WALLACE, Chief Judge:
The San Francisco Police Officers' Association (Union) appeals from a district court order approving the City and County of San Francisco's (City) proposal to "band" scores from the 1989 promotion examinations for police sergeant and assistant inspector. The Union challenges this scoring procedure on the grounds that it employs racial preferences in violation of (1) Title VII of the Civil Rights Act of 1964 (Title VII), 42 U.S.C. § 2000e-2000h-6, as amended by the Civil Rights Act of 1991, Pub.L. No. 102-166, 105 Stat. 1071 (1991 Act), (2) the equal protection clause of the Fourteenth Amendment, and (3) the consent decree entered earlier in this action. The Union also contends that banding is not a "substantially equally valid" alternative to rank order promotions under the Uniform Guidelines on Employee Selection Procedures, 29 C.F.R. § 1607 (1990) (Uniform Guidelines). The district court had jurisdiction pursuant to 42 U.S.C. §§ 2000e, 1981, 1983. We have jurisdiction over this interlocutory appeal pursuant to 28 U.S.C. § 1292(a)(1). We affirm.
This appeal arises out of the district court's August 21, 1991, order declaring "banding" a legally valid scoring procedure for promotion examinations to certain ranks in the San Francisco Police Department (Department). The order permits the City to treat scores that fall within a statistically derived "band" as substantively equivalent for purpose of the knowledge, skills, and abilities measured by the examination.
The City seeks to band some test scores as a means of complying with the consent decree entered earlier in this action. The consent decree settled two race and sex discrimination actions brought by the Officers for Justice and the United States pursuant to 42 U.S.C. §§ 1981 and 1983, and Title VII. The Union intervened as a defendant in both of these actions. All parties later entered into the consent decree, which the district court approved in March 1979. See Officers for Justice v. Civil Serv. Comm'n, 473 F.Supp. 801 (N.D.Cal. 1979) (Officers for Justice), aff'd, 688 F.2d 615 (9th Cir.1982), cert. denied, 459 U.S. 1217, 103 S.Ct. 1219, 75 L.Ed.2d 456 (1983).
The consent decree, which is based upon an undisputed history of discrimination, prohibits the City from unlawfully discriminating against any employee or applicant for employment with the Department on the basis of race, sex, or national origin. Id. at 811. It also provides for "specific, definable and good faith efforts to be made [by the City] to achieve certain goals for employment of women and minorities" within a specified time period. Id. at 810. It establishes as a target the appointment of minorities and women to sergeant and assistant inspector positions in proportion to their representation in the qualified applicant pool and sets a long term goal of 45 percent minority representation in the Department. Id. at 813-15. In addition, the consent decree prohibits the use of selection procedures that have an adverse impact on minorities and women unless they are proven to be valid under the Uniform Guidelines. Id. at 812.
This is not the first time we have examined the Department's promotion procedures. See San Francisco Police Officers' Ass'n v. City and County of San Francisco, 812 F.2d 1125, 1132 (9th Cir.1988), op. withdrawn, on reh'g, 842 F.2d 1126, 1128 (9th Cir.1988), amended, 869 F.2d 1182 (9th Cir.1988) (Police Officers' III), cert. denied, 493 U.S. 816, 110 S.Ct. 68, 107 L.Ed.2d 35 (1989). In Police Officers' III, we disapproved of a scoring system that changed the relative weight of different components of an examination as a means of minimizing its adverse impact on minorities. We held that result-oriented reweighting of examination results unlawfully discriminated against candidates on the basis of race and gender. 869 F.2d at 1184. We decided, however, not to penalize those individuals who had been promoted under the reweighting plan. Id.
On remand, the district court ordered the City to administer new examinations and to appoint 100 sergeants and 60 assistant inspectors on the basis of the results. The examinations were administered in November and December 1989. Despite the City's efforts to administer a fair examination, which included the hiring of outside professional test developers, the 1989 examinations again produced an adverse impact on minorities.
Rather than discard the examination results, or administer a new examination, the City proposed to "band" the scores in an effort to meet the requirements of the consent decree. The "band" is a statistically derived confidence range that is applied to the examination results. Differences between scores within the band are considered to be statistically insignificant due to measurement error inherent in scoring the examination. The City proposed to treat scores within the band as identical for purposes of determining promotion eligibility. For those within the same band, the City could then consider secondary criteria, not measured by the examination, in selecting candidates for promotion. When the top scorer within a band is selected for promotion, the band would "slide" down and the next highest score would serve as the reference point. In this way, no candidate would be denied a promotion in favor of another whose score is statistically significantly lower than his or hers.
