296 F.3d 524
7th Cir.
2002
POSNER, Circuit Judge.

This suit by whitе Chicago police sergeants and lieutenants challenges, as a denial of equal protection of the laws, the promotion in 1990 and 1991 of 20 black, Hispanic, and female sergeants and lieutenants to the rank of lieutenant and captain respectively. The challenged promotions were made pursuant to an affir mative action plan by which blacks, Hispanics, and women could be promoted “out of rank,” that is, promoted . even though they had a lower score than a white male on the test for the promotion. The district judge entered judgment after a jury trial and a partial retrial that he ordered, as he was authorized by Fed. R.Civ.P. 49(b) tо do by an inconsistency between two of the answers that the jury gave to the special interrogatories that it had been told to answer. See, e.g., Turyna v. Martam Construction Co., 83 F.3d 178, 181 (7th Cir.1996); King v. Ford Motor Co., 209 F.3d 886, 895 (6th Cir.2000); Austin v. Paramount Parks, Inc., 195 F.3d 715, 725 (4th Cir.1999). The judgment was for the City with .respect to all the promotions except that of the one Hispanic in the pool, who was promoted from sergeant to lieutenant. The plaintiffs appeal the ruling that the promotions of the blacks and women ahead of them did not deny the equal protection of the laws, while the City appeals the ruling that the promotion of the Hispanic sergeant ahead of the plaintiff sergeants was a denial of equal protection.

An initial question is the standard of review of jury findings in a racial-discrimination case. Although the plaintiffs’ position is unclear, it appears to be that we should review the findings de novo, which would amount to making our own findings on the basis of the evidence. The plaintiffs derive this position from judicial statements that to justify racial discrimination the defendant must have “a strong basis in evidenсe” for it. E.g., Miller v. Johnson, 515 U.S. 900, 922, 115 S.Ct. 2475, 132 L.Ed.2d 762 (1995); Johnson v. Board of Regents, 263 F.3d 1234, 1244 (11th Cir.2001). The plaintiffs have wrenched this language out of context: Racial discrimination even of the “affirmative action” sort, when practiced by a public agency and thus subject to the equal protection clause, requires proof, and not merely argument, that the agency had a compelling need to discriminаte and that it went no further in discrimination than necessary to meet that need. E.g., McNamara v. City of Chicago, 138 F.3d 1219, 1222 (7th Cir.1998); Wittmer v. Peters, 87 F.3d 916, 918-19 (7th Cir.1996); Associated General Contractors of Ohio, Inc. v. Drabik, 214 F.3d 730, 735 (6th Cir.2000). Argument in so sensitive an area of human relations must not, the courts believe, be allowed to draw on “common sense,” which might be inflected by stereotypes. See Danskine v. Miami Dade Fire Dep’t, 253 F.3d 1288, 1294-95 (11th Cir.2001); Contractors Ass’n of Eastern Pennsylvania, Inc. v. City of Philadelphia, 91 F.3d 586, 597 (3d Cir.1996); Hayes v. North State Law Enforcement Officers Ass’n, 10 F.3d 207, 214 (4th Cir.1993). But the requirement that there be рroof and not merely conjecture to justify racial discrimination even of the relatively benign, non-stigmatizing sort, and the allocation of factfinding responsibilities ‍‌​‌​​‌‌‌‌‌​‌​‌​​‌‌​​‌‌​​​‌​​​‌​​‌​​‌‌​‌‌​​​​‌​‌‌‍between trial and appellate court, are two different things. The jury in a discrimination case has the same responsibility. to resolve factual disputes that it has in any оther case, subject to the same standard of review. Worth v. Tyer, 276 F.3d 249, 266 (7th Cir.2001); Susan Wakeen Doll Co. v. Ashton Drake Galleries, 272 F.3d 441, 451 (7th Cir.2001); All Care Nursing Service, Inc. v. High Tech Staffing Services, Inc., 135 F.3d 740, 749 (11th Cir.1998); Tamez v. City of San Marcos, 118 F.3d 1085, 1094 (5th Cir.1997); United States v. Tolliver, 116 F.3d 120, 125 (5th Cir.1997).

