differences between class members mimics the Rule 23(b)(8) inquiry into whether common questions “predominate” over individual issues. And by asking whether the individual differences “impede” common adjudication, ante, at 350 (internal quotation marks omitted), the Court duplicates 23(b)(3)’s question whether “a class action is superior” to other modes of adjudication. Indeed, Professor Nagareda, whose “dissimilarities” inquiry the Court endorses, developed his position in the context of Rule 23(b)(3). See 84 N. Y. U. L. Rev., at 131 (Rule 23(b)(3) requires “some decisive degree of similarity across the proposed class” because it “speaks of common ‘questions’ that ‘predominate’ over individual ones”).9 “The Rule 23(b)(3) predominance inquiry” is meant to “tes[t] whether proposed classes are sufficiently cohesive to warrant adjudication by representation.” Amchem Products, Inc. v. Windsor, 521 U. S. 591, 623 (1997). If courts must conduct a “dissimilarities” analysis at the Rule 23(a)(2) stage, no mission remains for Rule 23(b)(3).
Because Rule 23(a) is also a prerequisite for Rule 23(b)(1) and Rule 23(b)(2) classes, the Court’s “dissimilarities” position is far reaching. Individual differences should not bar a Rule 23(b)(1) or Rule 23(b)(2) class, so long as the Rule 23(a) threshold is met. See id., at 623, n. 19 (Rule 23(b)(1)(B) “does not have a predominance requirement”); Yamasaki, 442 U. S., at 701 (Rule 23(b)(2) action in which the Court noted that “[i]t is unlikely that differences in the factual background of each claim will affect the outcome of the legal *377issue”). For example, in Franks v. Bowman Transp. Co., 424 U. S. 747 (1976), a Rule 23(b)(2) class of African-American truckdrivers complained that the defendant had discriminatorily refused to hire black applicants. We recognized that the “qualification[s] and performance” of individual class members might vary. Id., at 772 (internal quotation marks omitted). “Generalizations concerning such individually applicable evidence,” we cautioned, “cannot serve as a justification for the denial of [injunctive] relief to the entire class.” Ibid.
B
The “dissimilarities” approach leads the Court to train its attention on what distinguishes individual class members, rather than on what unites them. Given the lack of standards for pay and promotions, the majority says, “demonstrating the invalidity of one manager’s use of discretion will do nothing to demonstrate the invalidity of another’s.” Ante, at 356-356.
Wal-Mart’s delegation of discretion over pay and promotions is a policy uniform throughout all stores. The very nature of discretion is that people will exercise it in various ways. A system of delegated discretion, Watson held, is a practice actionable under Title YII when it produces discriminatory outcomes. 487 U. S., at 990-991; see supra, at 373-374. A finding that Wal-Mart’s pay and promotions practices in fact violate the law would be the first step in the usual order of proof for plaintiffs seeking individual remedies for companywide discrimination. Teamsters v. United States, 431 U. S. 324, 359 (1977); see Albemarle Paper Co. v. Moody, 422 U. S. 405, 415-423 (1975). That each individual employee’s unique circumstances will ultimately determine whether she is entitled to backpay or damages, § 2000e-5(g)(2)(A) (barring backpay if a plaintiff “was refused . . . advancement ... for any reason other than discrimination”), should not factor into the Rule 23(a)(2) determination.
*378* * *
The Court errs in importing a “dissimilarities” notion suited to Rule 23(b)(3) into the Rule 23(a) commonality inquiry. I therefore cannot join Part II of the Court’s opinion.