Banding is premised on the belief that minor differences in test scores do not reliably
Through the application of a statistical measurement of reliability, the City determined the width of the bands to be applied to the examination results in this case. The City computed the standard error of measurement by multiplying the standard deviation of scores from the examination by the square root of one minus the reliability coefficient. The City's application of two standard deviations established a confidence factor of 1.96 which, when multiplied by the standard error of measurement, produced a band of 4.0 points for the sergeant's examination and 5.48 points for the assistant inspector's examination. The sergeant band encompassed 52 of 747 candidates, the assistant inspector's 71 of 831.
The City then sought a declaratory judgment from the district court that its banding proposal was legal. The City proposed to make the 100 promotions ordered by the district court in strict rank order and 15 additional promotions pursuant to its banding proposal. The City also indicated that it would use race as the sole criterion for promoting candidates whose scores were within the same band. Minority candidates would be selected first from within each band. When no minorities remained, the highest scorers would be selected until the band "slid" down to encompass additional minority candidates. Race would thus be the paramount consideration in the selection of candidates from within the band.
The district court heard two and one-half days of testimony from five experts on the validity of banding as a selection method. The court ruled that the City's plan violated the restrictions on race conscious remedies set forth by the Supreme Court in University of California Regents v. Bakke, 438 U.S. 265, 98 S.Ct. 2733, 57 L.Ed.2d 750 (1978), and by this court in Higgins v. City of Vallejo, 823 F.2d 351 (9th Cir.1987) (Higgins), cert. denied, 489 U.S. 1051, 109 S.Ct. 1310, 103 L.Ed.2d 579 (1989). The district court indicated, however, that banding was an appropriate use of test scores and that the City's computation of the band was acceptable. The court also suggested that a modified proposal, along the lines approved in Bridgeport Guardians, Inc. v. City of Bridgeport, 933 F.2d 1140 (2d Cir.), cert. denied, ___ U.S. ___, 112 S.Ct. 337, 116 L.Ed.2d 277 (1991), would be considered. The City then proposed three criteria in addition to race — professional conduct, education, and training and experience — to be used in selecting candidates for promotion from within the band. The district court accepted this modification and declared banding and the proposed selection criteria "legally sound."
We first consider whether the use of racial preferences as a "plus" factor in promotion decisions violates Title VII as amended by the 1991 Act, the equal protection clause, and the consent decree. Whether the City's "banding" proposal comports with Title VII and equal protection are questions of law subject to de novo review, Bouman v. Block, 940 F.2d 1211, 1220-21 (9th Cir.), cert. denied, ___ U.S. ___, 112 S.Ct. 640, 116 L.Ed.2d 658 (1991), as is the district court's holding that the banding proposal is lawful under the consent decree. Vertex Distributing, Inc. v. Falcon Foam Plastics, Inc., 689 F.2d 885, 893 (9th Cir.1982). The burden lies with the Union to show that the City's voluntary affirmative action plan is unconstitutional or impermissible under Title VII. See Johnson v. Transportation Agency, Santa Clara County, 480 U.S. 616, 626, 107 S.Ct. 1442, 1448, 94 L.Ed.2d 615 (1987) (Johnson); Wygant v. Jackson Bd. of Education, 476 U.S. 267, 277-78, 106 S.Ct. 1842, 1848-49, 90 L.Ed.2d 260 (1986) (Wygant).
Title VII proscribes employer discrimination against any individual on the basis of race with respect to the terms and conditions of employment. 42 U.S.C. § 2000e-2(a). This proscription, however, does not bar the use of race conscious affirmative action programs when certain criteria are met. Johnson, 480 U.S. at 628-30, 107 S.Ct. at 1450-51. Under Title VII analysis, "voluntary adoption of a race-based remedy may be justified by a showing that a `manifest imbalance' exists, reflecting underrepresentation of women and minorities in `traditionally segregated job categories.'" Davis v. City and County of San Francisco, 890 F.2d 1438, 1448 (9th Cir.1989) (Davis), citing Johnson, 480 U.S. at 631, 107 S.Ct. at 1451; United Steelworkers of America v. Weber, 443 U.S. 193, 99 S.Ct. 2721, 61 L.Ed.2d 480 (1979).
The Union does not challenge the propriety of the affirmative action program under traditional Title VII analysis. Instead, it argues that section 107(a) of the 1991 Act, amending Title VII, prohibits the consideration of race as a factor in any employment decision. Because the 1991 Act was passed after the district court's ruling in this case, the question arises whether the 1991 Act applies retroactively. It is unnecessary to reach this issue, however, because the Union's argument on the merits is incorrect.