But what is that standard? The cases we just cited all use the clearly-erroneous standard to review jury findings, yet other cases say that in a federal civil case, by virtue of the Seventh Amendment, reviewing courts owe more deferenсe to a jury’s findings than to findings by a judge. See, e.g., District of Columbia v. Pace, 320 U.S. 698, 701, 64 S.Ct. 406, 88 L.Ed. 408 (1944); Artis v. Hitachi Zosen Clearing, Inc., 967 F.2d 1132, 1139 (7th Cir.1992); Richardson v. Suzuki Motor Co., 868 F.2d 1226, 1235 (Fed.Cir.1989). The standard of appellate review applicable to judge and jury findings is at least verbally different: a judge’s finding of fact can be set aside if clearly erroneous, Fed. R.Civ.P. 52(a), but, as explained in Artis, a jury’s determination can be set aside only if “there is no legally sufficient evidentiary basis for a reasonable jury to find for that party [the party opposing judgment as a matter of law].” Fed.R.Civ.P. 50(a). Yet the canonical formulation of the elearly-erroneous standard is that it requires for reversal that the reviewing court have “a definite .and firm conviction that a mistake has been committed.” Concrete Pipe & Products of California, Inc. v. Construсtion Laborers Pension Trust for Southern California, 508 U.S. 602, 622, 113 S.Ct. 2264, 124 L.Ed.2d 539 (1993); see also, e.g., United States v. Smith, 103 F.3d 600, 606 (7th Cir.1996). This is strong language; how different is its import from that of deciding that a jury’s finding was unreasonable? And if there is a difference, is it one within the cognitive capacity of a reviewing court to discern?, For we have remarked a number of times that there are limits to the fineness of the distinctions that judges are able to make. United States v. Hill, 196 F.3d 806, 808 (7th Cir.1999); United States v. Boyd, 55 F.3d 239, 242 (7th Cir.1995); Johnson v. Trigg, 28 F.3d 639, 643-44 (7th Cir.1994); Morales v. Yeutter, 952 F.2d 954, 957 (7th Cir.1991); Haugh v. Jones & Laughlin Steel Corp., 949 F.2d 914, 917 (7th Cir.1991); United States v. McKinney, 919 F.2d 405, 421-23 (7th Cir.1990) (concurring opinion). But this is a general issue of judicial epistemology, in no way special to ‍‌​‌​​‌‌‌‌‌​‌​‌​​‌‌​​‌‌​​​‌​​​‌​​‌​​‌‌​‌‌​​​​‌​‌‌‍cases involving racial discrimination; and, as we are about to see, it does not have to be resolved in this case.

Once the facts are found, the question becomes whether they demonstrate a forbiddеn racial preference. The fact of a preference, of discrimination, is just that —a fact. Pullman-Standard v. Swint, 456 U.S. 273, 287-90, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982). The question whether if is an unlawful preference is a question about the application of the law to the fact. As with many other such questions in constitutional cases, it is to be decided as if it were a pure question of law, that is, with no deference given to the finder of fact, whether judge or jury. Grutter v. Bollinger, 288 F.3d 732, 738 (6th Cir.2002) (en banc); Hunter v. Regents of University of California, 190 F.3d 1061, 1063 (9th Cir.1999); Wessmann v. Gittens, 160 F.3d 790, 795 (1st Cir.1998); Engineering Contractors Ass’n of South Florida Inc. v. Metropolitan Dade County, 122 F.3d 895, 905 (11th Cir.1997).

The findings of fact made by the jury in this case are neither clearly erroneous nor unreasonable (assuming these are meaningfully, ascertainably distinct criteria), and what they reveal, so far as the black and female affirmative-action promotions are concerned, is the following. Until Orlando Wilson became the City of Chicago’s police commissioner in 1960, black and white police officers were segregated, with black officers being confined to the parts of Chicago that were predominantly blaсk. However, blacks were hired roughly in proportion to their share of the Chicago population. Wilson desegregated the police force, ■ The City presented evidence that this resulted in a decline in the hiring of blacks. That may seem a para- doxieal consequence of desegregation, but the evidence dispels the paradox. The evidence shows thаt white officers didn’t want to serve with blacks. There were numerous acts of racial harassment of blacks, and black applicants flunked the police medical exam at rates suspiciously higher than whites. As a result of these circumstances, applications of blacks to the police force plummeted. After reforms in the mid-1970s that are acknowledged to have eliminated or at least greatly reduced racial discrimination in the Chicago police department, the rate of black applications climbed, and black applicants no ' longer flunkéd the medical exam at rates significantly higher than whites.

The evidence that we have briefly summarized justified, a finding that discrimination by members of the police force depressed the hiring of blacks during the 1960s, leading in turn to a deficit of blacks in senior positions in the 1980s. The affirmative-action promotions of blacks challenged here, promotions designed to remedy the discrimination that we have just described, involved thе promotion of 11 black sergeants out of a total of 182 promotions of sergeants and 3 black lieutenants out of a total of 50 promotions of lieutenants. These affirmative-action promotions resulted in .percentages of black sergeants and lieutenants that still were lower than would have been expeсted had there not been that decline in the entry-level hiring of blacks in tfie 1960s.