Section 107(a) of the 1991 Act provides that:
§ 107(a), 105 Stat. at 1075 (codified as amended at 42 U.S.C. § 2000e-2(m)) (emphasis added). The City asserts, and the Union does not dispute, that the savings clause of section 107 encompasses section 116 of the 1991 Act. Section 116 provides that "[n]othing in the amendments made by this title shall be construed to affect court-ordered remedies, affirmative action, or conciliation agreements, that are in accordance with the law." § 116, 105 Stat. at 1071. The Union argues that the phrase "in accordance with law" refers to the law as amended by the 1991 Act and that because section 107 prohibits the use of race as a motivating factor, "even though other factors also motivated the practice," the race conscious promotions at issue in this case are not "in accordance with law."
The Union's reading of the 1991 Act is predicated on an internal inconsistency: that Congress sought to protect affirmative action in section 116 while outlawing it in section 107. Such an interpretation should be "avoided if alternative interpretations consistent with the legislative purpose are available." Griffin v. Oceanic Contractors, Inc., 458 U.S. 564, 575, 102 S.Ct. 3245, 3252, 73 L.Ed.2d 973 (1982). The City properly argues that a more natural reading of the phrase "in accordance with law" is that affirmative action programs that were in accordance with law prior to passage of the 1991 Act are unaffected by the amendments. The language of the statute is clear, and the City's interpretation is consistent with that language.
In addition, the 1991 Act responded to several Supreme Court Title VII decisions, none of which addressed the question of affirmative action. In reversing the result of those decisions, Congress did not state that it also sought to overturn affirmative action. "[A]bsent a clear manifestation of contrary intent, a newly-enacted or revised statute is presumed to be harmonious with existing law and its judicial construction." Johnson v. First Nat'l Bank, 719 F.2d 270, 277 (8th Cir.1983), cert. denied, 465 U.S. 1012, 104 S.Ct. 1015, 79 L.Ed.2d 245 (1984). Therefore, we conclude that the 1991 Act does not alter existing affirmative action case law.
In its reply brief, the Union argues that banding is prohibited by section 106 of the 1991 Act, which provides that it is unlawful "to adjust the scores of, use different cutoff scores for, or otherwise alter the results of, employment related tests on the basis of race, color, religion, sex, or
The Union next argues that the City's banding proposal violates the equal protection clause of the Fourteenth Amendment. It contends that the effect of past discrimination and the present racial imbalance in the Department are addressed in the consent decree and therefore cannot serve as reasons for considering race in promotion decisions. It also contends that the mere imbalance in the results of the 1989 examination does not justify use of race as a factor without an affirmative showing that the examination actually discriminated against minorities. Because the City did not attempt to make this showing, the Union asserts that the City's use of race as a "plus" factor violates equal protection.
Public employers are permitted to use race as a factor in selecting between qualified applicants pursuant to a "narrowly tailored" affirmative action plan designed to remedy past unlawful discrimination. United States v. Paradise, 480 U.S. 149, 167, 107 S.Ct. 1053, 1064, 94 L.Ed.2d 203 (1987) (Paradise) (compelling state interest "unquestionably" exists where a race conscious employment program remedies past and present discrimination by a state actor). Under the strict scrutiny of equal protection analysis, a district court's finding must exhibit a "strong basis in the evidence" that remedial action was necessary. Davis, 890 F.2d at 1446. Statistical evidence of disparity sufficient to support a prima facie case under Title VII may, in some cases, constitute a strong basis in the evidence for believing that a voluntary affirmative action program was required by, and consistent with, the Constitution. Wygant, 476 U.S. at 292, 106 S.Ct. at 1856 (O'Connor, J. concurring); Hazelwood School Dist. v. United States, 433 U.S. 299, 307-08, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977); see also Johnson, 480 U.S. at 633-34 n. 11, 649, 107 S.Ct. at 1452-53 n. 11, 1461.
The Union's arguments must be viewed within the historical context of this litigation. At the outset, the district court found a severe underrepresentation of minorities in four areas of the Department. Prior to the signing and approval of the consent decree, the district court determined that there existed a prima facie case of discrimination against minorities and women in the upper ranks of the Department. In addition, the evidence before the district court when it entered the order presently appealed from indisputably showed that minorities continue to be underrepresented in both entry and promotional ranks. Finally, the results of the 1989 examination reflected an adverse impact on minority candidates. The district court determined that these statistics, coupled with the City's admission of past discrimination, established a prima facie case of discrimination. The district court's determination of prior discriminatory policies and conduct, coupled with continued evidence of discriminatory impact, satisfies the "strong basis in the evidence" test for voluntary affirmative action. See Paradise, 480 U.S. at 167 n. 18, 107 S.Ct. at 1064 n. 18; Davis, 890 F.2d at 1446-47.
The Union contends that banding, as a voluntary affirmative action program, is unlawful because there was no finding by the district court that the promotional examinations did not satisfy the requirements of the Uniform Guidelines. See 29 C.F.R. § 1607.3B (1990). The Union argues that the City may not use racial preferences in making promotion decisions without first showing that the examinations were invalid. It contends that use of race as a "plus" factor under these circumstances
We have not required that a particular examination be proven invalid in order to implement a voluntary race conscious affirmative action program designed to remedy prior discrimination. For example, in Higgins the employer never proved the examination invalid, yet we found the affirmative action plan constitutional. 823 F.2d at 358-60. The Union's argument is rejected.
The Union argues that the City's banding proposal violates the consent decree, which requires the City to "refrain from engaging in any act or practice which has the purpose or effect of unlawfully discriminating against any employee." Officers for Justice, 473 F.Supp. at 811. The Union contends that the banding proposal violates the consent decree because it unlawfully discriminates against nonminority candidates on the basis of race. Implicit in this contention is the claim that the consent decree prohibits the City from implementing a legitimate affirmative action plan.
The consent decree does prohibit the City from "unlawfully discriminating" in its employment decisions. However, it also contemplates race conscious relief, defined as "good faith efforts ... to achieve certain [numerical] goals." Id. at 810. One of the long term goals of the consent decree is to eradicate the effects of past hiring decisions with respect to the promotion of minorities and women. It aims to increase minority representation on the police force to resemble more closely the ethnic, racial, and sexual composition of the relevant labor force. Id. at 815. Therefore, the consent decree itself provides a proper rationale for race conscious promotions and serves as a valid defense against the Union's reverse discrimination arguments. Johnson, 480 U.S. at 626-27, 107 S.Ct. at 1448-49; Higgins, 823 F.2d at 355.
The Union also contends that banding is unlawful because it is not an equally valid alternative to rank order promotions and is therefore not required or authorized by the Uniform Guidelines. The Uniform Guidelines provide that "[w]here two or more selection procedures are available which serve the user's legitimate interest in efficient and trustworthy workmanship, and which are substantially equally valid for a given purpose, the user should use the procedure which has been demonstrated to have the lesser adverse impact." 29 C.F.R. § 1607.3B (1990). The district court found that the banding proposal "is more valid, or substantially equally valid to strict rank ordering." We will not disturb the district court's finding that banding is a valid alternative selection device unless it is clearly erroneous. Anderson v. Bessemer City, 470 U.S. 564, 573, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985); Clady v. County of Los Angeles, 770 F.2d 1421, 1427, 1429 (9th Cir.1985), cert. denied, 475 U.S. 1109, 106 S.Ct. 1516, 89 L.Ed.2d 915 (1986).
The Union argues that banding is not an equally valid selection procedure because strict rank order promotions are more likely to result in the selection of the most qualified individuals. The Union contends that even though a candidate is likely to receive different scores on subsequent administrations of the same examination, it is more likely than not that the relative positions of any two candidates will remain unchanged.
The district court, after hearing extensive expert testimony, decided that the City's computation of the band was appropriate, citing approval of the same measure of statistical deviation by the Supreme Court in Castaneda v. Partida, 430 U.S. 482, 496-97 n. 17, 97 S.Ct. 1272, 1281 n. 17, 51 L.Ed.2d 498 (1977), and the Code of Federal Regulations, 29 C.F.R. § 1607.14(B)(5) (1990). The district court heard evidence that the test scores themselves are imprecise measurements of future job performance and that by banding the scores and considering secondary criteria in making promotion decisions, the City may increase the probability that the most qualified
The Union also contends that the City's banding proposal is improper and that strict rank order promotions are required unless the City first determines that the examination scores themselves are invalid under the Uniform Guidelines. This argument misconceives the application of the Uniform Guidelines to this case. The Uniform Guidelines do not forbid the use of alternative selection procedures until one procedure is proven invalid.
On the contrary, before utilizing a procedure that has an adverse impact on minorities, the City has an obligation pursuant to the Uniform Guidelines to explore alternative procedures and to implement them if they have less adverse impact and are substantially equally valid to rank ordering. See 29 C.F.R. § 1607.3B (1990). It is undisputed that the results of the 1989 tests had an adverse impact against minority candidates for promotion. In this circumstance, rather than prohibiting banding, the Uniform Guidelines counsel in favor of its implementation to mitigate the adverse impact of the examination.
In 1989, following decades of concededly discriminatory promotional procedures, the City in concert with the Union, minority job applicants, and the court finally devised a selection process which offers a facially neutral way to interpret actual scores and reduce adverse impacts on minority candidates while preserving merit as the primary criterion for selection. Today we hold that the banding process is valid as a matter of constitutional and federal law. We do not overlook the fact that this resolution is the culmination of years of trial and error by the City and careful oversight by the district court. Although all parties were ultimately unable to agree to the proposal, we find that the efforts exerted in this process culminated in a unique and innovative program which succeeds in addressing past harms to minorities while minimizing future harmful effects on nonminority candidates. The successful efforts of all parties and the district court in reaching this resolution are to be lauded.