The evidence that the decline- was the result of discrimination was not conclusive. The plaintiffs presented evidence that the decline was the result of racial tensions in the 1960s that made blacks reluctant to become police officers because it would make them unpopular with other blacks. Part of the evidence consisted of a report on the Chicago police department by the Justice Department’s Law Enforcement Assistance Administration. Against this the City presented evidence that although racial tensions were not limited to Chicago during the 1960s, other cities did not experience a drop off in black hiring for their police forces, which suggests that something other-than racial-tensions probably accounted for the drop off in Chicago; that something other may have been discrimination. The conflict over this issue was one for the jury to resolve, subject to deferential review. For the issue was purely factual: the cause of the depressed hiring of blacks in the sixties. Since remedying past discrimination is a recognized justification for affirmative action, and since the action taken was modest — the promotion out of rank of a mere handful of blacks, resulting merely in delayed promotion fоr some whites rather than in anyone’s losing his job or failing (eventually) to get the promotion he sought and was entitled to — the conclusion that the defendant had not violated the equal protection clause followed directly from the jury’s factual findings. Chicago Firefighters Local 2 v. City of Chicago, 249 F.3d 649 (7th Cir.2001); Majeske v. City of Chicago, 218 F.3d 816 (7th Cir.2000), cert. denied, 531 U.S. 1079, 121 S.Ct. 779, 148 L.Ed.2d 676 (2001); McNamara v. City of Chicago, 138 F.3d 1219 (7th Cir.1998); Stuart v. Roache, 951 F.2d 446 (1st Cir.1991); Donaghy v. City of Omaha, 933 F.2d 1448 (8th Cir.1991).

The case for the affirmative-action promotion of the five women was even stronger. Until the 1970s women were formally barred from being hired for most jobs in ‍‌​‌​​‌‌‌‌‌​‌​‌​​‌‌​​‌‌​​​‌​​​‌​​‌​​‌‌​‌‌​​​​‌​‌‌‍the police department, including patrol officer. As a result, few were hired and many were deterred from applying because of the truncated career opportunities.

The plaintiffs argue that even if all this is true, the City’s effort to defend its affirmative-action promotions is barred by the doctrine of judicial estoppel, which to discourage fraud in the litigation process bars a litigant who has obtained a judgment on the basis of proving one set of facts from obtaining a second judgment by turning around and proving that the facts were actually the opposite of what he had proved in the prior ease. New Hampshire v. Maine, 532 U.S. 742, 749-51, 121 S.Ct. 1808, 149 L.Ed.2d 968 (2001); DeVito v. Chicago Park District, 270 F.3d 532, 534 (7th Cir.2001); Bethesda Lutheran Homes & Services, Inc. v. Born, 238 F.3d 853, 857-58 (7th Cir.2001); Lydon v. Boston Sand & Gravel Co., 175 F.3d 6, 12-13 (1st Cir.1999). In an earlier suit, in which the City was charged with discriminating against blacks in hiring for the police force, we reversed the district court’s finding that the City had violated the equal protection clause, on the ground that the City had not been proved to have engaged in intentional discrimination. At the time the case wаs tried, it was widely believed that discrimination did not have to be intentional in order to violate the equal protection clause. But while the case was on appeal, the Supreme Court held that it did have to be intentional, Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976), and that is why this court reversed the finding of an equal protection violation. United States v. City of Chicago, 549 F.2d 415, 435 (7th Cir.1977). Intentional discrimination had not been at issue in the case and there was no finding that the City had not engaged in intentional discrimination and hence no inconsistency with the City’s position in the present case, one of mea culpa. And anyway the City had not obtained a favorable judgment in the earlier case — -for we upheld the district court’s ruling that the City had violated Title VII and the relief ordered would have been no different had the City been found to have engaged in intentional discrimination as well. The City obtained no advantage from not acknowledging that it had engaged in such discrimination.

The City defends the affirmative-action promotions of the blacks and the women on the further, alternative ground that they were justified by thе operational needs of the police force' — a ground completely different from the remedial ground that we have been discussing. One ground is enough, so we need not consider the alternative ground except with regard to the Hispanic who the district court determined had been improperly promoted. For the City’s argument is not that his promotion was justifiable as a remedy against past discrimination against Hispanics — the disparity between the percentage of Hispanic policemen and the percentage of Hispanic Chicagoans in 1990 was due largely ‍‌​‌​​‌‌‌‌‌​‌​‌​​‌‌​​‌‌​​​‌​​​‌​​‌​​‌‌​‌‌​​​​‌​‌‌‍to the rapid growth of the city’s Hispanic population in the 1980s— but that it was justifiable in order to make the police force more effective in performing its duties. By 1990 the population of Chicago was almost 20 percent Hispanic but fewer than 5 percent of police lieutenants were Hispanic. Because there were only 14 Hispanic lieutenants before the affirmative-actiоn promotion of Sergeant Denk and the police work in three shifts with the result that the number of lieutenants per shift is small, on any given shift in 1990 only two or three Hispanic lieutenants were on duty in the entire city, with its population of 2.78 million in 1990 of whom more than 500,000 were Hispanic.

The evidence presented by the City, including the testimony of a competеnt expert witness, established a two-fold need for a larger number of Hispanic lieutenants. First, lieutenants and captains are the principal supervisors in the police department. They set the tone for the department. If there are negligible numbers of Hispanics in these ranks (and in 1990 the percentage of Hispanic captains was only half the percentage of Hispanic lieutenants), non-Hispanic police officers are less likely to be sensitized to any special problems in policing Hispanic neighborhoods. Second, the lieutenants and captains act as “ambassadors” to the various communities that make up Chiсago, of which the Hispanic community is an important one. Effective police work, including the detection and apprehension of criminals, requires that the police have the trust of that community and they are more likely to have it if they have “ambassadors” to the community of the same ethnicity. It is true that Denk does not hаve a Hispanic name, and for all we know does not even speak Spanish. But like William Richardson, President Clinton’s last Secretary of Transportation, Denk is accepted by the Hispanic community as a member of the community on the basis of the fact that his mother is Hispanic; there is an analogy to Jewish religious law, under which a person whose mother is Jewish is Jewish regardless of the father’s religion.

What we have recited are the facts and it remains to consider whether they justified discriminating in favor of Denk. As explained earlier, this is a judgment for us to make without deference to the district judge. Justifications of discrimination that are based on a public employer’s Operational needs are suspect, because they seem to have no natural limits, unlike remedial justifications, which cease when the last traces of the discrimination that gave rise to the remedy have been eliminated. Some discrimination, whether of the old-fashioned kind or the modern “affirmative action” kind, is vicious, ignorant, political, or otherwise invidious, but much is not. To allow discrimination on the basis that it was efficient or expedient would cause inroads into equal protection that the courts are unwilling to countenance.

As we pointed out in Builders Ass’n of Greater Chicago v. Cook County, 256 F.3d 642, 644 (7th Cir.2001), the question whether nonremedial justifications for affirmative action can ever satisfy the equal protection clause has in the absence of definitive resolution by the Supreme Court caused bitter divisions in the lower federal courts. Many courts, however, including our own have at least left open a small window for forms of discrimination that are supported by compelling public safety concerns, such as affirmative actiоn in the staffing of police departments and correctional institutions. Wittmer v. Peters, supra, 87 F.3d at 920-21; Barhold v. Rodriguez, 863 F.2d 233, 238 (2d Cir.1988); Talbert v. City of Richmond, 648 F.2d 925, 928-32 (4th Cir.1981); Detroit Police Officers’ Ass’n v. Young, 608 F.2d 671, 695-96 (6th Cir.1979). Especially in a period of heightened public concern with the dangers posed by international terrorism, effective police work must be reckoned a national priority that justifies some sacrifice of competing interests. If it is indeed the case that promoting one Hispanic police sergeant out of order is important to the effectiveness of the Chicago police in protecting the people of the city from crime, the fact that this out-of-order promotion technically is “racial discrimination,” though its impact, incidence, and motivаtion are remote from the impact, incidence, and motivation that have shaped the-current legal view of racial discrimination, does not strike us as an impressive counter-weight.

The imperative need for this discrimination had to be proved and not merely conjectured. It would not have done for the City merely to. hаve presented plausible argumentation or to have appealed merely to common sense. It had to substantiate its position with evidence. It did so. It proved that it has a compelling need to increase the number of Hispanic lieutenants; and the increase it defended — the promotion of one Hispanic sergeant — is the smallest increase it could have made.

The judgment is affirmed in part and reversed in part, and the case ‍‌​‌​​‌‌‌‌‌​‌​‌​​‌‌​​‌‌​​​‌​​​‌​​‌​​‌‌​‌‌​​​​‌​‌‌‍is remanded with instructions to enter judgment for the City.